ies on Innovative Intelligence - Vol. 8 9
.9
< 5 ^
Learning Support System JU7
Organizational Learning
Joachim P. Hasebrook Hermann A. Maurer
Learning Support Systems for
Organizational Learning
Series on Innovative Intelligence Editor: L. C. Jain (University of South Australia) Published: Vol. 1
Virtual Environments for Teaching and Learning (eds. L. C. Jain, R. J. Howlett, N. S. Ichalkaranje & G. Tonfoni)
Vol. 2
Advances in Intelligent Systems for Defence (eds. L. C. Jain, N. S. Ichalkaranje & G. Tonfoni)
Vol. 3
Internet-Based Intelligent Information Processing Systems (eds. R. J. Howlett, N. S. Ichalkaranje, L. C. Jain & G. Tonfoni)
Vol. 4
Neural Networks for Intelligent Signal Processing (A. Zaknich)
Vol. 5
Complex Valued Neural Networks: Theories and Applications (ed. A. Hirose)
Vol. 6
Intelligent and Other Computational Techniques in Insurance (eds. A. F. Shapiro & L. C. Jain)
Vol. 7
Intelligent Watermarking Techniques (eds. J.-S. Pan, H.-C. Huang & L. C. Jain)
Forthcoming Titles: Biology and Logic-Based Applied Machine Intelligence: Theory and Applications (A. Konar & L. C. Jain) Levels of Evolutionary Adaptation for Fuzzy Agents (G. Resconi & L. C. Jain)
Series on Innovative Intelligence - Vol 8
Learning Support Systems for
Organizational Learning
Joachim P. Hasebrook University of Luebeck, Germany
Hermann A. Maurer Technical University of Graz, Austria
\[p World Scientific NEW JERSEY • LONDON • SINGAPORE • B E I J I N G - SHANGHAI • HONGKONG • TAIPEI • CHENNAI
Published by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: Suite 202, 1060 Main Street, River Edge, NJ 07661 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.
LEARNING SUPPORT SYSTEMS FOR ORGANIZATIONAL LEARNING Copyright © 2004 by World Scientific Publishing Co. Pte. Ltd. All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.
For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.
ISBN 981-238-831-1
Printed in Singapore by World Scientific Printers (S) Pte Ltd
This book is dedicated to Nils.
This page is intentionally left blank
Preface
The chapters compiled in this book are based on articles and projects reflecting the implementation and evaluation of learning support systems and applied scientific research in the last seven years, 1997 to 2003. Most articles have been reviewed, mostly peer reviewed, and published in scientific journals or volumes, respectively. However, they are not always homogenous because the papers accompany and summarize relevant sections of our work with different projects in corporate, educational, and scientific institutions. We have to apologize, if this book does not always show the full coherence and homogeneity of an original scientific publication. However, we are optimistic that it is worthwhile to work through the text of this book, nonetheless: It clearly reflects not so much a scientific research program but the development of learning and information systems (in mostly) European academic and business environments. Mostly, the examples described here are taken from our work with the German Ministry of Labor, major European private banking institutions, Austrian academic organizations and a number of international companies. We want to thank all those who helped us to put together this book: The first author, Joachim Hasebrook, would like to thank Prof. Dr. Dr. Hermann Maurer, who encouraged him to become a member of international program committees and to write papers about his work; Maurer was also a thoughtful and helpful mentor during his academic career. He gratefully acknowledges the opportunity to teach courses for the online program 'Master of Distance Education' of University of Maryland University College (UMUC) and to become a member of
Vll
Preface
Vlll
UMUC's faculty; Gene Rubin, director of UMUC's online programs, and Dr. Urich Bernath, director of distance education at the University of Oldenburg, gave him this opportunity. He would like to thank his friends and colleagues who assisted in studies and statistical analyses reported here for their help, namely Prof. Dr. Gerd Doeben-Henisch, Dr. Louwrence Erasmus, Markus Gremm, Wolfgang Nathusius, and Jiirgen Wagner. Bank Academy, the non-profit organization for ongoing education of the German bank associations, has been a supportive and exciting work place. The director of the board of Bank Academy, Prof. Dr. Udo Steffens, and the member of the board of Commerzbank and director of the supervisory board of efiport Inc., Klaus Miiller-Gebel, gave him the chance to work in the new and emerging field of 'elearning' and to become a member of the board of the educational financial portal [efiport] AG, the e-learning company of Bank Academy and the major German private banks. The second author, Hermann Maurer, would like to thank Prof. Dr. Joachim for invaluable discussions and for the possibility to contribute in this book, albeit in a minor way; he wants to thank co-authors of his papers used as basis of some material in this book, particularly Eva Heinrich and Ron Oliver. He is very much indebted to Thomas Dietinger and Frank Kappe for the support of Hyperwave, and to Nick Sherbakov for many invaluable inputs. We finally want to make it clear that the basis of this book has been a set of papers by the two authors, with much additional material added and updated. The papers at issue are: Hasebrook, J., & Nathusius, W. (1997). An expert advisor for vocational guidance. Journal of Artificial Intelligence in Education. 8(1), 21-41. Hasebrook, J., & Gremm, M. (1999). Multimedia for vocational guidance: Effects of testing, videos, and photography on acceptance and recall. Journal of Educational Multimedia and Hypermedia, 8(2), 217-240. Hasebrook, J. (1999). Exploring electronic media and the human mind: A Web-based training. World Conference on Internet, Intranet and World Wide Web (WebNet), Honolulu, Hawaii. Hasebrook, J. (1999). Searching the web without losing the mind - traveling the knowledge space. WebNet Journal, 1(2), 24-32.
Preface
IX
Hasebrook, J. (1999). Web-based training, performance, and controlling. Journal of Network and Computer Applications, 22, 51-64. Hasebrook, J. (2000). Knowledge workers and knowledge robots. Invited paper. Proceedings of International Conference of Computer in Education (ICCE), Taipeh, Taiwan. Hasebrook, J. (2001). Learning for the learning organization. Journal for Universal Computer Science, 7(6), 472-487. Hasebrook, J. (2002). Cooperative and interactive distance learning: application of team-oriented and selective learning strategies in a European bank. Journal of Universal Computer Science, 8(9), 834-847. Heinrich, E. & Maurer, H. (2000). Active documents: concept, implementation and applications . Journal of Universal Computer Science, 6 (12), 1197-1202. Maurer, H. & Oliver, R. (2003). The future of PCs and implications on society. Journal of Universal Computer Science, 9(4), 300-308. Maurer, H. (2003, in press). Necessary aspects of quality in e-learning systems. Proceedings of
Quality in eLearning Conference, Geelong University, Australia,
February 2003.
We gratefully ackowledge the permission to reprint parts of the afore mentioned articles. Especially, we would like to thank the Association for the Advancement of Computer in Education (AACE, see www.aace.org), namely Gary Marks, and the editors of the Journal of Universal Computer Science (JUCS, see www.jucs.org) and the Journal of Network and Computer Applications (JNAC). Additionally, we cited some figures and tables from the following recent works of ours: Hasebrook, J., Rudolph, D.W. & Steffens, U. (2002). E-Learning Business Strategies & Opportunities. Chichester (MI): Datacom Research Report. Hasebrook, J., Herrmann, W. & Rudolph, D. (2003). European perspectives for elearning: Markets, technologies, and strategies. Thessaloniki: CEDEFOP (European Centre for Vocational Training).
Joachim P. Hasebrook & Hermann A. Maurer
Contents
Preface Prologue: Key Trends in E-Learning Benefits from Technology Key Enabling Technologies Key E-Learning Markets E-Learning for Economic Development Public and Private Expenditures for Education E-Learning in Developing Countries Developing Regions: Asia and Africa Access to Electronic Learning in Asia Access to Electronic Learning in Africa Advantages of E-Learning in Developing Countries Learning Support in Organizations
vii 5 7 8 13 17 19 20 22 22 24 26 28
Part 1: Managment Support: Introduction Beyond the Learning Organization
31 31
1. Implementing Organizational Learning: Learning for the Learning Organization Knowledge, Technology, Strategy Web-Based Learning: E-Learning Learning Organization and Organizational Learning The Market of Knowledge Applications Knowledge and Abilities Psychological Factors of Success Competences on the Balance Sheet Value Extraction and Value Creation Capital versus Talent
37 37 41 42 46 48 52 56 59 62
2. Implementing Educational Controlling: Web-Based Training, Performance, and Controlling More Training for Less Money The Cost of Training Learning Efficacy and Cost Efficacy The Process of Controlling Calculating Success Gains from Goal-Directed Planning
65 65 66 68 71 75 78
1
2
Learning Support Systems
Part 2: Performance Support: Introduction Quality of E-Learning Environments Active Documents and Active Communication Knowledge Management
81 81 83 85
3. Implementing Web-Based Training: Exploring Electronic Media and the Human Mind From CD-ROM to Internet From Help Pages to Performance Support Systems Learning to Learn The Role of Meta-Cognition Integrating Performance Support in Learning Systems Generic Performance Support
91 91 92 94 96 98 100
4. Implementing Electronic Courses: Collaborative and Interactive Distance Learning Collaborative Learning with Electronic Media The Notion of Active Documents Implementation of Active Documents The Heuristic Approach The Iconic Approach The Linguistic Approach Applications of Active Documents The Learning Environment Experiment 1: Collaborative Learning Strategies Participants of Experiment 1 Material and Procedures of Experiment 1 Design of Experiment 1 Results of Experiment 1 Comparison of WBT and Seminar Factors of Online Learning Expert Participation Team Interviews and Discussion Learning Culture
103 103 104 105 105 106 107 107 109 111 Ill Ill 112 113 113 115 118 118 119
5. Implementing Online Curricula: New and Emerging Media in Distance Education Learning Support Systems Event and Learning Management Conducting the NEMDE Course Computer Mediated Expert Communication Learning Strategies Course Structure and Objectives
121 121 123 129 130 131 134
Contents Course Objectives Course Development Assignments Focused Discussions Effective Tutoring
3 134 139 145 148 152
Part 3: Decision Support: Introduction
155
6. Implementing Expert Guidance: Expert Advisor for Vocational Guidance Career Decision Making Models for Vocational Guidance Vocational Interests of Young Adults Knowledge and System Engineering for Vocational Guidance Calculating the Goodness of Fit Matching Careers to Individual Interests Implementation of the Expert Advisor System Implementation and Product Development Evaluation of the Expert Advisor Career and System Options
159 159 161 161 162 164 165 167 170 171 177
7. Implementing Adaptive Multimedia: Effects of Individualized Testing, Videos, and Photography on Acceptance and Recall Media Effects Career Counseling Pilot Study Experiment 2: Photo and Video Participants of Experiment 2 Design of Experiment 2 Materials and Procedure of Experiment 2 Results of Experiment 2 Summary of Experiment 2 Experiment 3: Individual Information Participants of Experiment 3 Design of Experiment 3 Materials and Procedure of Experiment 3 Results of Experiment 3 Summary of Experiment 3 Field study: Comparing Electronic and Printed Media Summary of Field Study Mental Integration of Multiple Media
179 179 181 184 184 184 185 185 186 189 190 190 190 191 191 193 194 196 196
4
Learning Support Systems
Part 4: Self Learning Systems: Introduction
201
8. Implementing Knowledge Structures: Searching the Web Without Losing One's Mind Visionary Terabytes Reality Bites The Myths of Multimedia and Hypermedia Myth 1: More Media Leads to More Learning Myth 2: Hypertexts Convey Structural Knowledge Myth 3: Web is Easy, Print is Tough Complexity of Models and Reality SHOEs for Web Walkers Traveling Agents: Knowledge Robots Educating Knowbots for Education
207 207 208 210 211 212 213 215 217 219 221
9. Implementing Knowlegde Robots: Knowledge Robots for Knowlegde Workers Entering the Infoverse If It Works, It's not AI Neuroscience Aspect Information Science Aspect The Age of Intelligent Machines
225 225 227 228 234 238
Epilogue: Future Developments E-Assisted E-Learning in 2010 The Wizard in the Glasses E-Assisted E-Learning in the Future Virtual Keyboards Global and Culture-Fair Communication PDAs Revisited What We May Learn Planets of Learning Ecology of Mind Virtual Minds Maps and Minds Invisible Computing and Embedded Learning
;.. 241 241 243 245 245 246 248 249 251 253 254 256 258
Appendices
263
Bibliography
269
Index
287
Prologue
Key Trends in Global E-Learning
The major trends with the biggest impact on the global e-learning markets and learning support technologies are. * the increasing demand for academic degrees, * growing numbers of students attracted to educational hubs, and « the rapid growth of non-traditional, especially elderly, target groups. We are convinced that all effective e-learning scenarios will be centered around personal tutoring. Sustainable e-learning efforts will need sufficient private and public financing. Regular content updates by skilled subject matter experts as well as careful control of the didactical quality of the delivered content will be essential. Additionally, costeffective e-learning will only emerge from already existing systems and processes, such as corporate databases, human resource management or public administration and 'e-government'. In general, e-learning will make education more effective but not better, because technology is aimed to enhance the efficacy of processes whereas didactics' objectives are to enhance the quality of the steps and tools involved in the learning process. E-learning mostly is a piece or a system of software, although some hardware - like computers and networks - always has to be involved. Efficient software provides the opportunity to be more scaleable, flexible and personalized than without adequate software. Hardware, however, is measured - according to Moore's law - in terms of cost per unit (e.g. the price for one million instructions per second). Unfortunately, e-learning software has been 'sold' to the educational markets like a piece of hardware, promising it would cut costs for travel, accommodation, personnel and delivery of 5
6
Learning Support Systems
content. Many e-learning vendors, however, painstakingly learned that labor intensive tutoring, didactical adequate media and up-to-date contents are costly key success factors for e-learning. In a corporate environment, e-learning will fail like other forms of electronically supported learning, such as computer-based training (CBT), if it cannot become an integrated part of corporate knowledge and human resources management. In public and academic environments, e-learning will only flourish if it does not add too much effort and costs to the processes in place. E-learning will not be a 'killer application' for the further expansion of international markets for electronic devices. Instead, e-learning has to become one of the key drivers of a rapid international knowledge transmission and transition. This will lead to accelerated economic development and will give a multilingual and multicultural society incredible opportunities to support the creation of global alliances and wealth. One of our core assumptions is that e-learning does not replace traditional classroom education. Instead, it expands the market for education products and services. Thus, e-learning assists the growing population of non-traditional learners, many of whom must divide their time between work and school, to pursue an education. Further, e-learning solutions can be applied to non-education markets such as public relations, sales, and investor relations. The same tools developed to facilitate imparting knowledge to students can be used with great effect in persuading customers, investors, and commentators. Corporate training, career development, and expert enhancement are areas ripe for sustainable growth. E-learning facilitates cost-effective production and delivery of courses for specific companies, jobs, and skills. E-learning technology enables course authors and producers to readily re-use content in different courses or different versions of the same courses. Thus, e-learning courses can be more customized than traditional classroom teaching. The ability to address special needs with minimal effort ensures the broadest possible market for any given content. While there has been much interest in using the Internet for 'distance learning,' its use as a global distribution channel presents a much bigger opportunity. Highly specialized courses for which there is
Prologue
1
insufficient local demand may do well in the global market. The Internet can also increase the success of courses that do well locally. Benefits from Technology E-learning will change our minds about how much education we need, and when and where learning can take place. When education is a purely local affair, highly specialized courses are sometimes not viable due to insufficient enrolment. The ability to offer such courses to the global market makes a difference. There are also many people who would like to take courses but who do not have the time or cannot commit to attending a regular class. Education is already a big business. E-learning, by making it easy to impart information and skills to anyone, anywhere, anytime, and for any purpose will grow the education market. As always, the big winners will be those vendors that identify and serve emerging and sometimes hidden markets. Recently, the 'Organization for Economic Co-Operation and Development' (OECD; www.oecd.org) published a report e-learning: the partnership challenge (OECD, 2001) examining the status and growth perspectives of electronically supported learning and skills development in all 30 member countries and some of the more than 70 associate countries. The key findings concerning possible benefits of use of information and communication technologies (ICT) are listed in the following table 1. E-learning technology can be used almost anywhere and anytime. The lines between traditional education, self-improvement, and marketing are being blurred - just as the line between education and entertainment has blurred. The biggest growth segments unleashed by e-learning are education for non-traditional students and the use of educational methods in related areas such as public relations, sales, and investor relations. E-learning permits dramatic expansion of the education market. While 'distance learning' is the best-known example, we believe providing continuing education for busy professionals is even a much bigger opportunity.
8
Learning Support Systems
E-learning is primarily about superior solutions for self-study and online courses. These solutions, however, can be readily adapted to sales and public and investor relations. In both cases, the object is to get information across to the recipient. While the education industry correctly emphasizes the learner, that does not mean there is no longer a need for teachers. Teaching and persuading have many things in common and can share many of the same advanced tools. Corporate training and personal career development are segments ripe for considerable growth. Businesses need to impart both general skills and company-specific information to their employees. They need solutions that are highly reliable, consistent, and available in order to bring new employees quickly up to the required level of competency. Increasingly, corporations are realizing that the Internet and extranets can be used to train customers and business partners, as well as persuade investors, consultants, industry analysts, and potential customers. Table 1. Benefits of using ICT to deliver learning (OECD, 2001, p. 23) Things that cannot be done without technology the de-materialization of time and space - learning any time anywhere mass-education - access to learning for everyone Internet access to ever growing collections of educational resources and services input for task-based learning using fast search and retrieval software, or for research work learning on demand peer-group teaching / learning through distance learning via ICT Things which can be done better with technology the choice of learning style customized and personalized learning materials and services individualized tracking and recording of learning processes self-assessment and monitoring of learner performance interactive communications between participants and influences in the learning process interactive access to educational resources
Key Enabling Technologies Technology penetration has been greatest in the workplace, although other sites such as homes and community centers are increasingly wired
9
Prologue
up (see figure 1). OECD countries greatly increased their personal computer (PC) base in the 1990s with an average number of PCs installed per 100 inhabitants rising from 10 in 1992 to 24 five years later. In 1997, the Nordic countries, Switzerland, Australia, and the Netherlands had a higher ratio than all G7 countries but the US (OECD, 2001).
• 1992 total • 1997 total M1997 education
CO LI.
8 0)
•o
CD
F
£
(T)
Ger
b
8 c
O
ro
CO ,=
CO >
co
.E
D)
CO
> I s CO
^-
Q_
w
Z3
XI
? CO
Figure 1. Average PC installed base per 100 inhabitants and share in the home and education in the OECD area, 1992 and 1997(OECD, 2001; p. 13).
OECD countries have invested massively in ICT in education over the past four years. The OECD estimated that over the OECD area as a whole approximately 16 billion Euro were invested in 1999 - which is still not more than 1 % to 2 % of all education spending. Most of the investment has so far been in hardware and infrastructure. Figure 2 displays the percentage of households possessing a personal computer and presents the ratio of students per PC in upper secondary education for various OECD countries in 1998. As shown in figure 2, the percentage of households with a computer ranges from a high of 63 % in Denmark to a low of 20 % in Italy. Additionally, the same figure shows that the number of upper secondary students per PC varies from 4 in Norway to 35 in Portugal. Both figures reveal that there is no positive
Learning Support Systems
10
correlation in some countries between high coverage of PCs in homes and coverage of computers in upper secondary schools.
• Home • School
O
O
Z
.E ra Q. W
"ra
~xL
>
)
Sweden
8
Port
(0 CD 3
Figure 2. Home (percentage of households possessing a PC) and school (students per computer in upper secondary education) access to computers in selected OECD countries, 1998 (OECD, 2001, p.19).
Figures provided in the Kerrey report of the Web-Based Education Commission (2000) underline the rapid rising levels of school connectivity in the US, which grew from 65 % in 1996 to 95 % in 1999; classroom connectivity over the same period has increased from 14 % to 63 %. The same report has drawn attention to the fact that there is a huge gap between rich and poor schools within the same country or state. In the post-secondary sector, the Internet access is even more spectacular. Facts available for 1998 (Merrill Lynch, 2000) demonstrate that more than 50 % of college students in the US will have had Internet access from their homes and a virtually 100 % from campus. Over 90 % of college students access the Internet - with 50 % on a daily basis. At faculty level, nearly 40 % of all college courses are using Internet resources compared with 15 % in 1996. The OECD report (2001) clearly states that many European countries are still lagging behind the rapid
Prologue
11
development in the US, although Europe as a whole is quite close to international ICT use and access levels. Figure 3 lists figures from spring 2002 how many students in primary and secondary schools have to share one PC according to statistics published by Gallup Europe.
Figure 3. Students per PCs in primary and secondary schools in selected European countries (Gallup Europe, Eurobarometer Flash 118; spring 2000).
The biggest opportunities for e-learning are for content producers and distributors. E-learning enables those who have knowledge to reach a wider audience, both in terms of geographic reach and audience composition through the use of different versions or packages. There are also opportunities for software developers: Learning and testing platforms, authoring tools, and virtual campuses are among the best sold software solutions. Online courses and mobile learning create opportunities for software that is highly interactive and can intelligently suspend and resume lessons as needed. There is still much to be done in development of custom courseware that adapts to the learner's needs, knowledge, and skill set. Given the size and importance of the education market, one should not rule out the need for special hardware. For example, e-books should be adapted to the needs of students and
12
Learning Support Systems
trainees. Education, rather than the convenience of reading a flat computer screen in bed, may prove to be the bigger market for e-books. The gradual development of e-learning standards will lead to interoperability between content, learning platforms, and testing. This, in turn, will make it easier and less costly to develop courses on highly specialized topics and for niche markets. Thus, the best tools will be available and affordable for use with all topics. Networks will play a particularly important role, making it possible to deliver content and courses in new ways, and changing the way students, teachers, and others interact. The public Internet is central, but it is not the whole story. Private networks and virtual private networks (VPNs) will also play significant roles. The Web is both a global marketing channel and a delivery vehicle. While the lack of significant local markets has been an obstacle to highly specialized courses, it is not the only obstacle. Just finding such courses can be a major problem. For the first time, there is a fast and convenient way to both identify and access specialized courses. The Internet can be used for real-time classes and non-real-time classes - or a combination of the two. LANs enable classroom automation, and can be used to enhance the educational value of conferences and seminars. While schools have invested heavily in personal computers, they often put them in computer labs where students are taught how to operate them - but for no specific purposes. Wireless LANs enable distributing PCs throughout the school and moving them around as needed. Virtual private networks enable education institutions to create virtual campuses with whatever geographic reach is desired. VPNs also enable enterprises to extend education to workers, business partners, and resellers in the field. The continued growth of VPNs, however, depends on the industry's ability to prevent major security breaches. There are numerous opportunities for e-learning portals, online universities, and mobile learning (m-learning) services. These service providers can also benefit from specialized tools. As telecom operators strive to move up the value chain, education and persuasion will prove key markets. We would not be surprised to see virtual network operators emerge, focusing exclusively on education and persuasion.
Prologue
13
Mobile learning will make it possible for busy professionals to pursue continuing education. Public wireless LANs and third generation, 3G, wireless networks will enable users to remotely access courses, content, and other learning resources [www.fcc.gov/3G/]. Though some 3G wireless networks have been delayed, others have not, and the migration to 3G networks that support high-speed data is inevitable. The development of multimedia content for education will create demand for better ways to search for and identify video and audio content. Thus, training and education in general may provide the stimulus for new products and services addressing this problem. The use of meta-data and pattern recognition software are promising methods for cataloguing audio and video content. E-learning products and services may also be enhanced through the use of speech recognition, text-to-speech, and natural language technologies. The use of voice technology is particularly useful in teaching and practicing foreign languages. Voice technology will extend the reach of online e-learning courses to any telephone. Thus, learners could study a few minutes here and there, in different places and at different times. For busy professionals, voice technology could prove helpful in squeezing learning into their schedules (for example, while commuting to and from work). Copyright protection and intellectual property rights are key concerns for the e-learning industry. A number of solutions, generally referred to as digital rights management (DRM), do more than just protect content from unauthorized access and distribution. DRM can be used to control the number of times that material is viewed, whether it can be printed, how long it is accessible, and so forth. Thus, DRM will enable content producers to block sales of 'used' content for which they are rarely compensated. Key E-Learning Markets The ICT industry is mirroring other industry sectors in the trend towards consolidation. Vertical integration is a phenomenon affecting parts of the education market as well as the media market. Merrill Lynch (2000)
14
Learning Support Systems
identifies some potential impacts, which are possible to follow as the consolidation trend starts to influence the education and training markets (see Table 2). Table 2. Prospects for e-learning technology in various market sectors (Merrill Lynch, 2000; quoted in OECD, 2001, p. 33) Sector Early education
K-12 education
Post-secondary education Corporate training
Consumer/end user training
Impact As parents shift their children to center-based care, choosing providers with high quality educational content and brand names, we believe consolidation will finally come to the highly fragmented child care industry. The challenge of developing an effective sales channel into schools will be an asset and barrier to entry that will drive consolidation in this segment. The ability to access 'best in class' education from anywhere at anytime will challenge existing colleges and universities. Look for more specialization and consolidation as a result. The ability to offer a complete training solution to corporations taking a more comprehensive and proactive approach to training will encourage consolidation. Competition for retail shelf space and consumer 'mindshare' will drive consolidation and partnering in consumer products and services.
However, consolidation has not ended, yet. The educational and training markets are in transition and transformation into a more mature market model. In Table 3 we try to give a detailed review of the current status a growth prospects of the relevant market sectors (cf. Hasebrook, Rudolph & Steffens, 2002, pp. 12).
Prologue
15
Table 3. Status and growth prospects for e-learning technology in various market sectors (Hasebrook, Rudolph & Steffens, 2002, pp. 12) 1.5-3 years
0-1.5 years Traditional K-12
3 - 5 years
US Europe
medium low
medium low
low low
K-12 supplementary
US Europe
high low
high medium
high high
University undergraduate students
US
medium
medium
high
Europe
low
low
medium
University graduate students
US
low
medium
high*
Europe
low
low
low
US
medium
high
high
Europe
medium
high
high
Corporate training
Comments
Disillusionment in the US. Europe never really gets started. Homework assistance and special services for gifted children. Europe lags. Mainly e-books, online tutoring, and later licensed video lectures supplementing textbooks. In Europe: limited willingness to pay for education; teaching content continues to be less standardized. Low domestic potential due to research orientation and large share of self-studying via free online content. *High enrolment potential for foreign students at US e-learning centers. Europe will gain negligible share in this market. After initial disillusionment with cost-savings and quality, this segment will come backstrong with more targeted applications, a higher share of instructor-led courses, and quality content. Little difference between major US and European corporations.
16 Non-traditional Undergraduate
Learning Support Systems
US Europe
high medium
high medium
high high
Graduate
US Europe
high medium
high medium
high* high
Professional (updating knowledge)
US Europe
medium medium
medium medium
high high
Expert (seeking specialised knowledge)
US
low
medium
high
Europe
low
medium
high
Employee training, post M&A
US
medium
high
high
Europe
medium
high
high
Client/custom er training
US Europe
high high
high high
high high
Consumer (personal interest)
US
low
low
Europe
low
low
mediu m mediu m
Growth is mainly supply rather than demand constrained. Europe lags. Same as undergraduate plus: *High enrolment potential among non-traditional European students for (online) US e-learning universities. Big potential in medical research, law, finance, cross cultural marketing. Mainly constrained by dearth of experts with e-learning course design skills. Online tutoring by experts in niche applications such as data mining, engineering, quantitative finance, business plan design. Constrained by low market transparency and dearth of experts. After M&A, corporate-wide learning is often a matter of economic survival. Great potential in fields such as banking, insurance, automobile, aviation, hotel chains, chemicals, telecom services, software, consulting. Will become part of customer relationship management (CRM). Primary target group: age 55+. Big long term growth potential, but only modest medium term growth. Exceptions: financial planning and investing, foreign languages, and technology enabled communication among similar-minded, educated people.
Prologue Non-education US Public relations
low
low
medium
Europe
low
low
medium
Investor relations
US Europe
high high
high high
high high
Political campaigns
US
medium
medium
high
Europe
medium
medium
medium
17
Long term growth potential for select applications including (private sector) hospital and nursery home evaluation, school and college evaluation, environmental efforts of oil and chemical companies, and working conditions in high growth firms. Potential in public sector: SEC, consumer protection (e.g. mad cow disease), public pension plans, and tax law changes. This is the most promising application for e-learning outside traditional education. Offers potential for small parties and special interest groups (e.g. agriculture) exercising greater pressure.
E-Learning for Economic Development While many segments of the Internet economy are struggling with the severe downturn of the industry, the various e-learning initiatives targeted at developing countries are experiencing accelerating support from various institutions. Most notably is the commitment of the World Bank and the Australian government. The magazine University business [at www.universitybusiness.com] reported in its September 2001 issue that both institutions formed a partnership that will bring distance education via state-of-the-art technology to developing countries. The two institutions will spend 1.5 billion US-Dollar to promote education in developing countries. The initiative will cover the entire range from training of primary-school teachers to advanced courses for policymakers. The World Bank devotes 1.3 billion US-Dollar over five years and AUSAid contributes 200 US-Dollar million to the new program called 'Virtual Colombo Plan.' The VCP is the virtual remake of an Australian educational program, called the 'Colombo Plan', launched in
18
Learning Support Systems
the 1950s that brought students from abroad to Australian universities. The objective of the VCP now is to bring the Australian universities to students living abroad. On a global scale, the World Bank's educational consulting arm Edlnvest attempts to facilitate education entrepreneurship and international investment in developing countries. Apart from the many advantages that e-learning offers for transition and developing countries in delivering education, there are several important driving forces that will foster the growth of these emerging e-learning markets over the next decade. Some of these demand driven factors can be briefly summarized as follows: In developing countries, there is a large gap in incomes that induces many of the highly qualified teachers to migrate to OECD countries. As a result of this, transition and developing countries continue to experience what has been called 'brain drain.' E-learning could counteract these tendencies and give these countries a fertile 'brain gain'. Regional income disparities continue to be a source of concern for policy-makers. While it is true that regional differences are reflected in Internet connectivity, it is also true that once a classroom is wired, its students have equal access to the very same educational resources that are available on the Internet as those who live in the most privileged areas of the world. Corporations continue to internationalize their business operations and their market scope. They have to deliver training to their employees and sales force in many different countries. These international corporations need employees with certified qualifications from high quality education providers. In many developing countries, such education providers are not available locally. IT prices continue to fall worldwide and make it affordable for larger shares of the world population to become wired. For example, the US producer price index for computers declined in 1999 by nearly 20 %. Since 1997 that decrease was no less than 20 % in each of these years. The other major cost component of Internet access are the rates for telecommunication. Here too, prices are coming down all around the world. Deregulation and competition make this happen. For example, until 1999 it cost around 300 Euro in the Philippines to establish a land line telephone connection. Today, it only costs 50 Euro. Despite this
Prologue
19
dramatic drop of ICT costs low developed countries still pay a 'poverty premium' for access to basic ICT services (Prahalad & Hammond, 2002). It is this 'poverty premium' and the fast growing number of inhabitants in the low developed regions which stand for the biggest opportunity and the largest 'digital divide. By far, the biggest 'brain drain' we are witnessing on a global scale is the fact that women in low developed regions are not educated properly: In some regions of Central Asia and Africa the average time of schooling for men is two to five years, the average time of schooling for women is below two years or even one year. Compare this to the about ten to twelve years of schooling for women and men alike in the high developed countries. Linking people to knowledge is the key success factor to link them to profit, as well. Linking as many people as possible to profit, then, will be the key success factor to grow the global economy (Rosenzweig, 1998). Public and Private Expenditures for Education Despite the rhetoric about the central importance of education for the competitiveness of countries, the overall public expenditure for education as a proportion of GNP more or less stagnated between 1980 and 1997. In developed countries the share (5.1 %) remained unchanged. In the developing countries this share was substantially lower (3.9 % in 1997) and similarly stable (3.8 % in 1980). Most alarmingly, the least developed countries that have to do the most in terms of catching up, have the lowest share of public expenditure for education as a proportion of Gross National Product (GNP) with 2.0 %. And this share has even declined from 2.8 % in 1980 to 2.0 % in 1997. A particularly important segment in the education industry for the growth prospects of e-learning is the private education sector. Private firms have a customer orientation and the economic freedom to make decisions and can, therefore, expected to be among the early adopters of this course delivery mode. As the graph depicted in figure 5 shows, it is not true that only highly developed countries have a share of the private sector that is comparable with the US private sector share. In the
Learning Support Systems
20
schooling market the share of the private segment is substantial in several developing countries. E-learning in Developing Countries International trade in education is already a reality. For example, students participating in the e-learning courses at the University of California Extension come from many different countries such as Russia, Mexico, Japan, and Canada, and at the online campus of the University of Phoenix, students from 21 different countries are enrolled. E-learning eliminates trade barriers to export education abroad. Before the emergence of e-learning, students had to leave their home country, if they were either looking for a highly specialized degree program that was not available in their home countries, or if they were born in less developed countries. In both cases, they typically had to pay much higher costs of living during their studies than at home.
World Avg.
Developed Countried
Developing Countries
Least Developed Countries
Figure 4. Trends in public expenditure on education as a proportion of GNP in % (UNESCO World education report, 2000; from www.unesco.org/courier).
To get an idea of the real market potential for high quality postsecondary education outside the US and the willingness to pay for education of students abroad, it is instructive to take a look at the foreign student population in the US: Nine out of the 15 countries with the largest foreign student population come from Asia. And these countries account for almost half of the foreign student population in the US.
21
Prologue
China alone sends more students to the US than all western European countries combined. For educational markets, it is the number of people that matters for the potential of investment opportunities. Here, Asia is by far the number one foreign market for education. In the US, international education is an important economic sector. Based on data about cost of living and tuition expenses provided by the Association of International Educators, [NASFA, see: www.opendoorsweb.org], the total combined foreign expenditures on tuition and cost-of-living exceeded 12.3 billion US-Dollar in 1999. Many of the most talented foreign students receive scholarships from US universities and colleges. However, the vast majority of international students (74.7 %) receive most of their funding from sources outside the United States.
India
Argentine Indonesia
Thailand Phillipines
Brazil
Egypt
Figure 5. Percentage of private primary and secondary school enrolment in selected countries in 1997 (OECD education database, 2000; from www.unesco.org/courier).
Especially older students that have already job experience and seek a higher degree are the primary target groups for e-learning. While generalizations are very difficult to make, it might be safe to say that the older a student is, the more do academic and professional objectives dominate goals that have to do with developing one's personality. Among the countries of origin of the 15 largest student groups in the US two European countries (Russia and France) emerge, which have relatively high shares of graduate students. Also, the share of graduate students among the student populations vary widely within this group of countries between 24 % (Indonesia) and China (81 %).
22
Learning Support Systems
Developing Regions: Asia and Africa
Access to Electronic Learning in Asia In Asia there is a growing number of firms and institutions that emerge in the e-learning industry. According to the eAsia Report, which is a study done by the New York City research firm eMarketer, Asia can be expected to have 61 million Internet users by the end of 2002. This will be 22 % of the world's total. The Internet community in China, for example, grew to 6.7 million by the end of 2000 (up from 890 000 users in 1998). The research firm IDP Education Australia predicts that the number of university students in Asia will reach 45 million by 2010 and rise to 87 million by 2025. NextEd Ltd, which is an Internet start-up in Hong Kong, estimates that the e-learning market in Asia accounts already for 6 billion US-Dollar and grows at a rate of 25 % per year. The potential for post-secondary education providers in Asia is enormous, but these emerging education markets are difficult to enter. For individual universities, colleges and commercial training institutes in OECD countries it is simply too time consuming to deal with accreditation and regulation issues in emerging market countries. For example, while China is the biggest market for higher education, delivering education in the PRC requires the approvals from at least three ministries - the Ministry of Education, the Ministry of Information, and the Ministry of Foreign Trade and Economic Cooperation. Two Snapshots of emerging e-learning firms, alliances and institutions show that e-learning is becoming a reality in the developing countries of Asia. It is certainly true that government regulation, relatively low bandwidth and high telecommunication prices (as compared to the US) are important impediments for more rapid growth of the new education economy in developing countries. But it is also true that one can cite numerous cases where private education firms in developing countries have been operating successfully for years. In February 2000, the Laguna College of Arts and Trade School in Manila, Philippines, Cisco Systems' Networking Academy, the World Bank Group and the International Youth Foundation announced the
Prologue
23
launch of their joint partnership project. Cisco Systems' Networking Academy program is an e-learning environment, delivering a foursemester curriculum targeted at high school, college and adult students. Cisco Networking Academies now operate in all 50 US states and in 60 countries worldwide. There are already more than 30 Cisco Networking Academies in the Philippines. These regional academies are located at Ateneo De Naga University, Philippines Science High School and University of Cebu [www.cisco.com/edu]. Table 4. Internet use in Asia (Source: Asiaweek, 18-25 August, 2000). Location Asia/Pacific
Brunei China Hong Kong India Indonesia Malaysia Philippines Singapore South Korea Sri Lanka Taiwan Vietnam
Number of users 75 million expected to rise by 2003 to 233 million 16 000 8 807 590 978 180 4 5 million 413 000 1.4 million 335 000 1011206 10 078 560 14 000 4 800 000 40 000
EduQuest [http://www.eq-campus.com] is an alliance between Asia Pacific College, IBM Philippines Inc., and SM Equicom Computers. The focus of the partnership is on developing curriculum content using courseware and e-learning technologies for the K-12 and higher education segment. The alliance grew out of a pilot project in 1986 to improve the curriculum and learning activities at a depressed, deprived and underprivileged elementary school in Metro Manila. The project was extremely successful and significantly raised students' achievements. By now, EduQuest partners with 29 private and government schools, colleges, universities and training institutions.
24
Learning Support Systems
Access to Electronic Learning in Africa A study of the Africa's Information Society Initiative (AISI) from July 2002 indicates that the use of the Internet has grown relatively fast in most urban areas in Africa, in much the same pattern as the adoption of the mobile phone which followed shortly after. As an indication, five years ago, only a handful of countries had local Internet access, now it is available in every capital city. But although these are encouraging trends, the differences between the development levels of Africa and the rest of the world are much wider in this area than they are using more traditional measures of development: Of the approximately 816 million people in Africa in 2001, it is estimated that only 1 in 4 have a radio (i.e. 205 million), 1 in 13 have a TV (i.e. 62 million), 1 in 35 have a mobile phone (i.e. 24 million), 1 in 40 have a fixed line (i.e. 20 million), 1 in 130 have a PC (i.e. 6 million), and 1 in 160 use the Internet (i.e. 5 million). Because of the large number of shared accounts, and the high use of public access services, it is difficult to measure the total numbers of Internet users. The rates of growth in users seen in the 1990s has slowed in most countries as the bulk of the users who can afford a computer and telephone have already obtained connections. As of mid 2002, the number of dialup Internet subscribers was close to 1.7 million, 20 % up from last year, mainly bolstered by growth in a few of the larger countries such as Egypt, South Africa, Morocco and Nigeria. In Africa, each computer with an Internet or email connection usually supports a range of three to five users. This puts current estimates of the total number of African Internet users at around 5-8 million. This is about one user for every 250 to 400 people, compared to a world average of about one user for every 15 people. The UNDP world development report figures for other developing regions in 2000 were: 1 in 30 for Latin America and the Caribbean, 1 in 250 for South Asia, 1 in 43 for East Asia, 1 in 166 for the Arab States. The lists and newsgroups are almost entirely hosted off-continent except for a number in South Africa, North Africa and Kenya. There is a list for almost every nation as well as others on more general topics ranging from African cinema to post colonialism. In the area of ICTs in Africa, AFRIK-IT is the only notable public list, and it is run from Ireland by the University College of Dublin.
Prologue
25
Figure 6. Connectivity and Internet access in Africa in 2002 (International Development Research Center, IDRC, from www.idrc.ca/acacia/divide/).
An Internet training program for institutes, schools and other agencies of higher learning in Francophone and Lusophone sub-Saharan African counties called Internet pour les ecoles inter-etat d'Afrique de I'ouest et du centre has been established in a related effort under the Diderot Initiative. Comnet~IT? established by the Commonwealth Secretariat (ComSec) in Malta, supports ICTs In Commonwealth developing countries, also a number of ICT support activities, such as the provision of scholarships for Commonwealth country students to obtain postgraduate degrees In computer science. The African Virtual University (AVU) provides training in computer and Internet applications and programming languages to its 29 university campuses In 18 countries In
26
Learning Support Systems
Africa [www.avu.com]. International volunteers are being seen as an increasingly important vehicle for training and technology transfer. This has been supported by UN volunteers and other similar nongovernmental organizations (NGO). Advantages of E-Learning in Developing Countries The list of highly advantageous features that the new learning technologies offer for e-learning applications with developing countries is long. Below we give a short list of some of the typical problems that beset the delivery of education to developing countries and the potential that e-learning offers for solving these problems. Table 5. Advantages of e-learning for education in developing countries (cf. Hasebrook, Rudolph & Herrmann, 2003, p. 25). Problem of delivering education to Problem solving potential of e-learning transition and developing countries High diversity of educational backgrounds The hyperlink structure of the Internet of learners make face-to-face teaching makes it easy and inexpensive to especially difficult and impede the learning incorporate free background material and effectiveness. glossaries that help to bring all learners up to the same standard. High travel costs and seminar fees result in The learning effectiveness of a hands-on rather infrequent and short courses. approach is much higher than learning Weekend seminars give little time for through listening. But a hands-on approach participants to solve exercises and sample takes time and cannot be taught in problems, which induce them to apply the weekend seminars and 'crash courses.' newly learnt techniques and concepts. E-learning is ideal for a 'learning-byWithout such exercises, the new knowledge doing' teaching mode that allows learners presented does not become part of to acquire the new knowledge step by step participants' active skills. and incorporate it into their daily work. Especially studying more advanced course In e-learning courses, students do no material which improves students' cognitive longer 'get stuck', they can always quickly knowledge that facilitates understanding ask for help. E-learning allows the frequent can be quite frustrating without sufficient and cost effective communication between communication with the instructor. Such the learner and instructor. Moreover, frustrations result when students 'get stuck' e-learning facilitates communication and they severely endanger the learning among students as well and, thereby, success. students can share their learning experiences and problems, which improves their motivation and learning success.
Prologue Learners often hesitate in face-to-face seminars to ask 'stupid questions.' Yet such hesitation hinders the learning effectiveness. Most potential learners do not know, what they do not know and what is out there that they could and should learn. In the past, education used to be one of the sectors with the lowest market transparency.
27
The barriers to ask such questions is much lower using e-mail than in class discussions.
A large number of e-learning portals provide unprecedented market transparency of educational offers. Online bookshops and the course syllabi posted on academic Web sites make it extremely easy to get an review of today's knowledge base for any field. The relatively high costs of textbooks make There is an enormous array of high quality it difficult for instructors and teachers in teaching resources on the Internet that is transition and developing countries to gain freely available, access to state-of-the-art teaching content. Human capital investment decisions are Modularization of course content allows insufficiently vertically integrated. As a the cost effective provision of continuing result, highly educated people find no education at all levels of the value chain adequate employment opportunities and hierarchy, because their potential co-workers lack basic education.
This short list already makes it clear that e-learning is an extremely promising innovation for fostering education in developing countries. So it should come at no surprise that the World Bank gives high priority to e-learning projects. Teaching and learning across cultures is an application that calls for a 'blended approach,' i.e. the combination of traditional face-to-face instruction and e-learning elements. Teachers and learners must learn from each other. Learning is anything but a one-way street. The very essence of learning when teachers and learners come from different cultures are the learning experiences from human interaction - only face-to-face meetings can provide such experiences. Most discussions of e-learning applications for developing countries take the rather narrow view of selling established degree programs or certificate courses without any alterations abroad. In other words, the functionality of e-learning is reduced to its cost advantages in delivering content. Little or no attention is paid to cultural differences in learning and teaching styles, the economic pressures that students face in less privileged countries,
28
Learning Support Systems
and the vastly different realities in developing countries that can turn the meaning of content developed in OECD countries upside down. Without major alterations of the curriculum, content and course design, the success of e-learning offers in developing countries must necessarily fall short of expectations. Learning Support in Organizations In what has been discussed so far, we tried to give a global perspective of digital media, access to ICT, and its use in education. In what follows, we want to focus on ICT to support learning processes fostering the success of corporate, governmental, and non-governmental organizations. We have to beware that technology solves technological problems, only. Learning, however, is not a technological problem but delivery of educational services just-in-time and on-demand is. Learning is a social process. Didactics aims at greater quality, technology at greater effectiveness. Therefore, ICT does not help to overcome didactical faults and shortcomings - it may even magnify them. According to Keil-Slawik (2003) we would like to distinguish three distinct levels of digital media in education: « The primary level provides general infrastructure and software features used for learning (e.g. Web browser). « The secondary level comprises specific software features for learning (e.g. testing facilities and learning management platforms). ® The tertiary level introduces self learning and adaptive software features (e.g. Intelligent Tutoring Systems, ITS, and knowledge robots). Knowledge is not a factor of production, such as financial capital or soil. A 'war for talent' has started indicating that knowledge and its proper application produces a more severe shortage to economic growth than the access to financial capital and financial mediation (Martin & Moldovenau, 2003). Knowledge has to be available like the ubiquitous power of electricity. Education is the transmitter of knowledge. New media are the latest technology for transmission. New media enables
Prologue
29
'wireless' education. Education can be carried into low developed regions using a lean physical infrastructure. Linking people to knowledge. Linking people to knowledge will link them to profits. Global trends indicate that new media can help to built the link. The structure of the book reflects the three levels or layers of digital media in education: Infrastructure, special software and systems, and self-learning systems. Each layer is adapted or applied to learning processes in organizations. Thus, the four main sections of this book are: ® Management support (infrastructure layer), * performance support (infrastructure and special system layer), * decision support systems (special system layer), and * self-learning systems (self-learning system layer). Additionally, the nine chapters subsumed under one of the four sections are proceeding from closely related issues, such as educational organization and controlling, to the implementation of course modules (or Web-Based Training, WBT), co-operative learning, and curricula; the later chapters are describing methods and applications for the adaptation and individualization of educational software as well as implementation of strategies for self-learning software agents.
This page is intentionally left blank
Parti
Management Support: Introduction
Beyond the Learning Organization On the one hand, we often think of organizations as persons who can learn (e.g. being a learning organization), and who can have social relationships (e.g. celebrating a wedding with another organization). On the other hand, psychology is widely ignored in the literature about organizations when computer scientists assume that organizations have much data, some information and hardly any knowledge and not assuming that persons in organizations have much more knowledge than the storage of any corporate database could hold. Economists of famous business schools examine the 'learning paradox' which is the unwillingness of employees to learn for a more effective organization while the most striking learning paradox is that this economic literature has ignored about fifty years of psychological research. Many psychological disciplines have contributed to a better understanding of organizations and the persons therein: Work and Organizational Psychology examines the relationship of individuals and their work environment including tools, teams, and workplaces. Social Psychology helps researchers to understand the conditions of social behavior within and between groups of persons - in the workplace or in entire nations. The Nobel prize winning research of Daniel Kahnemann and Amos Tversky (and others) has shown that the rational decision making process guiding the behavior of consumers or managers is a mere invention of economists in their ivory tower (1979, 1984): Real life 31
32
Learning Support Systems
decision making operates with incomplete data and under uncertainty; it is therefore based on simple rules and shortcuts or heuristics. It turns out, however, that the mostly selfish decisions made under uncertainty can result in perfectly rational behavior, quite frequently. Moreover, humans are not as selfish as many economists like to believe: Some of our best reasoning skills are dedicated to detect social fraught and to prevent peers in our social group from cheating. In general, we can deduce from research that human decision-making is not overall erroneous but strongly depends on the context and the format in which underlying data is provided. Cognitive Psychology teaches us that we are not the emotionally detached and rationally based 'homo economicus' who we might like to perceive ourselves. It is the realm of Clinical Psychology to remind us that there are strong links between emotion and cognition. Emotion is the basis on which all cognition is operating: Research has shown that we tend to make judgments about a person or a social situation within a few seconds; it may then take us hours or even days in order to rationally explain our decision to ourselves. We further need cognition and cognitive labeling in order to understand our own emotions and to make them a part of our conscious decision making process. Therefore, it seems reasonable that methods of diagnoses and treatments developed by clinical psychologists can help us understand how organizations can be treated effectively, and how problems in organizations can be diagnosed and handled properly. Whereas Cognitive Psychology has been applied to economics and the analysis of professional and corporate organizations, Clinical Psychology has not been adequately incorporated into the psychology of organizations. There are some exceptions which have been examined to some extend, such as mentor protege relationships, leader membership exchange (LMX), and negative social behavior, such as 'mobbing'. However, clinical psychological diagnoses and treatments are still mostly restricted to individuals outside the context of organizations or groups supposing that abnormal behavior or mental illness is an individual problem, which has to be treated individually. Family and Group Therapy has shown that there are ways to regard social context in order to facilitate diagnoses and to enhance therapies.
Learning Organization
33
We are suggesting that methods and research findings of clinical psychology can be and should be applied to organizations, such as private companies and governmental or non-governmental organizations. This can be accomplished in two ways: First, one can ask what positive impact clinical psychology can have on the psycho-diagnosis and psychotherapy within organizations. We suspect that rapid and permanent change, extreme workload, and high risk can impose 'traumatic events' of permanent distress and helplessness on people. Constantly changing management methods based on democratic and - at the same time - charismatic leadership as well as transparent and direct communication find their limits in the social skills of the managers sometimes eliciting social phobias or even more general anxiety disorders. Second, clinical psychology can provide a scientific framework to find answers to the question if and how organizational behavior equals individual behavior of persons or the behavior of individuals in social groups. Answers to this question will help to support general organizational competence, such as organizational learning, mental fitness, innovation and creativity, and social and emotional 'intelligence'. Clinical psychology is usually seen as the psychology of the (mostly impaired) individual - and therefore it has hardly been applied to professional organizations. The research about psychology in and for organizations, such as private companies, governmental and nongovernmental organizations (GO and NGO), has been conducted and discussed mostly outside the psychological expert community: Economics and sociology, management training, even applied computer science has adopted basic psychological research about cognition and emotion. The careful review of existing methods and results, the conduct of innovative and inter-disciplinary research, and the application and its sound evaluation in organizations might help scientific psychology to have another substantial impact on economics: The Nobel Prize winning research of Daniel Kahnemann and Amos Tversky has shown that cognitive psychology can considerably contribute to a better understanding of economics; the contributions of clinical psychology are yet to be discovered. In this book, however, we want to focus on methods
34
Learning Support Systems
and applications of computer science. Therefore, we want to start with a managerial perspective of learning organizations (if such a thing exists) and systems supporting these organizations while taking into account basic psychological knowledge. Educational controlling and performance management on a corporate level as well as the (technical) environment to build a learning organization are the main topics of this section. The first chapter examines ways of knowledge and competence management. Humans are not able to cope with the exponential growth of information and the increasing speed of information and business processes fostered by information and communication technologies. Technical support not only for information storage and retrieval but also for information selection, process planning, and decision support is needed. Most of the ICT investments, however, do not foster innovation or productivity. Recent studies show that ICT-based training is the main instrument of knowledge management. Organizational learning is a contradiction in terms (or oxymoron), because 'organization' can be understood as a rule describing a desired function or as a social society. From a psychological point of view, however, learning cannot be organized and knowledge cannot be managed. A 'learning organization', therefore, can only try to reduce roadblocks to learning and to improve informal ways of communication using the most efficient organizational structures and processes as well as the most efficient ICT supporting them. E-Learning is able to deliver more valuable training for less money only if it is part of an integrated knowledge and skills management system. Two case studies of knowledge management and meta-data management systems are discussed. The second chapter of this section deals with educational controlling. On-line media and self-directed learning environments are among the most effective training solutions in terms of cost, time and logistics. In the last few years, the percentage of employees participating in training courses increased. At the same time, there has been a decline of training budgets. Therefore, careful educational controlling must guard direct and indirect costs listed by amount, mode and source. Additionally, the success of on-line courses must be measured in terms of training and functional sectors. Most educational departments generally use
Learning Organization
35
controlling processes which consist of consecutive phases. Modern economic controlling, however, focuses on the permanent improvement of process outcomes. Recent examples of Web-based training courses and on-line tools illustrate how continuous feed forward and feed backward loops may be implemented in on-line learning. Controlling the transition from learning to working and checking the sustainability of learning outcomes improve quality and success of training and make it possible to calculate the success of training applications. Thus, modern educational controlling approaches should be introduced: they comprise exact calculation of financial investments and gains, optimal planning of organizational processes and goal-oriented definitions of strategic and operational learning objectives. Web-based training will then allow the educational and IT staff to take over a strategic role to establish an innovative learning and working culture. Investing in IT and other tangible assets does not guarantee increase of productivity and competitive advantages. Therefore, intangible assets are becoming a more important factor of the measurement of the goodwill of a company. International Accounting Standards (IAS) are incorporating recommendations for the measurement of goodwill and intangible assets. IAS will be introduced to the European Union in 2005. Measurement and presentation of intangible assets has been developed during the last twenty years. Recently, it has been complemented by dynamic process models and rating of intellectual capital according to international rating standards. Investing in human capital has also been seen as a reduction of risk and an increase of stability of a company. Even considering considerable advances of the discussion about intangible assets, it should be pointed out that competence management will not be established before educational controlling is as normal as financial controlling and investing in education is as important as investments in marketing.
This page is intentionally left blank
Chapter 1
Implementing Organizational Learning: Learning in the Learning Organization
Knowledge, Technology, Strategy For several years, high investments in information and communication technology (ICT) were considered to guarantee increased productivity. For the last ten years, however, surveys have not been able to show a significant increase in productivity by further computerization, (Bubic, Quenter & Rupplet, 2000). The authors point out, that the companies invested 'defensively': they reacted to technically induced changes by completely or partly automating processes, but neither redefined valuechains nor positioned themselves in new market segments. In fact, 70% of all ICT-investments are used to safeguard service and delivery, 20% to increase performance, and only 10%) to improve the competitiveness of the enterprise. Thus, modern ICT is more like a narrowing dead-end than like a useful support for a dynamically growing and changing business. Studies over the past decade have pointed to several factors that contributed to the productivity paradox. First, some of the benefits of ICT were not picked up in the productivity statistics. This is mainly a problem in the service sector, where most ICT investment occurs. ICT may have aggravated the problems of measuring productivity, as it allows greater customization, differentiation and innovation in the services provided, most of which is difficult to capture in statistical surveys. A second reason is that the benefits of ICT use might have taken a considerable time to emerge: The diffusion of new technologies is often slow and firms can take a long time to adjust to them A third 37
38
Learning Support Systems
reason is that many early studies that attempted to capture the impact of ICT at the firm level were based on relatively small samples of firms, drawn from private sources. Moreover, several studies have suggested that the impact of ICT on economic performance may differ between activities. Therefore, it is up to more recent studies based on large samples of data and covering several industries to find an impact of ICT than earlier studies (OECD, 2003). The banking sector is leading the development and adoption of electronically supported workplace training and is now moving towards distance education and integrated multimedia learning environments to accommodate the scale of vocational training and communication requirements needed. Today, many distance learning projects are realized by means of conventional media, such as printed mater and telephone hotlines. There is an emerging consensus that companies must invest in and provide access to a complete range of training programs for their employees from basic skills to high-end management and technical skills training. Electronic performance support systems (in the form of on-line media and self-directed learning environments) are among the most effective training solution in terms of cost, time and logistics. Therefore, some international companies have already reduced their face-to-face training courses by approximately 30% p.a. and all major companies are now introducing Web-based training (WBT) as a means for cost effective training. American universities which use Web-Based Training (WBT) or 'electronic learning' (in short: e-learning) learned that total costs of ownership (TOC) are higher for e-learning than for on-site classroom teaching. The main reasons are added technical and tutorial services as well as considerable up-front ICT investments (Bernath & Rubin, 1998). Will there be no gains in productivity because of investments in ICT and e-learning? Naturally, ICT itself does not create values. Values are created by its goal-oriented application in enterprises. The strategic areas of application of ICT can be sketched as in figure 7.
39
Learning Organization
Business Segments
jL
1. Redefinition of Products (e.g. electronic cash)
2. Change of Products (e.g.. integrated transaction management)
3.
Support of Products (e.g. improved credit handling)
Knowledge Measurement and Evaluation (Management Information System, Balanced Score Cards etc.)
Knowledge Selection and Distribution (Decision Support System) Provision and Storage of Information (Performance Support System)
--^**Business Control
Business Processes 1. Support of Processes (e.g. more efficient accounting)
2.
Change of Processes (e.g. helpdesk and user support)
w 3.
Redefinition of Services (e.g. integrated knowledge management)
Figure 7. Strategic Areas of Application of Information Technology according to (Bubik, Quenter & Ruppel 2000, p. 105).
Therefore, 'knowledge management' is a new definition of internal business and service processes with direct effects on rating and control mechanisms in enterprises and indirect influences on the new definition of business and market segments. In fact, knowledge management is conceived as an internal service which helps to collect, distribute, use, and evaluate explicit knowledge which can be represented in ICT structures (Probst, Raub & Romhardt, 1998; Kopp, 1999). Business segments with a very high demand of knowledge like banking or consulting industries are leading users of knowledge management; chemical, pharmaceutical and mechanical engineering enterprises are slower in taking up this new technology. One example is shown in the following figure 8. Of course it is no proof of more efficient or innovative business processes when a company -according to its self-assessment- claims to use more or less knowledge management. The so-called new economy marks the shift from the information to the knowledge economy, and asks for a different approach towards efficient business processes: The new economy was thought to temporarily overcome the business cycles
Learning Support Systems
40
leading to unequalled growth without substantial inflation. These expectation have led to a more dynamic performance of the stock market and of net products (Lochel, 2000; Hasebrook, Rudolph & Steffens, 2002). Others ^ ^ ^ H
Mechanical Engineering
~ ^ ^ ^ _
Insurance
....
H planned
1
1
Automobile
• no issue 1
|
"
I
1
Chemical+Pharma
|
IT " Consultancy
• initiated
1
1
1
1 •
Banking/Finance i 20
i 40
h60 80 (Specifications in %)
Figure 8. Status of Knowledge Management in Different Business Segments (Source: IT Research Report, March 2000).
The market for information- and knowledge-based -mostly digitizedgoods is only beginning to grow as yet, and training and education will be very important in this context. According to Merrill Lynch analysts, ICT-based training processes (or better: knowledge acquisition processes) are degrading what is presently meant to be the 'new economy' to the 'old ICT-based economy', because as before it is the industrial production of digital goods that is of prime importance, and not the control and support of knowledge-intensive business processes (cf. figure 9).
Learning Organization
41
Web-Based Learning: E-Learning In order to employ multimedia in an efficient way, it is at least as important to prepare the learning environment adequately as it is to choose the proper media mix and the instruction methods. Figure 10 summarizes the research data and meta-analyses on learning efficiency and duration of learning (Hasebrook, 1995). 'old economy' Termination in 4 years (university etc.)
new economy Termination in 40 years (lifelong learning) Training seen as cost center Training seen as competitive advantage Mobility of learners Mobility of contents Distance learning in self-study Distributed, co-operative learning E-mail, letter and printed products Multimedia service centers with online media Generalized offers for everybody Tailor-made, individual offers Regional vendors (mostly international) trade names and well-known individuals Training just in case Training just in time Self-study, personal responsibility Learning partnerships, organizational learning Figure 9. Learning in the Knowledge Society (Source: Merrill Lynch. The Book of Knowledge, 1999).
On the left, the average study time (time) and learning rate (learning) of a textbook (text only) equals '100' and serves as a benchmark for other types of media. The values in this figure are only approximate values for the usage of the media in question, because above all the learning success depends on the carefully controlled and adequate use of media, and only little on the medium of learning. However, the frequently cited 'pyramid of retention rates', which simply summarizes media effects, does not portray reality correctly. This fact can easily be illustrated by an example: Nobody is able to learn foreign-language vocabulary effectively when the radio is playing the latest hits, TV is broadcasting economic news and vocabulary is read at the same time. But that is exactly what a mere addition of media effects is assuming.
42
Learning Support Systems
Figure 10. Learning efficiency and duration of learning; learning rate and study time for textbooks (text only) are set to 100 (cf. Hasebrook, 1995).
Learning Organization and Organizational Learning Learning is the acquisition of a relatively lasting change of behavior or the potential for it. Learning means exploring and investigating new things, being curious and leaving routines. Organizing, on the other hand, implies laying-down standards and routines, and a restriction of the behavioral spectrum (in order to improve efficiency). On this understanding, every combination of 'learning' and 'organization' form a so-called oxymoron, i.e. a contradiction in terms (Weick & Westley, 1996). Since the term organizational learning had been introduced by Argyris and Schon, several attempts have been made to reconcile the unfortunate couple. (Argyris & Schon, 1978):
Learning Organization
43
When comprehending 'organization' as a rule describing the desired functionality of a social system, an 'instrumental perspective' is adopted which contains prescriptive specifications regarding the characteristics of the organization ('How-to'-approach; Tsang, 1997). When viewing an organization as a social learning society comprising many individuals, an 'institutional perspective' is adopted based on descriptive approaches on how collective learning processes actually look (Tsang, 1997). Moreover, there was a discussion whether management knowledge is a combination of know-how and know-why (Kim, 1993) or a mixture of explicit and implicit or tacit knowledge (Nonaka, 1994). Kluge and Schilling propose to define 'organizational learning' (OL) as cooperative learning in a social system, and 'learning organization' (LO) as the formal framework which allows continuous life-long learning (Kluge & Schilling, 2000). Kluge and Schilling (2000) come up with the following conclusions: Organizational learning as a change and adaptation of the organization members' mental models takes place as direct, mostly informal interaction. Information technology is of little importance as far as information take-in and information evaluation is concerned, but is of high importance as a means of storage and transfer. There are organizational processes which improve information processing and transfer, e.g. learning orientation, trial and error-learning, team work, and standardization. It is essential to hold a balance of 'old' and 'new' personnel. Social relationships support the organization but at the same time innovations are prohibited. Social relations tend to stabilize within a few months only, but an organization can only learn by leaving familiar paths. Nonaka (1991) argues that much of the knowledge accumulated in the firm is made out of experience or organizational memory (Conklin, 1992), Therefore it cannot be communicated by workers under traditional, extensively formalized management procedures. And yet the sources of innovation multiply when organizations are able to establish bridges to transfer tacit into explicit knowledge. By doing so, worker experience is communicated and amplified to increase the formal body of knowledge in the company. Adler and Cole (1993) argue that in an
44
Learning Support Systems
economic system where innovation is critical, the organizational ability to increase its sources of all forms of knowledge becomes the foundation of the innovative firm. This organizational process requires the full cooperation of workers in the innovation process so that they do not keep their knowledge solely for their own benefit. It also requires a subtle balance of flexibility and stability of the labor force in order to ensure a constant flow of innovative and approved knowledge at the same time. This apparently simple mechanism lies at the root of a profound transformation of management-labor relationships (Senge, 1990; Oppermann, 1996). In a recent study the German company unicmind.com surveyed the top 350 German companies, 102 of them responded to the call for information about e-learning and knowledge management. The results show that 90% of the companies employ e-learning, such as computerbased trainings, but only 25% use WBT; e-learning is implemented in order to reduce costs (70% agreement) whereas knowledge management is implemented to exploit the knowledge resources optimally (72%); elearning and knowledge management projects mostly are initiated by human resource departments (= 62%, IT dept. = 23%) but the systems are mainly used by marketing and sales forces (= 69%, IT dept. = 52%). But, knowledge collection, selection, and distribution mainly depends on informal communication. Active model learning and coaching is only possible where the management span is small (fewer than 10 persons) and the organizational culture supports social learning. Strong personal support at the working-place has a multitude of positive consequences: There is a growth in satisfaction and self-efficiency as well as in the identification with the enterprise. Contrary to some theories regarding OL, informal relationships cannot be forced by means of organizational measures, but mostly develop as informal relations. Job rotation, overlapping project groups and systematic succession plans help to establish informal relationships (Blickle, 2000).
45
Learning Organization
Feedback
single loop
mental models double loop
individual actions
shared mental models
environmental response , e.g. computers single loop organizational actions double loop
Figure 11. Single Loop and Double Loop Learning Cycles (Kim, 1993). One of the most popular concepts of organizational learning is a learning cycle consisting of observation, assessment, design, and implementation (= OADI). Simple learning (single loop) can be supported technically, organizational learning demands a specific human resource management (double loop). At the beginning of this chapter, learning was defined as a relatively lasting change in behavior, triggered by experiences. It is true that by communication people can exchange information. Experiences, thoughts and feelings, however, cannot be directly transferred from person to person. From a psychological point of view, a definition of learning is only possible and sensible when there is a learning objective or learning task to distinguish the continuous change of mental processes from specific learning processes. Learning in this sense is meant to be every relatively steady change in visible or potential behavior deliberately acquired by an individual through experiences. Contrary to most concepts of OL, learning does not only include verbally describable expert or fact knowledge. Neither is it determined by observation and imitation. Learning is a general, creative analysis of the environment and the body in order to enhance one's own potentials and abilities. The knowledge obtained therewith is not a collection of objective facts: We cannot distinguish between the world we are experiencing and our knowledge about the world because our knowledge is our world (Johnson-Laird, 1993).
46
Learning Support Systems
Assuming that the terms OL and LO are oxymorons, the compound 'knowledge management' is as senseless as e.g. the term 'virtual reality' for real-existing three-dimensional computer graphs. To direct and to control 'knowledge' from aspects of managerial objectives would imply that the experiences of the world of a group of people can be directly manipulated. A simple separation of data, information and knowledge, or terms like 'explicit' and 'implicit' knowledge is too restrictive and suggests a clarity that does not exist: knowledge cannot be managed and learning cannot be organized. Accordingly, the empirical results in this area are little encouraging. However, the psychological view of learning as a way of systematically changing or enhancing one's knowledge - and with that one's experiences and views of the world - enables us to influence this behavior to the advantage of the organization (and its members). The Market of Knowledge Applications One of the most successful enterprises of the world used to be Cisco Systems which is producing most of the Internet's routing systems. In 1999, the CEO of Cisco, John Chambers, stated: 'Investing into technology was the first wave. Investing in services is the second wave. E-learning will be the third wave\ Traditional means of education are no longer adequate to meet the needs of life-long learning. Continuous education for large numbers of people appears to be unrealistic if conventional strategies are pursued. Even where available, the quality of education does not meet the high standards of international business. Furthermore, in many countries public and private funding for educational services are declining while costs rise faster than income levels and tax revenues. Therefore, electronic distance education will become a major source for ongoing education in the international knowledge-based economy (Romer, 1993). In 1998, the German Economical Institute (IDW) published the results of its study about on-going education in Germany between 1992 and 1995: About 75% of all employees participated in training courses, this percentage increased about 10% during the course of the study. At
Learning Organization
47
the same time, the educational budgets were reduced by nearly 10% resulting in 21 billion Euro - at an average of 1000 Euro per employee (Hasebrook, 1998). In a current study, the Gartner Group estimates that in 2002 60% of the e-commerce enterprises used skills management, and that 80%> of them experienced a considerable increase in productivity as a result. For the purposes of recruiting, employing and training of staff there will be a vast increase in the usage of the Internet. According to market analyses of the IDC, the turnover of skills management solutions will increase tenfold from about 565 million US-Dollar in 2000 to almost 5.6 million US-Dollar in 2006. Most technical systems concerning the human capital of a company focus on the administration of personnel and training, such as SAP Human Resource modules, Peoplesoft or SABA - just to mention a view of them. An US-American study lists about 300 systems for training administration and delivery (Hall, 2001). But finding matches of needs and demands certainly means more than matching keywords to indices or user profiles to database requests. The knowledge economy is not so much about information, it is about people. In a joint initiative of several partners we implement a Skills Management Information System ('SMIS', see figure 12), which enables the user to select learning modules according to her or his individual demands, prior knowledge, and time schedule. To make most of the knowledge and competence resources of ones staff and to motivate them for life-long learning is a key success factor of any organization. In former times, a good administration of personnel was sufficient, but nowadays effective personnel marketing and active personnel development are needed. Ilogos Research report that in 1998 only 29%) of the international top 500 enterprises chose e-cruitment for the purpose of recruitment, whereas the percentage rose to 60% in 1999 and to almost 80%> in 2000. There is a similar development as regards the usage of online-media for qualification: In 1997, Internet and Intranets together amounted to only about 2.4% of the total turnover of the education market. The IDC (International Data Corporation), however, estimates that there will be an average annual increase of 62% and 140%>
48
Learning Support Systems
respectively in 2002. Thus, in 2002 about 39% of the turnover of the education market have been realized on these platforms. Knowledge and Abilities "The only capital an organization has that is irreplaceable is the knowledge and skills of its people. How productive this capital is depends on how efficiently people share what they know with those who can put that knowledge to good use,5 noted the US-American entrepreneur Andrew Carnegie as far back as 1930. In 1999, the HRXML Consortium Pittp://www.hr-xml.org] was founded which - in derivation of the XML 1.0 Standard - is elaborating a meta-data standard for human-resource-related e-commerce.
PERSOMLICRER WO WSOtECK Folgende Tabello enlhalt die empfohlenen Thernen tur einen Beraier PBC im Private Banking Ihro petsonliche Bildungshistone ist hietin nicht beruckan-htigt Sie haben jetzt die Moglichkeit, die vorhegende Emschatzung su adjustieien
?hro runMiort JAssistenz flu", s'grtuMung
~3 c-evftjrjpchter Tesmip j'ouartaiiJQOnT^
MedieftKotnptsterc. Rechtl Rahrnenbedmgungen BWUWVL Sleuern Korvto urcd Dapot Ktcditgeschaft Hetttsn AHien
Jill
I
I"
1 f§§ in AnsJfeen 2 | | | in weiten Teilan
3 H deutfoh 4 H§ auSerst deutSdt unci Icons stent in der Ausfiihrung 5 :?:> in heraurragender ' Ausbildung und voibildfchar Ausfuhrung
Figure 12. Screen shot from the Skills Management Information System 'SMIS' adapted for Deutsche Bank/Private Banking (Germany) - selection boxes help to identify job roles, prior knowledge, and time schedules (left), a colored table indicates skills covered by recommended training courses (right).
Learning Organization
49
The group defines itself as follows: The HR-XML Consortium is an independent, non-profit association dedicated to the development and promotion of a standard suite of XML specifications to enable ecommerce and the automation of human resources-related data exchanges. The mission of the HR-XML Consortium is to spare employers and vendors the risk and expense of having to negotiate and agree upon data exchange mechanisms on an ad-hoc basis. By developing and publishing open data exchange standards based on Extensible Markup Language (XML), the Consortium can provide the means for any company to transact with other companies without having to establish, engineer, and implement many separate interchange mechanisms (cf. figure 13). Members of this consortium are big ICT companies like IBM, Cisco and Oracle, as well as software companies like SAP and Peoplesoft, and staff agencies like Randstad and Manpower, or financial services like Charles Schwab & Co. At the moment, about 100 companies belong to the consortium. The introduction of this standard and the early adaptation of the solutions in store put both vendors and users of software solutions and services in the areas of e-cruitment and e-learning in an exclusive position among the competitors and enable them to organize global markets. Internationally acknowledged and XML-based standards for the description of knowledge products are being developed for e-learning. The most important standard is Learning Object Metadata (LOM). Here is the self-description of the LOM Consortium [http://www.manta. ieee.org/pl484]: 'The mission of the consortium is to develop technical Standards, Recommended Practices, and Guides for software components, tools, technologies and design methods that facilitate the development, deployment, maintenance and interoperation of computer implementations of education and training components and systems. Many of the standards developed by LTSC will be advanced as international standards by ISO/IEC JTC1/SC36 - Information Technology for Learning, Education, and Training.'
Learning Support Systems
50
Name Description Competence ID (prim, key)
required (Y/N)
Taxonomy Entry Competence
Competence Evidence Application Domain Subordinated Competence
Primary Key T Description T Owner
Key • Description T Owner
Figure 13. Structural model for the description of competences according to Human Resource XML - white boxes indicate simple data types (e.g. strings), grey boxes indicate complex data types (e.g. records); source: 'Competencies (Measurable Characteristics) Recommendation, 2003 February 26'.
There are several providers for software and consulting in the field of e-learning and especially in the development and management of skills. Beside the modules for the administration of human resources (HR) by SAP and PeopleSoft, there are specialized providers, e.g. Meta4 from Spain or Infinium, and SkillsScape from the USA. Since the majority of enterprises confine themselves in their staff development to the handling of biographical and administrative data, most of the offered software solutions support the administration of human resources only. However, since the rise of the information society and its evolution into a knowledge society there is a need for a strategic reorientation from the administration to a more active development of skills in human resource management (HRM). In this context, skills-oriented management
Learning Organization
51
paradigms have lately come into existence which are now considered in corresponding management information systems (MIS). The vast majority of providers of software and service solutions in the field of HRM are still concentrating on administrative solutions, which are of little help in the proactive planning and usage of skills in enterprises.
fc++l
'mm
Matchergebnfe Kir *>uchkriterien Kutsbeschrefbgng i- fur
frortqeschrittene
Mfe;j
D o k u m e nte n m a n a oe m e nts
M
Programmierunq qrafischer Benutz&roberflachen
tmll [Flektronische GeschaftsabwicUunq irn v
S3
Dei dieses seminars ist die Verttefung Ihrtr j*j j s-ushjndenen v++-K'enntuiiJe. Dot- Seminar behandfiitfort-gfschnttenfl TltMnen und Kon-^pte von C + + , sowie
Throne- M#ir'd am Rechner sofort in die Praxis umgesetzt und dur'cb Ubungeri vercieutlicht, C»r Arrteil der praktischen Ub'jnqen betragt rnehr sis SO %. In ftinetn durcbgan^igeh Be-tsptel entwtckeln Sie erne funktionsfahige Anweodung. Aufcor/Vortragender Dr. Merke! Medien Sprechen en A*3B Notebook,
Figure 14. Combined Meta Data Management for Learning Object Metadata (LOM, right hand) and Human Resource Standard Exchange Protocol (HR SEP, left hand) based on HR XML binding (skills and learning management system of efiport Inc.).
The skills management system developed by the first author (cf. figure 14) is supplied with a thesaurus of some 1000 skill designations covering general skills (with a strong focus on banking and finance),
52
Learning Support Systems
personal skills (e.g. languages) and social skills (e.g. communication abilities). These skills can be put together to create profiles for job and function descriptions, project tasks and job advertisements. The profiles in turn can be combined to form model career paths and then linked to offers of training. Employees thus have the opportunity to assess their own existing skills and their target profile in the context of an internal online job mart. Comparing themselves with existing profiles will help them to assess how further training and project involvement might best contribute to developing their career. To this end the system not only makes all profiles and the links between them available: it also provides a gap analysis (target-actual comparison) relative to all these profiles, together with proposals as to suitable training measures that would move a person's profile toward their desired target. All data are stored as XML data sets and are based on the international standards LOM and HR XML. Psychological Factors of Success Bill Gates, founder, former CEO, and present head of research of Microsoft Inc., once claimed that modern technologies of information and communication form the 'nervous system' of a modern enterprise. At least in the near future, this will presumably be an apt remark. Certainly, personal communication is not replaced by them, but forms next to this 'nervous system' - the genuine circulation, the basic foundation of life. From 1997 until 1998, a research group at the University of Karlsruhe surveyed 47 enterprises and 58 persons regarding factors of success and failure which they knew from their daily experience (Gemiinden & Babin, 1999). They distinguished factors within project teams and factors within the organization. The essential results are summarized in figure 15. Gemiinden and Babin (1999) drew the following conclusion from their study: ® Conflicts are inevitable and are no substantial threat for the success of a project.
Learning Organization
53
® Conflicts should be approached and discussed in an open manner, and not be covered up, a solution should not be enforced, and conflicts should under no circumstances be avoided. * Teams with good group dynamics experience less conflicts and tend to develop a practicable approach to managing conflicts. ® American citizens are accustomed to an open, direct way of managing conflicts. In Germany, the existence of conflicts is commonly acknowledged and has considerably less effects on the success and satisfaction of team members. The theories X and Y, formulated by McGregor as early as 1960, and supplemented by Schein by a 'social theory' (Schein, 1988) are well known. Summarized briefly, the ideas state the following: Theory X: employees are basically not motivated to do their work and must therefore be encouraged to do their work effectively by external rewards and control. Theory Y: employees are seeking development and confirmation of their selves at their working-place and therefore will be striving to do their work as successful as possible. Social Theory: at their working-place, employees are looking for social contacts and want to be respected and appreciated by their colleagues; therefore they will try to do their work in a way, that will bring them as much confirmation as possible in their environment. In surveys, approximately 10% of the interviewees tend to support theory Y, 40 % support theory X and the remaining 40% back the social theory. This weak emphasis on the aspects of self-development and responsibility has consequences: Most enterprises have not succeed in winning their employees for their goals. Only 10% have gained agreement with the enterprise's goals by processes of social learning, 5% had chosen the enterprise because of its goals and more then 50% of the employees disagree totally or in part with the enterprise's goals (Berkel & Herzog, 1997).
54
Learning Support Systems
0 5 10 15 20 25 30 Figure 15. Central Interfering Factors in Projects of Enterprises as Experienced by Project Teams and the Entire Organization. (Gemiinden & Bobin, 1999).
Commitment to performance is not only based on 'the hope for success'. There is also 'the hope for failure', when the goal of a task is disapproved of. There is 'the fear of failure', if one does not feel fit for the task, and there is 'the fear of success', if the employee fears, that the activities will result in disagreeable consequences or demands. Landy combined these basic ideas in the goal-setting theory (Landy, 1985). It lays stress on four factors of success: 1. The determination to reach a joint goal depends especially on whether it agrees with the individual goals. The employee will try to obstruct or prevent a goal that he or she disapproves of personally. 2. Employees have to know in advance, which consequences they will have to face, if milestones are reached or not reached. Undefined consequences lead to a more observing attitude. It is highly important, that it is transparent, at what time which consequences will follow and how they are defined. 3. Employees work most efficiently, if there are precise milestones and they can adapt their behavior accordingly. Orders like 'Do your very best!' are not helpful or even annoying. 4. Difficult goals encourage efforts and competition to prove one's expertise. If goals are too simple, they undermine the intrinsic motivation.
Learning Organization
55
The simple technical means of communication most frequently produce the most important effects: For instance, Sproull and Kiesler (1991) showed in several studies, that simple e-mail conferences had several advantages over personal discussions. The communication was more balanced and focused, and more often led to unanimous results. However, co-ordination via e-mail took more time than direct communication. Weisband and Atwater (1999) observed that the selfappreciation of participants in electronic communication is more uncertain and more restricted than that in direct communication, because social feedback and sympathy values are lacking. The success of electronic communication is highly influenced by the strategic employment of seniors and specialists. Ogata and Yano (1998), e.g., found out that electronic discussions among colleagues tend to be scarcely attended and to 'die out' quickly. If experts and seniors participate, both activity and drop-off rates are considerably higher. Electronic communication is increasingly used to support learning processes. Boiling and Robinson (1999) examined co-operative learning by comparing three learning groups: individual learning with print products, group learning with print products, and group learning with multimedia. Considered all together, co-operative learning turned out to be the best method; multimedia was best for students with a high amount of previous knowledge. Information and communication technologies have led to the fact, that local and temporal distances are of less importance. It is more important than ever to practice the principles of management, that are usually restricted to ceremonial proclamations. A management structure, that is based on an open exchange of information also known as 'leader member exchange' (LMX), leads to results, that are significantly better than those of a management, which is predominantly focused on organizational and cultural changes. This is true for on-site teams relying on face-to-face communication and on-line teams which only communicate by computers (Howell & Hall-Meranda, 1999). If the management participates actively in group processes and defines ambitious (but not unattainable) goals, this enhances both the motivation and the team spirit of the group (Tesluk & Mathieu, 1999). Informal communication, based on personal confidence, is the force that enlarges
56
Learning Support Systems
the knowledge of an organization, as studies regarding LO and OL have shown. In the first instance, technology only solves 'technical' problems of data storage and distribution. In an economy working interdependently and relying on direct knowledge transfer these logistical tasks are not to be underestimated. The application of these technologies must not be used as an excuse for not consequently realizing the principles of a democratic management and a transparent culture of communication. Competences on the Balance Sheet The measurement of competences and their impact on the intangible assets and therefore on the goodwill of a company are still in their infancy. Mostly, measurements based on tangible assets are pursued. Table 6 displays the shift of the paradigms from tangible to intangible assets - and hence from financial to intellectual capital as a major source of the goodwill of a company.
Table 6. Past and future of controlling regarding 'intangible assets' Valuation (financial capital) Tangible assets Valuation at one point in time Feed back (judging the past) Cost management Pure Financial Valuation Periodic controlling Rules and Routines
Valuation (intellectual capital) Intangible Assets Valuation of processes Feed back & feed forward (predicting the future) Value Based Management Blended Valuation Strategies Permanent controlling Management of Change
There is an increasing number of recommendations of the International Accounting Board (IAB) how to reflect intangible assets in the goodwill of a company properly. This is not an easy but an important task given the high number of mergers and acquisitions (M&A) in several industries which are based on an adequate estimation of the companies' goodwill. The first professional measurements of intellectual
Learning Organization
57
abilities and job related abilities and aptitudes, however, were psychological testing procedures. Psychological testing procedures including computer-supported diagnosis are used to conduct aptitude tests in personnel selection (e.g. Ghiselli, 1973; Sweetland & Keyer, 1984; Funke, 1993), adaptive testing optimizing economy and performance of personality, aptitude, and ability tests (e.g. Cronbach & Gleser, 1965; Park & Tennyson, 1983; Weiss & Vale, 1987; Bennett, 1993), and decision analysis applied to management diagnostics (Nagel, 1993; Sonnenberg, 1993). Additionally, there are numerous tests which check for individual interests (e.g. Todt, 1967; Irle & Allehoff, 1984), ability (ITB, 1988a+b, Deidesheimer Kreis, 1993), and aptitude (Fock & Engelbrecht, 1986). The different and differentiated ways of saying intangible assets and intellectual capital evolved during the last 20 years (Sullivan, 2000): A starting point was the work of the Japanese Hiroyuki Itarni about 'Invisible Assets' at the beginning of the 80s. Later on, a number of economics developed different approaches about the valuation of technical commercialization. It was Brian Hall who founded the first company in the U.S.A. to perform such valuations. In 1986, David Teece introduced a method how to calculate the value of a technical innovation. Thomas Stewart in his well known article in the management magazine 'Fortune' coined the term 'intellectual capital', in 1992. One year later, Hubert St. Onge added the idea of valuating 'Customer Capital'. Baruch Lev, a colleague of David Teece at the Universiy of New York, launched systematic research about intangible assets in the middle of the 90s. At the end of the 90s, the Swedish researcher Karl-Erik Sveiby started to study 'human capital' in a comprehensive way, although some approaches of calculating human capital had been developed from a cost management perspective some years before (Flamholtz, 1981, 1985). In 1997, Leif Edvinsson developed the 'Skandia Navigator' which - for the first time - put intellectual capital aside to financial capital and put intellectual, human und customer capital in a logical order (cf. figure 16; Edvinson & Malone, 1997; Edvinson & Briining, 2000; Edvinson, 2002).
58
Learning Support Systems
Value
Intellectual Capital
Financial Capital
Structural Capital
Human Capital
Customer Capital
Organizational Capital
Innovation Capital
Process Capital
Figure 16. Structure of the Skandia Navigator (cf. Edvinsson & Malone, 1997; Edvinsson, 2002).
The internationally best known examples of the valuation of intellectual capital as a part of the annual business report are two Swedish companies: The pioneering insurance company, Skandia, and the educational consulting firm Celemi, which runs offices in five different countries. Next to the usual financial statements based on International Accounting Standards (IAS), Celemi displays three additional domains: (1) Growth and Renewal, (2) Efficiency, and (3) Stability. These domains are divided in subdomains: « Growth and Renewal ® Image Enhancing Clients « Research & Development / Revenue « Average Professional Experience
Learning Organization
59
«
Efficiency « Revenue / Client ® Revenue / Admin Staff « Value Added / Expert * Stability ® Repeat Orders * Rookie Ratio (proportion of staff with less than two years of professional experience) * People Satisfaction (according to Celemi's annual satisfaction survey) The business report of Celemi points out the transition processes from static valuation of financial capital to dynamic valuation of intellectual capital. The fictitious example displayed in table 7 is based on recommendations of the International Accounting Board which puts together the International Accounting Standards (IAS), which will be applied in the European Union starting in 2005. The example of a balance sheet comparing the years 2010 and 2011 shows that intangible assets are regarded as 'real assets' which can be borrowed or activated as a part of the active corporate capital. Shareholder value, then, will be partly based and driven by intellectual capital. Value Extraction and Value Creation Approaches for the valuation of technical commercialization and technical innovation are focused on 'Value Extraction' in order to separate a certain piece of the value chain as a part of the goodwill of a company. Comprehensive models, such as the Skandia Navigator, are dealing with 'Value Creation' in order to shape the entire value chain and enhance the overall value of the company (Sullivan, 2000; Edvinsson & Malone, 1997). Value creation is not seen a static 'snapshot' of a evaluated status but as a dynamical process (cf. Tuomi, 1999; 2003). A different and complementary approach is the valuation of Intellectual Capital according to international rating standards which are developed and applied by rating agencies, such as Standard & Poor's and Moody's.
Learning Support Systems
60
Therefore, Intellectual Capital, for instance structured according to the Skandia Navigator, can be rated based on international rating standards. An example is displayed in figure 17 (Edvinsson & Kivikas, 2003).
Business Model
Relationship
Human Resources
Organization
A
'' Intellectual Property (IP)
Management
Staff Personell
Network
Brand, Image
Customers (Stakeholder)
BB
B
A
BB
BBB
Figure 17. Example of a rating based on the Skandia Navigator and international rating standards (cf. Edvinsson & Kivikas, 2003).
From this point of view, investing in intellectual capital can be seen as a means of controlling risk and Weighted Average Cost of Capital (WACC) . This is the crucial part of Value-Based Management which seeks to optimize margins by reducing and controlling risks. The Economic Value Added (EVA) and their Market Value Added (MVA) are the major economic target measurement to control the entire company. Thus, investments in human resource development and intellectual capital are compared to and evaluated in the same way as financial investments (Bender & Rohling, 2001).
61
Learning Organization
Table 7. Fictitious example of a balance sheet based on IAS comparing the years 2010 and 2011; all figures are given in millions of US-Dollar, all lines related to intangible assets are highlighted. Assets Current Assets Cash and equivalents Accounts receivable Inventories Total current assets Investments net Plant & property net Intangibles Brands Intellectual property Processes and systems Sustainability and environment Goodwill Total intangibles TOTAL ASSETS Liabilities and Stockholders' Equity Current Liabilities Notes and accounts payable Debt due within one year Accrued and other current liabilities Total current liabilities Long-Term Debt Other Non-current Liabilities Borrowings on intangibles Deferred income tax liabilities, non-current Pension and postretirement benefits, non-current Total other non-current liabilities Stockholders' Equity Common stock Retained earnings Intangible capital Net stockholders'equity TOTAL LIABILITIES AND STOCKHOLDERS' EQUITY
2011
2010
245 1624 1414 3283 3703 6621
220 2044 1342 3606 3489 6896
1829 6072 4513 2708 328 15450 29057 2011
1805 4311 4373 3087 615 14191 28155 2010
2941 130 675 3746 2951
3875 101 719 4695 2106
2324 1369 1220 4913
1916 1330 1015 4261
781 3539 13126 17446 29057
781 4036 12276 17093 28155
Learning Support Systems
62
Capital versus Talent Investments in ICT and other tangible assets are do no longer guarantee an increase of productivity and the value of a company. Therefore, intangible assets are moved into the main focus of economic valuation. Nicolas Carr (2003) wrote an article for the Harvard Business Review entitled 'IT doesn't matter' assuming that ICT is a commodity and a necessity and does not add any competitive advantages, any more. But not only ICT investments are considered to be mere commodities: In another article of the Harvard Business Review Martin and Moldovenau (2003) are debating the conflict of financial and intellectual capital. The authors think that the so called 'war for talent' indicates that financial resources are much more available than intellectual resources. If the value of a companies is much more related to intangible assets than to tangible assets, it concludes that in a conflict of 'capital versus talent' the winner will be talent, that is, intellectual capacity is the limiting factor of the companies' growth perspectives.
600 j500 — 400 — 300 — 200 — 100 » 0 — 1960
1970
1980
1990
2000
Figure 18. Ratio of CEOs' income as compared to their companies' net profits (income level of 1960 set to 100; sources: BusinessWeek's Executive Compensation Scoreboard and enterprise profits according to Standard & Poor's; cf. Martin & Moldoveanu, 2003).
Learning Organization
Martin and Moldovenau (2003) try to support their view comparing the income of chief executives and their companies' profits during the last 40 years. It turned out that the income of managers (playing the role of 'talent') is no longer related to companies actual profits and therefore the shareholders return financial investments (cf. figure 18).
63
by net the the on
This page is intentionally left blank
Chapter 2
Implementing Educational Controlling: Web-Based Training, Performance, and Controlling
More Training for Less Money Traditional means of education are no longer adequate to meet the needs of life-long learning. Continuous education for large numbers of people appears to be unrealistic if conventional strategies are pursued. Even where available, the quality of education does not meet the high standards of international business. Furthermore, in many countries public and private funding for educational services are declining while costs rise faster than income levels and tax revenues. Therefore, electronic distance education will become a major source for ongoing education in the international knowledge-based economy (Romer, 1993). The marketplace for many educational services within the corporate training sector will expand international or even global - with great increases in the quality of education available to the individual at lower real costs per capita than conventional education today. Computer, television, satellite, fiber optics and other technologies combine to create a vast interactive communication and information network (Hasebrook, 1998). The banking and insurance sector is leading the development and adoption of electronically supported workplace training and is now moving towards distance education and integrated multimedia learning environments to accommodate the scale of vocational training and 65
66
Learning Support Systems
communication requirements. Today, many distance learning projects are realized by means of conventional media, such as printed matter and telephone hotlines. There is an emerging consensus that banks must invest in and provide access to a complete range of training programs for their employees from basic skills to high-end management and technical skills training. Electronic performance support systems (in the form of on-line media and self-directed learning environments) are among the most effective training solution in terms of cost, time and logistics. Therefore, some German banks have already reduced their face-toface training courses by approximately 30% p.a. and all major banks are now introducing Web-based training (WBT) as a means for cost effective training. During the same period of time, German banks experienced a considerable increase of training costs: Major German banks and bank associations spent between 88% and 136% more money on training in 1996 than in 1989 (Ausbilderhandbuch, 1998). A closer look at these data reveals there was a dramatic increase of training costs between 1989 and 1993. Since 1994, there has been a decline of training budgets at the same rate as in other business sectors. Banks are spending 6% of their personnel budgets for training but 15% to 30% of their administrative and operational budgets for information and communication technology (ICT). For instance, the largest of German banks, Deutsche Bank, spent 205 million Euro on training and 1.3 billion Euro on ICT. This translates into approximately 2,500 Euro for training and 4,000 Euro for ICT per employee (Moormann, 1999). Banks and insurance companies have always been among the pioneer users of new technologies. In Germany, 30% of all insurance companies and 15% of all banks are actively using multimedia. More than 50% are planning to do so in the future. More and more, tele co-operation is applied in order to support training-on-the-job, counseling, and customer services (Hasebrook & Woerlein, 1996). The Cost of Training The overall cost of training and on-going education are divided into direct and indirect costs. Direct costs consist of external costs, such as
Educational Controlling
67
rooms, travel, accommodations, material and trainers, as well as internal costs, such as salaries and administrative costs. Indirect costs are elicited by lost-production costs and consecutive costs, such as higher wages of highly trained employees. Some simple guidelines may help to calculate the costs of web-based training as compared to face-to-face courses: Computer-based training (CBT) and Web-based training takes 50% to 70% of the study time of a traditional learning course (Hasebrook, 1995). Highly efficient WBTs may even reduce the study time to 30% to 50% of the original study time. One hour of learning requires 40 to 50 pages in the WBT, approximately one month of development and a budget of 30,000 Euro. Given these rough cost estimations and the average costs for travel, accommodations and trainer income, it is possible to compare the costs of face-to-face courses and WBT: If 100 employees participate in a three day face-to-face training, the costs will be between 420 and 840 Euro per participant. CBT and WBT may be provided for a rate between 360 and 660 Euro per participant. It must be considered that sophisticated WBT, with the support of on-line tutors and moderated chats or bulletin boards, may result in even higher costs than simple face-to-face training (Bernath & Rubin, 1998). On one hand, there is clearly a trade-off between cost reduction and support quality. On the other hand, cost reduction is not the only ends of WBT. In fact, one of the outstanding cost factors is quite frequently overlooked: time-to-market, which is the time from product development to market introduction, or time-to-operation (TTO) which is the time spent to introduce a fully operational corporate ICT solution, such as an ERP system. A complete cost analysis takes into account all direct and indirect costs listed by amount and mode. A crucial point is to identify the source of the costs: e.g. the department which demanded the development of a certain training course. Some WBT systems in German banks use individual 'training accounts' which gives each employee the opportunity (and the obligation) to monitor the cost of his personal training course. A comprehensive cost analysis helps to identify and calculate alternative forms of training courses, forms the basis of prognoses and planning processes and is necessary to perform comparational analyses (Kredelbach, 1998).
68
Learning Support Systems
Learning Efficacy and Cost Efficacy Many problems arise when a systematic controlling of on-line learning processes shall be employed. A major obstacle is an inadequate learning and communication culture which does not support the free flow of information but facilitates competition between employees who in turn are not ready to share their knowledge with others. 'Learning in electronic media' is a WBT which helps to overcome deficits in computer literacy and learning culture (Hasebrook, 1999). A major German bank uses the course modules to introduce a platform for WBT administration and delivery (cf. figure 18). Bank Academy, therefore, designed the course as a generic Web-based training system that can be used as a stand-alone training course or as an integrated help system supporting other Web applications. Thus, it can be used as a Performance Support System (PSS) to enhance utilizing electronic media in learning and in working environments (McGraw, 1994). The PPS provides four modules, a comprehensive glossary, a keyword and a full-text index as well as a graphical overview with brief summaries of all modules. The table of contents comprises the following topics: Learning with multimedia, information from the Internet, e-mail and computer conferences, and learning strategies for WBT. Each module is accompanied by a brief self-test enabling the user to check his or her knowledge on the subject matter. In order to motivate users to apply learning strategies, we integrated fifty so called brain tests: Each brain test consists of short psychological experiments which can be easily conducted within a few seconds and illustrate important features of human perception and human memory. Initial experiences ascertained that it is both amazing and highly motivating for students to test their own perceptions and learn about human cognition. Another problem is that the departments which are responsible for Human Resource Development (HRD) mostly do not properly define strategic and operational goals which can guide the training and controlling processes. An obvious short coming in many training courses is a lack of knowledge about the target audience and its needs. Therefore, it is suggested to implement an user-centered design process which involves end users and does not solely rely on expert information
69
Educational Controlling
(Glowalla & Hasebrook, 1995). It is helpM to control for homogenous learning teams with the same status of prior knowledge and motivation. Integrating users in the design process may enhance the motivation and ascertains to assume the correct level of prior knowledge. Furthermore, it is important to underline the importance of new ICT in education as an investment in the future of the corporation - and not 6just another IT rollout5. Thus, educational controlling must consider not only the costs of the training courses but also the outcomes. The crucial success factor of training is the transfer of information learned from the WBT into practice. However, this will require removing learning processes from its isolated position in training centers and dedicated training PCs; learning must be integrated into the workplace environment. It is clear that WBT, as an integral part of the coiporate Intranet, is best suited for providing a combination of information, learning, and communication services which support working and learning. i^ji^^MMti^l^te^ii,
v,
,
.... „.,...„.„..„„..,.„...„„,..„.„..
,
Gtossaf I SHcbwBtiar I
Lerrierr-wilWli®ii Wedien Uberslcht
Hivtnfbmiittioiten I ,^_ ¥f£?
i|r-fti|lr;:i||!jj
^Hi^i^liP
l
^MM^^mmMi
Figure 19. The Web-based training 'Learning in electronic media' helps bank employees to grasp the basic navigational skills and learning techniques to successfully cope with Web-based learning environments (copyright by efiport Inc., developed in charge of Deutsche Bank AG).
70
Learning Support Systems
• The success of on-going education Is measured In terms of taking and functional sectors. Training success consists of user acceptance (e.g. a survey with Individual acceptance ratings) and learning performance (e.g. knowledge tests before and after a training course). Important parts of functional success are the Individual performance at the work place and the performance gains of the department or corporation as a whole. They can only be measured, if educational goals are pre-defined. Assessment tools, such as surveys, questionnaires, and knowledge tests, support educational controlling optimally, if they are an integral part of the'WBT. Fina* Test/Test
1 f
,* Fe^a Es&ate U « i tt:a; p . « iipfr mt, of Ks p.schv,?
I p*
!"%.M ESVSM? L<".a»\s » s c-ffy' oifww: w8h tls*a Mere**
II J""
Motf*s-v^spisni fee- is «-iiVgsc!ft>-£;*><*ShSr-te (x^iv
B p ~ . Vs;i&We ofv! u-wj irt*n<^ r-st*>f- :$•& afiwii for Pes! rs'sts t w *
« Metiers Pon^<J
£t <m Sf;;p bsc->"
\os ivs.ii.y.m1i
Figure 20. A Web-based training developed by efiport in charge of Raytheon Training and General Motors Acceptance Company combines off-line and on-line elements, such as CBT-like naYigational aids, search facilities, bookmarks and a test database.
The first author has developed a test database to create, maintain and deliver exercises and tests within various environments from one source: Print-outs, stand-alone applications, and Java applets to be integrated in
Educational Controlling
71
WBTs (Klemme & Hasebrook, 1999). A broad range of interaction types have been implemented which enable the test database to provide surveys concerning acceptance ratings, to conduct knowledge tests and to store and process the input data. Figure 20 displays an example for a test evaluation form of a WBT which has been developed in charge of General Motors Acceptance Company (GMAC): After having passed an on-line exam, the user has the option of filling in his or her name on an evaluation form and to submit the learning results voluntarily. Of course, it is possible to store the data mandatory and automatically by the system itself. However, in many European countries strict rules of personal data security and individual performance controlling must be regarded which do not allow for automatic data collection and storage. The Process of Controlling Most HRD and educational departments generally use stepwise controlling processes which consist of the following consecutive phases: « Needs analysis - define strategic and operational goals and needs * Conception and planning - define learning objectives, decide to make or to buy, apply adequate didactics and organizational preparations ® delivery and conduction - control and support learning processes, collect evaluation data on-line and offline * Evaluation and transfer - analyze evaluation data and control transition to the work place Although this schedule appears somewhat comprehensive, it does not address the needs of an on-going learning process which permanently delivers information, learning, and communication services to learning stations and work places (Paul & Siewert, 1996). Modern economic controlling does not focus on strict schedules but on the permanent improvement of process outcomes. Therefore, modern controlling is a highly interconnected process which starts with the definitions of goals, problem analyses and prognoses, which rely on benchmarks and judgement of reasonable alternatives. A decision is made based on the data from these early stages and the realization is evaluated in order to
72
Learning Support Systems
come up with an detailed comparison of problem analyses and goal definitions. This comparison is used as feedback to adjust the following goal definitions. The feed forward loop from goal definition to evaluation is used as a fact finding process whereas the feed backward loop is used to store relevant data and lessons learned (cf. figure 21). Feed forward - collecting information
Evaluation
Goals i
Problems
»
y
i i
X^ Decision
Prognoses
Implementation
'' Alternatives
eans-end a lalyses
-
"
Comparison
* Feed backward - Storing information
Figure 21. A comprehensive model of a modern on-going controlling process (Paul & Siewert, 1996).
A WBT about currency management developed in charge of a major German bank illustrates how feed forward and feed backward loops may be implemented in on-line learning. The WBT is based on the Hyperwave information server (Maurer, 1998) and the learning platform TrainingSpace (formerly GENTLE; Dietinger, 1998, now eLS, see www.hyperwave.com). This software stores and maintains the user interface (e.g. buttons, frames), the structure (e.g. links, hierarchy of pages) and the actual content (e.g. HTML-pages, images) separately. Thus, all complete WBT pages are composed on demand and may contain individual information, such as notes and user defined links, without interfering with the contents of the WBT delivered to other users. We conducted a pilot study with 70 participants. Half of them were pooled in learning teams with 5 persons each and the other half studied
Educational Controlling
73
individually. All participants were allowed to take notes and write contributions to the discussion forum. All notes and contribution were typed according to their contents, that is the user decided whether she or he wants to type in a question, an answer, an agreement, a disagreement or a simple remark. All notes containing questions were sent as an email to an expert. All public notes were copied to the discussion forum. The WBT consisted of five modules consisting of approximately 120 pages each. Each user took an average of four notes per module and additionally wrote one or two messages to the forum. The notes did not only support the learning process by motivating the users to discuss the subject matter of the WBT. They also provided an tremendously useful source of information for the adjustment and improvement of the system (cf. figure 22; Kunz, Drewniak & Schott, 1994). The results of this study are reported in full detail in chapter 4 of this book. WBT clearly has the capability to support feed forward and feed backward processes: Notes can be linked to particular pieces of information in the WBT and copied to a forum. Colors and icons can indicate the content and access rights of notes and forum messages; messaging systems support the flow of information between individual learners, learning teams and subject matter experts; frequently updated information and background libraries allow the WBT to be up-to-date even in rapidly changing environments. In order to account for the various roles and uses of WBT in a modern business environment, two areas of controlling data should be regarded (Meier, 1995): A. Controlling of learning outcome * Acceptance data (e.g. surveys, interviews, group discussions); 48% of all German banks collect acceptance data. * Performance data (e.g. study time, learning outcome, stability and sustainability of the learning outcome); about 25% of the German banks collect performance data systematically. * Problem solving and transfer (e.g. case studies, workshops, observation and judgements in the work place); 73% of the German banks use at least one of these methods. « Career development (e.g. individual discussions with employees, evaluation of the individual career steps); only 52% of all banks regularly evaluate the career development of their employees.
Learning Support Systems
74
Corporate success (e.g. Controlling of corporate and.department goals, sustainability of the success); probably nearly 100% of the banks are engaged in regular evaluation of their success measurements.
mmmummmmmmmmmmmmm 41
#
^rak.k
V»
% feu
iaiferi
A
£*
Atfong
Sirh>#i
gJ'B*fca*»aia»teHo X j j * * * * * '«5 M-a1w>h#M ^
! 1 L-sWwpn 2Ui
;
* -Siida
•Dmcktaa
frtoroa^^^J
SiJftirltaiC
M
HachiKhten jj
1ST -^j
fiffiC
^ ^
11.1.4 Warum wird der Euro eine starke Wahrung? Hire Kunden haben oft die Wahl zwischen verschi'edenen Wahrungen und Anlagestrategieti.
.c-itolgreichsn "Fificenkundenbeire .. 1 V W ; P;f--.,'arte.i un&fcra
zoom?; Ihre Kunden werden Ihnen daher Fragen stellen wis:
? Was fun *»» fur r*if- Kunden'' 3 Warum profiti'?;-"'! y;--!it-it- s^irtden vcw Euro'7 1.1.4 Warum w i r d der Euro eine starke Wahrung?
h
In welcher Wahrung sollen wir Gesehafte abschlieden?
fc
Sollen wir bei der D-Mark bleiben und lieber den Euro nehmen?
**
Sollen wir anders europaische Wahrungen bevorzugen, etwa das tflBritische Pfund und den Schweizer Franken?
Urn Ihren Kunden eine Oberzeugende Argumentationskette bieten zu konnen, miissen Sie sich Ihre Antworten Schrrtt fOr Schritt uberlegen. Die Verflffentlichungen unserer Bank helfen Ihnen d-abei. Wir haben einige Teste zusammengestellt, die Antworten auf die zentralen Fragen Infer Kunden beinhalten. Versuchen Sie nun mit Hilfe dieser Teste Ihre eigene, Oberzeugende Argumentation aufzubauen.
saa
..AA^A.-.-^.'..^.: Figure 22. A WBT produced in charge of a major German bank comprises five training modules on 'foreign currency management', search facilities including additional background libraries, note taking, discussion forums and an e-mail-based messaging system to exchange questions and answers between students and experts (marked with a 4 ?P icon in the text).
It has to be regarded, however, that most of the data mentioned above (e.g. acceptance and performance ratings) are independent and do not correlate with one another significantly (Kunz, Drewniak & Schott, 1994). Therefore, missing data cannot not be estimated referring to
Educational Controlling
75
another set of data. The other area of controlling is described in the following list: B. Controlling of transfer In the work place « Overall acceptance ratings of the learning environment * Observations (e.g. team discussion, work performance) « Judgements (e.g. regular evaluation meeting and individual judgements of employees) » Follow-up (e.g. workshops of regional teams which adapt generic training courses to their particular needs) Directly after training « Knowledge tests (e.g. self tests, examinations, comparison of preand post-tests) * Discussion between employee and employer ® 'Transfer partnership' or learning groups discussing the learning outcomes Controlling the transition from learning to working and checking the sustainability of learning outcomes are the crucial factors which improve quality and success of training - and makes it possible to calculate the success of training applications, doing away with guess work. Calculating Success Pichler (1996) conducted a national study on marketing and counseling training for retail bankers of a German bank. He found costs to be approximately 840 Euro per participant and day. The net income resulting from the training was about 2,600 Euro per participant in the first year. Comprehensive scoring models must be employed in educational controlling in order to come up with reasonable calculations of training costs and income. Kaplan and Norton (1997) have suggested such a model consisting of 'Balanced Score Cards', which take into account financial data, innovation rates and customer satisfaction. All cards are rated using scores derived from historical corporate data and benchmarking. Benchmarking and the use of control groups without training help to estimate the gains of training investments. However,
76
Learning Support Systems
there is no linear correlation of learning effort and learning outcome. A rough estimation states that 50% of the efforts produce about 80% of the results, that is: There is much more effort needed to get slightly better results, if the level of performance is already high (cf. figure 22; group 1 and 2). The graph depicted in figure 23 illustrates the non-linear correlation of learning effort and learning outcome. Having defined the learning objectives once, the graph identifies the investment needed to achieve these goals. Moreover, the probable learning outcome may be estimated if the actual investment has been pre-defined. The graph does not only provide a rough understanding of the correlation of the different factors. It may also be used to calculate numerical parameters of a statistical function that provides a sound estimation of the financial benefits from various levels of efforts and outcomes (e.g. Arnold, Cooper & Robertson, 1995; pp. 159). Critical parameters are: the percentage of persons of the target group that is able to fulfil the learning objectives (criterion), the percentage of persons of the target group who pass the test or exam (selection) and the correlation of criterion and selection procedure (validity). In order to calculate the following terms all individual scores are given as the z-score: x, + u z = ———
with
8 // = — V ^
and
n
Therefore, the z-score is obtained by subtracting the arithmetic mean from the individual values and division by the standard deviation. This procedure ascertains that all values are normally distributed with a mean of zero and a standard deviation of 1. With perfect selection, the average z-score of the selected learners would be the average work performance resulting from the training investment. The work performance, however,
Educational Controlling
11
will depend on the accuracy of the selection, that is the validity of the test procedures involved.
Group 1 „ (high level)
Outcome actual gains
Group 2 (low level)
Outcome (objective)
Efforts/ Investment Figure 23. The non-linear function of learning efforts (investments) and learning outcome - if a learning result has been defined, the necessary effort can be identified (and vice versa).
The validity of the knowledge test can be estimated by calculating the linear correlation of the test scores and a independently measured criterion, e.g. using methods of transfer controlling. The linear correlation of two values is given by: r=
P* SS„
with
n A conservative rule of assigning financial values to performance measurements is to assume that 40% of the salary is assigned to each standard deviation of work performance. This leads to the following estimation of the financial benefits from training investments:
Learning Support Systems
78
benefits
= — rtestcriterion • dyear • cyear with n N = number of candidates in the training (e.g. during one year) n = number of selected candidates by test scores (e.g. in one year) r = validity of test d = duration or number of years candidates will stay c = number of candidates tested per year Gains from Goal-Directed Planning Several conclusions can be drawn from this course of reasoning: The financial gain is enhanced if there are many adequate candidates, a low selection ratio and tests of high validity. All these parameters cannot be estimated and controlled in a short period of time. Educational controlling, therefore, demands an on-going collection of adequate data during a given period of time. Initial positive results cannot be expected within the first six months. However, collecting and analyzing educational data is worth the effort: Computer programmer aptitude tests were used to select computer programmers in the US; with a selection ratio of 50% gains of between 13 and 37 million US-Dollar can be expected in one year (Arnold, Cooper & Robertson, 1995). Using educational controlling procedures may also have such enormous positive effects. All relevant data needed to calculate the financial gains can only be collected and evaluated in an on-going controlling process. Additionally, data derived from such a controlling process support the evaluation of historical corporate data as well as planning the future. Thus, Web-based training should not be viewed as a simple extension of traditional computer-supported learning approaches. It should be used to introduce a modern controlling approach which comprises exact calculation of financial investments and gains, optimal planning of organizational processes and goal-oriented definitions of strategic and operational learning objectives (Hasebrook, 1999). Web-based training will then allow the educational and IT staff to take over a strategic role to establish an innovative learning and working culture within the corporation. Web-based training will be a successful complement to
Educational Controlling
79
traditional ways of delivering training if it proves to be a solid basis for goal-oriented planning and cost-effective training solutions.
This page is intentionally left blank
Part 2
Performance Support: Introduction
The empirical evidence does not support the mad dash to use multimedia content in education. A recent meta-analysis examines 248 research studies on computer-aided learning. 150 studies failed to show any significant benefits. The other studies showed only a slight advantage over textbooks or lectures: error rates for simple retention tests were reduced about 10%, problem solving was hardly enhanced, and study time was reduced about 30%. Though multimedia seems to save time and reduce simple errors, it has not been found very effective as a problemsolving tool. Reviewing several meta-analyses, it seems clear that the use of multimedia is not the main factor influencing learning: Measured learning gains are most likely due to instructional methods. Fortunately, there are some studies showing that multimedia can facilitate the learning process. The Software Publishers Association (1995) reviewed the impact of instructional technologies in 133 school studies from 1990 to 1994. They found better test results, increased selfreliance, and closer interaction between students and teachers. Similarly, Boettcher (1993) collected 101 success stories. Many other studies show: Multimedia can enhance communication and motivation. This does not necessarily lead to improved learning but it can facilitate learning beyond the classroom. Quality of E-Learning Environments It is worthwhile that we start this section with three points that we strongly believe in: 81
82
Learning Support Systems
First, e-learning must not be seen as tool to teach larger groups in a stratified way, but rather to provide individualized teaching at the right level of knowledge and cognitive skill of the individual student involved. This does not mean that it can't be cost effective and that it cannot be used for large groups, but it does mean that the material must adapt itself to the users: it must not provide a 'one size fits all'- kind of solution. Second, although most members of the e-learning community have slowly started to agree that e-learning material must provide more than a slightly interactive electronic book, it is less widely understood that digital libraries (be it self-produced or purchased from publisher) can and should be used to provide important background information. Putting it differently, courseware consisting of stand-alone units that do not make use of existing digital libraries, be it local or on the WWW, provide a focus that is much to narrow. Third, it must be recognized that no university or company can or should compete with Hollywood or TV Studios when preparing e-learning material. Good and pleasing content is necessary, but impressive multimedia material is not the answer (indeed may be distracting in some cases), but the answer is the use of a suitable e-learning environment. In this section we just want to address a few features of e-learning environments from a user perspective that are often not considered seriously enough. For a full set of functions required in a good e-learning environment see e.g. the e-learning Suite of Hyperwave [www.hyperwave.com ] (Maurer, 1996) or the efiport learning support system [www.efiport.de] (Hasebrook & Otte, 2002). First, any good e-learning system should provide pre-tests to determine the knowledge of the student involved, and if possible also the most suitable cognitive style. Note that as far as cognitive style is concerned there is widespread agreement that the performance of learners may well depend on how material is presented, yet there are few attempts to systematically exploit this in the knowledge transfer process. Second, a good e-learning system must support all paradigms for learning that we often hear about: no single paradigm is ideal for all applications. For fact learning, behaviorism (drill and practice) remains as valid as ever and should not be looked down upon, the 'cognitivistic' approach is the one often best suited, unless constructivism is a viable
Performance Support
83
alternative. But there are other approaches like implicit learning (Holzamer, Pichler, Ahner & Maurer, 2001), situated learning (Maurer & Pivec, 2001), that also must not be ignored. Third, users should be able to work with the material to an extent that goes beyond all learning theories: we will return to this in the next section. Fourth, when e-learning is used in a networked environment (and notnetworked attempts have been successful only in very isolated cases) the network must be exploited for communication and collaboration to the fullest, including discussion forums, chats, shared work spaces and the like. Note in particular that although discussion forums are in widespread use by now, such forums rarely are powerful enough to handle largescale discussions, including the re-structuring of discussions, the merging of discussions or the extraction of parts of a discussion as a special resource used elsewhere. Systems with powerful discussion forums are e.g. eLS and WBT-Master, but some basic rules have already been formulated much earlier (Maurer, Rozsenich & Sapper, 1999). Fifth, any kind of courseware or teachware should not be seen in isolation, but always in conjunction with sufficiently large digital background libraries. Such libraries can consist of material generated in the course of other activities, or can be purchased in the form of libraries available on the WWW or material on CDs and DVDs. As is pointed out by Maurer and Tochtermann (2002) using techniques from knowledge management to be discussed in a later section, automatically generated links even in a way that they can be visualized as 'knowledge maps' such as the ones in the multimedia encyclopaedia 'Brockhaus Multimedial' should and can be used. For some aspects of digital libraries and their use consult Marchionini & Maurer, 1995; Lennon & Maurer, 1995; Maurer, 2001). Active Documents and Active Communication It is generally accepted that passively observing material on a computer screen, no matter how many pictures, diagrams, animations, movies, audio material etc. are used is not enough to create a productive learning
84
Learning Support Systems
situation. Much more interaction is essential. This is, after all, the basis of many learning theories. Only by letting students solve problems, collaborate with each other etc. will they be sufficiently involved in a process conducive to serious learning. However, what is often overlooked is that students should also be allowed to 'massage' material that is presented to them, by adding notes, adding links to the WWW or a background library, by attaching files, by highlighting, supplementing or erasing parts of what is shown to them etc. In each case, such changes will either be just for one student (producing a particular student's view) or for a group (producing a special view of the group collaborating). It is our experience that in this way material offered to students will expand, different persons or groups of persons ending up with often surprisingly different versions of the original teaching material (which of course is never modified as such: the modifications are only superimposed and only visible to those who have authority to see them: nobody except for the original author can change the underlying substance.) Communication is not just important to break the isolation of students in an e-learning environment but also for a much more basic reason: whatever one person says or writes, the receiver of the information will always interpret the information in the receiver's personal context, created through upbringing, culture, language, etc. This does often lead to deep misunderstandings. Our favorite example is the story of a fish who, when hearing of a flying animal does not think of a bird as we know it, but of course of a fish with wings; or when hearing of a 'four legged animal with an udder with milk' is more likely to imagine a frog with an udder with milk than a cow, simply because frogs are probably the only four legged creatures fish know. It is often claimed that a picture says more than a thousand words. And this may well be true, but although at times miss-conceptions might be resolved using pictures, this is by no means always the case: a fish who happens to see a person drinking a glass of water will be quite dumbfounded, for such action does not seem to make sense to the fish; a nomadic person in the dessert (who has never seen anything but the dessert) will not understand the picture of fog rising over a lake, nor will the traditional Indian in the Amazon jungle be able to make much sense
Performance Support
85
of trees covered with lots of snow when shown a picture of a winter scene from Austria. The only way to make sure that information is properly understood is not by reading, hearing, seeing, but by being able to check if things have been understood and by asking questions: this is why an e-learning system that ignores the importance of communication will not work. There is one more subtle aspect about communication: We believe that communication should not be restricted to communication between persons but should be extended to cover communication between students and documents. To formulate it in an exaggerated way: We would like to see systems where a student who sees something on the screen can type in any question whatsoever and the document gives the answer. Although this sounds like absolutely impossible, the situation described can be approximated quite well if the information on the screen is viewed by many thousands of persons before it changes. In this case the concept of 'active documents' (Heinrich & Maurer, 2000; Heinrich, Johnson, Luo & Maurer, 2001) can be applied: when the first few hundred users ask questions, the answers are given by experts, but both questions and answers are stored in a database. Later questions that can be recognized to be semantically identical with earlier ones by the system can then be answered by the system, i.e. the documents. In large applications we have found that over 99.5% of all questions can indeed be answered without human intervention. The 'Knowledge Information Center' of Hyperwave is one such software module. Indeed, many of the features discussed so far, and including the knowledge management aspects to be discussed in the next section have been successfully handled by extending Hyperwave. For background information consult (Maurer, 1996; 1998). Knowledge Management Teaching and learning are clearly involving.knowledge transfer, and hence e-learning is clearly a small subset of the fairly new but increasingly important area of Knowledge Management (KM). Thus, elearning must not be dealt with in isolation but with at least a minimum
86
Learning Support Systems
of information about KM and its tools. Rather than defining KM, it is easier to explain KM by quoting the famous statement: 'If our employees only knew what our employees know we would be a much better organization.' Thus, the original challenge of KM is to extract knowledge from persons (without burdening them with extra work), storing the information in a computer system, and making the knowledge available to users when they need it (even if they have not asked for it). It is exactly the two parenthesized remarks in the last sentence that distinguish KM systems form ordinary information systems or databases: in ordinary information systems, information has to be input and requested explicitly. Surprising as this may sound, the automatic extraction of knowledge without imposing extra work on the persons whose knowledge is desired, and the provision of relevant information at the right moment is indeed possible to an increasingly high percentage. For details we refer to Maurer and Tochtermann (2002) and the references therein, e.g. (Ives, Torrey & Gordon, 1998; Meersmann, Tari & Streus, 1999). However we would like to at least mention three of the many tools that are currently used in KM that directly apply to elearning: the first two, 'knowledge maps' and 'active documents' we have already briefly mentioned above. And the active document concept contains in it the seed for something much larger: after all, to discover if two questions are semantically the same, one basically needs mechanisms to discover if two documents are similar. It is this similarity- recognition that proves to be invaluable in KM and in e-learning. Let us look at a number of simple examples to try to prove our point, the first one from the commercial world, the others from e-learning. In a large, world-wide distributed manufacturing company a new project is to be undertaken. The engineers draw up, according to quality assurance procedures, a detailed description of the product in a fairly standardized manner. The Hyperwave system that we have installed in one such instance translates the specifications from whatever language into (rather mediocre quality) English. This English document is compared to the English descriptions of all projects, planned or in progress in all locations of the company at issue. If it detects a strong similarity, it alerts both groups involved to avoid potential duplication.
Performance Support
87
This 'simple' procedure has saved the company at issue more than ten million US-Dollar in a single year. Suppose you are writing a paper. As soon as you have finished your extended abstract you switch your system into supervisory mode. It now starts to check what you have written against all material available in back- ground libraries world- wide. If it finds some 'suspicious' similarity, it will alert you: you might be frustrated since your 'novel idea' turns out to be not quite as novel as you thought (but better you find out now than later) but it may also help you to show you results you can put to good use. There are many applications that directly apply to e-learning. Similarity recognition (SR) may alert a student that another one who is doing similar stuff. SR may help students to find fellow students who are experts in topics they are currently interested in. SR may help a teacher to find out an incident of plagiarism. SR can help to short-cut discussions in forum by pointing out that the same topic has already been treated exhaustively some time back. Without going deeper it should be clear by now that tools developed for KM are very much applicable to e-learning, and must not be ignored by the e-learning community as has largely been the case with some exceptions like Hyperwave and 'http://coronet.iicm.edu' mentioned earlier. Groupware is hardware and software that enables groups of people to work together. For example, groupware enables a team to access a database containing everything related to the work process, including past discussions, memos, and meeting reports. Most groupware programs use a local area network; some run over the Internet. The main functions of groupware are: « knowledge sharing; * group calendar keeping and scheduling (for project due dates, meetings, and conference calls); « real-time meetings (e.g. chat and video-conferencing); ® bulletin boards (asynchronous discussions over long periods of time that can be stored and retrieved); and * workflow management. Applications for groupware are not just limited to teams in the same company. Many projects extend beyond the narrow confines of the
88
Learning Support Systems
company. For example, architectural firms typically work with a large number of subcontractors when designing or renovating buildings. But technology is not everything. It also takes social skills to make virtual teams successful. While conventional work teams have many ways to convey information, virtual teams do not. In a virtual team, the leader must tell members early on what they are supposed to do, who makes the decisions, who represents the team inside and outside of the firm, how information is processed, and how communication should take place. The following two chapters try to examine the key success factors of learning support by using multiple media and collaborative learning systems. Effective Web-based training (WBT) has a need for adaptation and contextual information, but most WBT modules provide not more than some simple help pages or hypertext facilities with keyword indices. Help pages mostly provide information how to access functions but not how to apply them - and why. Therefore, a generic Web-based Performance Support System (PSS) was designed which can be used as a stand-alone training course about 'Learning in electronic media' or as an integrated help system supporting other WBTs. The PPS provides four modules, a comprehensive glossary, a keyword and a full-text index as well as a graphical overview with brief summaries of all modules. In order to motivate users to apply learning strategies about fifty so called brain tests were integrated: Each test consists of short psychological experiments which can be easily conducted within a few seconds and illustrate important features of human perception and human memory. First experiences ascertained that it is highly motivating for students to test their own perceptions and learn about human cognition. Major companies, especially banks, invest in interactive distance learning replacing face-to-face training. Research in this field has shown that the choice of media does not influence learning very much. Learning gains are mostly due to a shift in instruction. In this study a WBT about currency management of a major German bank was examined. The communicational features of the WBT comprise a discussion forum, note taking, and automatic messaging of questions and answers between experts and students. The experimental design compared a face-to-face seminar with WBT learning. The results show that WBT participants
Performance Support
89
learned as much as the seminar participants, but in about 70% of the seminar's study time. Young seminar participants performed better than older ones, while WBT learning did not produce an age effect. The results of the study demonstrate that the learners in the bank tend to choose traditional learning strategies and do not profit from co-operative and selective learning strategies, although they tend to appreciate audiovisual media. Experts were not very much engaged in the discussion process. Communicational features, however, were used quite frequently. The users who were experienced in using a CBT and showed high self esteem gained most from WBT learning.
This page is intentionally left blank
Chapter 3
Implementing Web-Based Training: Exploring Electronic Media and the Human Mind
From CD-ROM to Internet Learning effects of multimedia in education are disappointing, quite frequently. Van den Berg and Watt (1991) compared multimedia in competition to a classroom lecture, multimedia supplementing a lecture and multimedia replacing a lecture. They came to the conclusion that students would prefer to use multimedia as a supplement to lectures and books. Meta-analyses support statements like these (Kulik & Kulik, 1991; Hasebrook, 1995). Although, multimedia seems to save some time and reduce simple errors, it has not been found to be very effective as a problem solving tool (Mayer & Anderson, 1992). Many vendors and users prefer a stepwise migration from 'old' to 'new' technologies. For instance, Bank Academy has implemented a multimedia CBT in charge of the financial department of an international automobile manufacturer and dealer (cf. figure 24) which was implemented in five different languages and delivered on CD-ROM. One of the challenges of this project was to produce off-line and on-line training courses in a single production process. Therefore, we implemented the different CBT versions using the Hyperwave Information Server (Maurer, 1998) in order to maintain the multimedia elements. Hyperwave directly delivers Web-based training, because it includes a complete Web server, and allows to produce a 'snapshot' of the 91
Learning Support Systems
92
database which can be delivered on a CD-ROM. In the first release, the training course did not provide more than a traditional multimedia CBT. But since the year 2000, the course had been put on-line and, therefore, integrates Hyperwave's on-line features, such as note taking, discussion forums and bulletin boards.
Lesson 1 /Chapter 1 (11/59)ft&ain Purpose of Commercial Lending With which activities pur$a$ $f#3gcure the mmn pur$ma of Commercial Lending?.;
Please mark tfie correct aciwiti«% •:.. r r r r r
to grant sachrequestedcredit to provide and promote various types of funding., which meet dealers' nescis to determine the eligibility of each customer to maintain effective relationship between GMAC and dealers to convince dealers to on! . ... ..*• .,,•."„' ,.~";T~*~
'B$?:s^e#-
Figure 24. Advanced Web servers like Hyperwave allow to introduce Web-based training starting with off-line multimedia CBTs; this approach has been chosen in a project of Bank Academy in charge of an international automobile manufacturer.
From Help Pages to Performance Support Systems
Duchastel (1992, pg. 69) claims: 'Adaptation is essence of what is known as pedagogical knowledge1. Many researchers aim to make their multimedia systems more adaptive - a n d therefore more 'pedagogical1
Web-Based Training
93
(e.g. Cox & Bma, 1995). Expert systems and Intelligent Tutoring Systems (ITS) adapt to the learner's demands, abilities and knowledge especially in subjects which can be described in formal structures (Bastien, 1992). There is an increasing number of adaptive computer programs which are equipped with media like videos and photographs. As of today, a diverse spectrum of techniques, approaches and philosophies impede the progress in intelligent learning environments (Self, 1992). There are promising results, however, supporting positive effects of intelligent learning environments teaching mathematics and programming (e.g. McGraw, 1994). In general, effects of adaptation and system-controlled tutoring have been small or medium sized, yet (e.g. Schulmeister, 1996). Despite these insights about the need for adaptation and contextual information many Web-based training modules provide not more than some simple help pages or hypertext facilities with keyword indices. Help pages mostly provide information how to access functions but not how to apply them in different learning contexts - and why to apply them. Effective learning needs a good deal of verbal and visual literacy, whereas computer literacy seems not to be the most influential factor (cf. Mayer & Sims, 1994; Mayer & Anderson, 1992). Thus, most help systems do not support learning strategies to cope with linked multimedia elements, and they do not motivate to use electronic media as an serious learning tool. Effective help systems should support the user to overcome his or her weaknesses and take advantage of her or his strength. We therefore designed a generic Web-based training system that can be used as a stand-alone training course about 'Learning in electronic media' or as an integrated help system supporting other Web applications (cf. figure 24). Thus, it can be used as a Performance Support System (PSS) to enhance utilizing electronic media in an learning and in working environment (McGraw, 1994). The PPS provides four modules, a comprehensive glossary, a keyword and a fulltext index as well as a graphical overview with brief summaries of all modules. The table of contents comprises the following topics: Learning with multimedia: Advantages and disadvantages of computer-based training - Appropriateness of multiple media - Learning
94
Learning Support Systems
strategies for multimedia - Combining dynamic and static media - Self test 'Multimedia expert'. Information from the Internet: Basics about the Internet - Addresses in the Internet - Search engines and search strategies - Self test 'Internet expert'. Email and Computer Conferences: Basics about email - Writing emails - Mail and list server - Asynchronous and synchronous computer conferences - Video conferencing - Self test 'Email expert'. Learning strategies for CBT: Browsing hypertext and multimedia Using navigational tools - Using bookmarks and note taking - Graphical browsers, maps and overviews - Strategies for learning and re-learning Self test 'CBT expert'. Learning to Learn Many authors suggest that deeper understanding means that sequential verbal information is highly interconnected with analog pictorial information (e.g. Mayer & Anderson, 1991, 1992). Supporting understanding, then, demands the construction of semantically connected pieces of text and pictures, activating appropriate pre-knowledge, providing learning strategies for multimedia, and changes of media and learning perspectives to support the construction of comprehensive mental models (Albrecht & O'Brian, 1993). Research (e.g. Mayer & Sims, 1994) support the consideration of individual differences in abilities and interests in order to enhance the understanding processes. In two studies with 75 subjects we were able to confirm that individually adapted information enhances motivational and learning processes within computer-supported learning environments: Audiovisual media produced only a small effect, individual generated information, however, was very effective and was independent of subject variables like computer experience and usability judgements (Hasebrook & Gremm, 1999). These data are explained in full detail in chapter 7 of this book. Glowalla and Hasebrook (1995) conducted studies with 52 students which participated in a hypermedia learning course, all of them were
Web-Based Training
95
novice hypermedia users. In the first lesson they are 'unskilled learners', in the last lesson they were 'skilled learners'. Four month later, 43 of these students attended a re-learning course. All students received exactly the same course materials and configuration of features of the hypermedia system as in the learning sessions. Therefore, in the first lesson they were skilled learners, but 'unskilled relearners', and in the last lesson, they were 'skilled relearners'. The results show that browsing tools, such as paging and hypertext links, were used most frequently by skilled relearners, informational tools, such as a glossary and a keyword index, were used more often during learning than during re-learning. jgtefcwj
^ni-ji^-iLM-aigi— mmmm mm a ; ; iaBiiiiii; ijbung: Blister koytnsn bsim Lozm logischer Problems- kelfm. Glauben Sie nicht?Dann versuchen Sis diefolgende !• Aufgabe zu Iosen.
--/
)
tjbung: Sis sehen zwei imfersckiedliche Graphiksn. Bifts beurt€iiey] SHe gteichzzitigbade Orafiken, Welch's gut im Sinne von $us:>age,krtiftig, und weichs schleckt, ah.o wem'g aussriSpkriifligisi.
JSSL
S^UHJigb "T^S^p J5T p1iaf? W»J = . >ij> gr e^v La>\ 1* v. *"
gut
^
schiecht
i;l|l^^gi|S|||
" •* ^'f»*f
"Hi
~ "
- £;i _
-. ~
• a a** i ti i }•<• Y H C
giit
zu
C
parten schlecht
|;:||||^|i|i|i:::||
Gelurntest: Farbenfioh Gelurntest: Ordentliche Gedaiiken Nachdem unsei Sehsystem nach sineiri srsteri Blick erne Organisation der Figuren eiseugt hat, beginntiittt demsiveitenBlici; die Suche und dc.r VergJeJch von infomiatjomn, ttifiist aufgrumi einei beslimmtenFragestellung. Eei dec Analyse von Diagrammen eollten zunachst die visr grundlegendenBestaridteile von Diagiartimeii betf&chtet werden:
;;; i;iBe|iii^i|*:;y^iie!i;;:;:;
iiiMiigli Figure 25. A content screen of the Web-based training course 'Learning in electronic media' provides brief explanations, interactive exercises (indicated by the mouse icon) and psychological self assessment (indicated by the brain icon).
96
Learning Support Systems
In conclusion, multimedia information is first encoded in simple text and image bases; using more sophisticated elaboration and inference processes mental models can be generated based on the information in the text and image bases (Hasebrook, 1999). Higher levels of processing can help to generate appropriate static and dynamic mental models (Hegarty, 1992). The Role of Meta-Cognition Self-regulated learners govern a broad variety of cognitive and metacognitive strategies to fulfill their learning tasks. They monitor and - if necessary - modify their learning strategies. They are motivated, independent, and meta-cognitively active controllers of their own learning processes (Zimmermann, 1990). It is not easy to acquire knowledge about knowledge acquisition, but it is even more difficult to transfer this knowledge to every-day learning tasks. Therefore, it is essential to practice how to apply study techniques and to motivate the use of sometimes time consuming learning strategies. Otherwise, the learners most likely prefer simple study techniques, such as accessing all pages in sequential order. Our aim is that is to use 'the computer as a tool for learning through reflection', as Collins and Brown (1988) put it. We tested the correlation of pre-knowledge and acceptance of software tools using an on-line expert system for vocational guidance. There is a positive correlation between the students' judgements about ® how well the information provided by the system match their interests, » how well they know subject, and ® how well they can imagine important aspects of the subject matter. However, there is a negative correlation between all these variables and the actual state of information: That is, the more information the students have got, the less they are willing to accept system advisory and the less they have got a notion of knowing. Therefore, information
Web-Based Training
97
leads to more skepticism and criticism (Hasebrook & Nathusius, 1997). Studies related to this issue are reported in chapter 6 of this book. In order to motivate users to apply learning strategies we integrated about fifty so called brain tests (cf. figure 26): Each brain test consists of short psychological experiments which can be easily conducted within a few seconds and illustrate important features of human perception and human memory. First experiences ascertained that it is amazing and highly motivating for students to test their own perceptions and learn about human cognition. The list of the brain tests comprises the following subjects:
Fortgesetzte Bewegung Oft wird behauptet. Tennis spieler verfolgten den Tennisball genau rait Augen. Unsere Augenbewegungen sind jedoch vielzulangsantfur den schnellfliegendenBall. Ein Tennis spieler schautnicht dem Ball hint ether, sondem et schatzt unbewufit die Position, die der Ball vermutlich haben wirdund blickt dann auf diese Stelle, bis det Ball auftaucht (b ei Anfangem geht das haufig - im wahrst en Sinne de s Worte s - " daneb en"). In demfolgenden, kleinen Versuch, konnenSie ausprobieren, wie das GehimBewegungen abschatat Sie werden feststetten," dafi wir feste Erwartungen haben, welchen Verlauf eine Bewegung haben sollte (mid wit waren seht etstaunt, wenn eine Kugel sieh nicht so verhalt, wie wir es uns vorstellen).
Start- dben
Start f.mim
Bild sum Abschltzen von B ewegungen- Drucken Sie dazu auf"Start oben" oder"Start uaten", urn vom token Bildrand eine "Kugel abzuechieBen" Vetsuchen Sie nun genau dort hinzuscfaauen, wo diese Kugel fainter del Abdeckung sum Vors chain kornmen wild Weitere Versyche; Spiralbewegung Phi-Bewegung Veiwandte Themen: Fatbenfioh
Links im Internet http ://wvAV.lihim. 01 g/s ens e s/
Figure 26. 'Brain tests' enable the user to participate in brief psychological self tests and learn about her or his perceptional and memory system; the screenshot depicts an experiment about the perception of movement.
98
Learning Support Systems
* Specialization of the brain hemispheres « Grouping visual stimuli * Processing visual features (angles and distances) * Processing three-dimensional scenes ® Perceiving colors « Perceiving movements » Acoustical illusions » Short-term and working memory « Memory sets and schemes & Cognitive illusions * Mood and memory » Social influences on cognition « Illusions of awareness and consciousness * (Very) long-term memory * Unconscious and implicit learning. Additionally, self tests for each module help the users to check their expertise on each topic and enables them to review directly the parts of the course which are linked to the questions. Integrating Performance Support in Learning Systems Web-based learning systems combine various advantages: access to huge amount of data, up-to-date information, and guidance provided by (self) tests and expert systems. This does not necessarily mean that students enjoy working with electronic media. This is the lesson we learned when comparing four media for vocational guidance (Hasebrook & Wagner, 1997): two of them are multimedia applications and the other two products are printed matter. We measured individual acceptance ratings after having used the four different products with 75 students participating in this study. The results show that printed matter are preferred. This result is statistically independent of sex, education, and experience in using a computer. Thus, the students enjoyed using electronic media, but they rely on printed matter. Expert advice provided by the system, however, clearly increases acceptance and performance of electronic media: Users pick up more
Web-Based
99
Training
information and they consider this information to be more valuable. The more information they have gathered and elaborated the more they lose their notion of knowing and develop a critical approach to expert advisory. Multimedia applications should not be designed to provide 'something for everyone', but they should provide exactly that piece of information which is needed in a particular stage of the decision making process. The effects were enhanced, if individual preferences were regarded. Generic: 'Learning in electronic media' Self-experiments (Meta-cognition)
Specific Web course
Selftests
L^Z
Learning with electronic media (4 modules)
Graphical Overview " * and summaries
Glossary (technical terms)
Web-based training course Interface between WBT andPSS
Help pages
Full-text index Index (key word index and full-text index)
Figure 27. Architecture of the Web-based PSS 'Learning in electronic media' - generic topics and exercises can be linked to specific on-line courses by an interface using the search and the index facilities of the PSS.
A major German bank uses a special version of the PSS 'Learning in electronic media' as a kick-off course and motivational aid to introduce Web-based training in the bank. The aim of the PSS here is not to achieve pre-defined learning objectives but to avoid a 'learning culture shock' by helping the user to become a self-regulated learner. We aim to fully integrate the PSS as a background library and help system for any
100
Learning Support Systems
specific Web-based training course (cf. figure 27). Therefore, we develop an interface that uses keywords from the current HTML pages of the specific course and search terms entered by the user to search the glossary, the keyword and the full-text index of the PSS for relevant explanations and exercises. The entire system can be stored on an ordinary 'open' Web server, such as Apache. New browser version allows for minor adaptations of colors and fonts using style sheets. Additionally, the system runs on a Hyperwave server which adapts the complete graphical user interface of the PSS to the user interface style of the target system: All buttons of the target system are active, styles guides concerning font, color, and navigational tools such as table of contents, are applied automatically. In this way the PSS becomes an integral part of any specific Web-based training course without re-implementing or modifying its contents. Generic Performance Support The learning effect of multimedia has been disappointing, so far. This seems not to be based on a lack of computer literacy but on general deficits in media literacy and in learning strategies that support the integration of knowledge from different sources. Simple help pages are insufficient to cope with these deficits and to strengthen users' ability to become self-regulated learners. We therefore suggest to design and implement generic Web-based Performance Support Systems (PSS) that help the learner to understand and practice appropriate study techniques. Such an PSS has to enable the users to transfer the use of general learning strategies to their actual learning tasks - and it has to motivate this additional learning effort. In two different studies we examined the impact of different types of navigational and meta-cognitive aids on our WBT program. In the first study 51 students with an average age of 17 years participated. The used either an advanced graphical network navigation or a simple folder tree (or hierarchical table) to browse through the WBT. It turned out that they liked the graphical network, although they found it hard to handle. Moreover, prior knowledge and having experiences in using CBT or
Web-Based Training
101
WBT helped them to learn with the program. In the second study we found that our brain tests add 'fun' and therefore acceptance to our WBT. If they are replaced by information which is not related to metacognitive learning strategies, however, the performance in the self tests is not influenced. The PSS 'Learning in electronic media' comprises explanatory texts, pictures, and animations as well as interactive exercises, self tests, and brief psychological self experiments which give an vivid impression why and how elaborated learning strategies should be applied. Furthermore, the PSS discussed here can be integrated into specific Web-based courses by using style sheets, advanced Web servers, and an interface applet that browses the search facilities of the PSS and automatically points the user to appropriate information.
This page is intentionally left blank
Chapter 4
Implementing Electronic Courses: Collaborative and Interactive Distance Learning
Collaborative Learning with Electronic Media Major trends on the educational market are focussing on core competencies and outsourcing of training services, certification of (software) product related skills by the product developers, increase in the importance of international grades and certificates, and growth of the market capitalization of training companies. In 1997 Internet- and Intranet-based training accounted for only 2.4% of the total cash flow of the educational market. Johnston and Moretti (1998) estimate the annual increase in these training technologies to be 140% and 62%, respectively. Today, Internet- and Intranet-based training represents about 15% to 25% of the educational market. Experts play an important role in online discussions. Ogata & Yano (1998) found out that the presence of an expert led to more direct participation in an online discussion, but also to a higher drop-out rate while peer-to-peer discussions suffered from poor active participation if the participants were not directly invited to join in the discussion by their peers. Boiling and Robinson (1999) compared three different learning groups: 1. individual learning with printed matter, 2. co-operative team learning with printed material and special instructions, and 103
104
Learning Support Systems
3. team learning without special instructions using multimedia courseware. Taking into account the prior knowledge of the participants, the authors found co-operative learning to be the most effective training method. Individual and multimedia team learning did not differ significantly. The best performance was observed among participants of the co-operative learning group with high prior knowledge. These and similar findings are in line with recent research results indicating that group cohesion is enhanced when group members are actively managed and master high performance barriers (Tesluk & Mathieu, 1999). LeaderMember-Exchange (LMX) produces higher follower performance as compared to transformational leadership irrespective of physical distance (Howell & Hall-Meranda, 1999), and only content goals with a clear skills improvement focus have been found to support performance in training programs (Brett & VandeWalle, 1999). The Notion of Active Documents Collaboration and communication are crucial elements of any good elearaing system. It is our belief that the communication should not be restricted to person to person communication but the notion of 'communication' should also be possible between persons and documents. This notion we have called 'Active Documents'. It has been used successfully in a number of applications and will be explained in what follows. The idea behind active documents is very simple: whenever a user sees some information on the screen of a networked computer the user can ask or type in arbitrary questions and the document immediately provides the relevant answers. Putting it this way, the notion sounds impossible to implement: How can a document answer any conceivable question? However, although the idea cannot be fully implemented it is surprising how well it can be approximated in large scale distributed networks. The reason for this is that important documents are viewed by a very large number of persons, hence the same questions (albeit in possibly different wordings) come up over and over again. Indeed, experimental data with one of the first users
Collaborative Learning
105
of the Hyperwave e-learning Suite confirms that after some 500 to 1000 users new questions come up only exceedingly rarely (Maurer, 1996). Thus, the basic idea behind active documents is simple: When new documents are added to a server, questions asked are initially answered by experts, ideally but not necessarily immediately and around the clock. All questions and answers are recorded in a database. As a new question is asked, it is checked whether this question is semantically equivalent to an earlier one. If so, the system can provide the answer immediately. As the number of semantically new questions decreases (and indeed approaches almost zero) the experts become superfluous. In the very rare case that a really new question arises the system might answer apologetically: 'This is a very good question. We will forward it to our experts and you will receive an answer within two days'. It is extremely difficult to determine whether two questions, which may show different wording, are equivalent on a semantic level. For instance, how can a piece of software recognize that the question 'Please explain to me how the picture compression techniques GIF, BMP and JPG work and compare them to each other' is really equivalent to T do not understand the difference between GIF, BMP and JPEG coding'. In the next paragraphs we show that this task, as hopeless as it might seem at first glance, has indeed both pragmatic and systematic solutions. Implementation of Active Documents
The Heuristic Approach When a question is asked it is compared with some heuristic algorithm to find an earlier question that seems to be similar. Similarity can be determined by using a number of techniques that can become quite sophisticated: from comparing if different words match independently of their order, to the use of synonyms, to stemming algorithms that take care of the flexion of words, to semantic nets, to syntactic analysis using such nets, a variety of techniques do exist that are surprisingly powerful. In each case, once the system has determined that one question being
106
Learning Support Systems
asked is likely to be the same as another question asked earlier, the system will say 'Do you mean: [the earlier question]?'. It is now up to user to decide whether indeed the answer to this question is what is desired or not. If not, the system may offer other alternatives, but if none satisfies the user, the apology mentioned earlier will be displayed: 'This is a very good question. We will forward it to our experts and you will receive an answer within two days'. As time goes by (in the sense that a document is visited by more and more persons), chances that a question is asked that is not only identical semantically but is also similar in form decreases, thus reducing the amount of time a human expert has to help out, and hence reducing the number of times a question is not answered immediately. The Iconic Approach Users ask questions by selecting some pieces of information on the screen: Their question may refer to information in the selected area. When a question has been answered, some icon or highlighting shows to other users that other persons have asked questions concerning this piece of information and that experts have answered them. If another user also has a question concerning the material at issue, one click suffices to show all questions and answers that have occurred so far. Again, when a sufficient number of users have sent in their questions (and have received their expert responses) the chances are good that all questions of interest have indeed been answered. The iconic approach is clearly particularly easy to implement and has the advantage that semantically equivalent questions that are formulated in different ways will not often arise. The advantage of this technique is, that feedback is provided to the authors of documents as to where users have questions. After all, this may often mean that some explanation is not clear enough or information is missing - hence allowing the improvement of the documents. The iconic approach has been sometimes belittled by just saying that this is not more than FAQ's, and in a way it is, of course. However, the FAQ's are not collected in a long unusable list but in a short list directly attached to where the problems occur. If that list gets too long, something is likely to be wrong with the document, and the document should be improved.
Collaborative Learning
107
The Linguistic Approach The most satisfying approach to handle active documents would be to develop techniques that actually prove the semantic equivalence of questions. Heuristic techniques can only provide guesses whether two questions are equivalent, they cannot prove their equivalence. Both, the heuristic and iconic approaches, use the intelligence of the user to determine whether previously asked questions are relevant or not. It would require a complete understanding of natural language in order to prove that two pieces of text are semantically equivalent. However, we can consider a compromise: Rather than allowing a full natural language, we restrict our attention to a simplified grammar and to a particular domain for which an ontology is developed. Clearly, sufficiently restricting syntactic possibilities and terms to be used will allow to actually prove the equivalence of pieces of text. Attempts to adjust it to the active document situation are currently been carried out (Heinrich, Patrick & Kemp, 1999). Up to now, the restriction on the wording of questions and the domain specificity are serious problems. It seems clear that this technique will not be suitable for the naive user, yet. Applications of Active Documents We have stated that 'really new' questions usually do not arise after a document has been used by some 500 to 1000 users. This figure comes from a large e-learning experiment with a multi-national company with some 200.000 employees. Typically, the WWW or Intranets also contain information that changes over time. What we have discussed so far does not handle this situation at all. The notion of active document is only applicable to fairly static information and large number of users. Question and answer dialogues should be time-stamped so that they disappear automatically when they are invalid. Actually, the situation is better than it seems: An indirect step can alleviate the problems in a very elegant way. The answer to the question 'How much snow can I count on' should not be '30 inches' but rather 'Find information on current skiing conditions under 'snow report" (where 'snow report' provides a link to server which is updated on a regular basis).
108
Learning Support Systems
We have encountered the notion of active documents the first time in connection with e-learning experiments and have found them to be very useful in this connection. Note that the 'critical number' of users of the same document will usually only be reached if the courseware at issue is offered on the WWW for a large audience, or in a substantial Intranet, but will usually not be reached when using material for e.g. teaching typical university classes. However, even in this case, the mechanisms described eliminate many duplicate questions and are hence useful even if the 'saturation level' is never achieved. However, in addition to e-learning there are three other areas where active documents seem to be particularly useful: * one is support for software, ® the other is in connection with help desks, and * the third are digital libraries. Let us elaborate this again by means of examples. Suppose a company releases a new software product to a large number of customers, with the corresponding documentation on the WWW. In the past, support staff would always receive a great number of questions. Support staff would then have to consult a manual, locate a certain page and a certain paragraph or line, and examine the situation - only to find out that this error or 'bug' had been pointed out many times before, and that the development team already fixed the error (or the manual, for that matter). If each page of the documentation would have been an active document, this situation would never have occurred: After the problem at issue is pointed out the first time, other customers have no need to ask the question any more. The situation is similar in help-desk situations when customers do not understand a manual. There are two interesting additional aspects in this case. First, customers may ask the question not via an active document on the WWW but by telephone. Help desk staff may, however, use the active document to find the answer to this question. Second, in this case the linguistic approach could come in handy: The knowledge domain is limited, and the staff of the support center may well be expected to be able to translate customer queries into queries supported by the linguistic approach. We have argued in Maurer (2001) that active documents will also play a fundamental role in digital libraries in the future, since those
Collaborative Learning
109
libraries will turn more and more from static information repositories into interactively and collaboratively used centers of human knowledge. In what follows, we show that even partial implementations of 'active documents' (namely, the heuristic and the iconic approach) lead to considerable improvements of the learning process. The Learning Environment A WBT about currency management was developed by Bank Academy in charge of a major German bank. The WBT is based on the Hyperwave information server and its learning platform GENTLE (Maurer, 1998). This software stores and maintains the user interface (e.g. buttons, frames), the structure (e.g. links, hierarchy of pages) and the actual content (e.g. HTML-pages, images) separately. Thus, all complete WBT pages are composed on demand and may contain individual information, such as notes and user defined links, without interfering with the contents of the WBT delivered to other users. The WBT consisted of five modules comprising approximately 100 pages each. About one third of the pages contained animations or interactive exercises, such as calculators and interactive telephone orders. Important content areas, such as definitions, examples, exercises, and team instructions, were marked by special icons. Half of the participants were automatically pooled in learning teams with five persons each by the system and the other half studied individually. All participants were allowed to take notes and write contributions to the discussion forum. All notes and contributions were typed according to their contents, that is, the user decided whether she or he wanted to type in a question, an answer, an agreement, a disagreement or a simple remark. All notes were linked to a particular phrase or page in the WBT. Additionally, different access rights could be attached to each note: Public, learning team (if available), and private. Private notes were marked with gray icons, public and team notes with green icons. All notes containing questions were sent as an e-mail to an expert who decided whether he or she wanted to respond to that question. The notes which had been responded to by an expert were marked with a blue icon. All public notes were
Learning Support Systems
110
automatically copied to the discussion forum with a link in the note enabling the user to access the anchor of the note by clicking on that link. EMMS.
mmm &if*-¥-
^gjMZi -ty&Mf. yMpM^.
jftaasgichma
JM
P&& Modul i fur BIWB.'G • .3 Kj?gahar<del 1 31 Proi.3- o-icr Menoftfirioiigruriq 1.3 2 Elrrfiugr.e a>,£clen Delf oner am Tehran
LiiEUM:
r
Unser Kunde vertouft EUR 100.G00.0Q unri katifl. USD -m 1.1 WO K ^ s ^ valuta 15 ,1-anusr
r
Uii3erKunoe ksufl Eii.R 100 000,00 unci ve:l:-airfl USD zu 1.0355 K33se \'3'ijt3 15 januar
r
Unser Kunde kairft EUP 100.000.00 unci verkauft USD zu 1,1000 Ka*se Valuta 15 J^nusr
Dtskusf ion zu Motliil I Worte>bufl'.
vvird ells ' Handlers" dan Auftrag korrekt ausrunren? Joictitm Hap jtorook
B
a
© • ^^:^'P#^^^^^i^-:^
Figure 28. The design of the learning environment: Navigation and table of content (left), general tools (above), and an interactive audio exercise (centre) is displayed by the Hyperwave system.
The study reported here was conducted with this WBT. The WBT consisted The notes did not only support the learning process by motivating the users to discuss the subject matter of the WBT. They also provided a useful source of information for the adjustment and improvement of the system, because the users took lots of notes which described technical or design problems. Furthermore, a background library of encyclopaedias and news services enabled the user to access a vast amount of background information and most recent information without leaving the WBT environment.
Collaborative Learning
111
Experiment 1: Collaborative Learning Strategies
Participants of Experiment 1 Outlets of the bank all over Germany were asked to nominate trainees of their corporate finance departments for a two-day seminar about currency management. Seventy persons were assigned to the one-day WBT, thirty persons to traditional face-to-face seminars resulting in 64 complete data sets of the WBT users and 30 complete data sets of the seminar participants. Only ten of these 94 persons were female; the mean age was 35.2 years (standard deviation = 12). Material and Procedures of Experiment 1 The WBT learners used the WBT described above. The WBT was based on the printed material, such as papers and slides, used in the seminar. Additionally, the trainer of the seminar groups served as the subject matter expert of the WBT development. In the beginning, all subjects filled in a survey about personal data, that is, gender, age, professional experience, prior knowledge, WBT experience and their personal expectations. Furthermore, they responded to 16 multiple-choice questions about currency management. While learning with the WBT, the users' inputs were automatically recorded by the system. All WBT participants learned about the WBT features conducting an introductory module which took them about 20 minutes to complete. Each module started with a brief overview and offered a multiple-choice self test. After having finished a module, the WBT offered an evaluation form with questions about the correctness, jobrelatedness and user-friendliness of the WBT module, which could be filled-in voluntarily. After the training, all seminar and WBT participants filled in a second survey about their experiences with the training course and responded to a multiple-choice test with 24 questions: 16 questions were taken from the pre-test, 8 questions were newly introduced. The survey was paper and pencil work, all multiple choice questions were presented at the
112
Learning Support Systems
computer and were rated by an expert team according to their difficulty. Test and survey were filled in anonymously and without observation in order to avoid social desirability distortion (cf. Richman et al., 1999). It took the participants about 40 minutes to fill in the survey and respond to the multiple-choice test. The WBT course took about 8.5 hours (standard deviation = 1 ) and the seminar about 12 hours of net study time to be finished. All WBT learners took part in a moderated team discussion about their experiences using the WBT. The results of these discussions were recorded by the moderator. Design of Experiment 1 The first experimental factor was the comparison of the between factor 'seminar vs. WBT learning' with respect to acceptance and performance criteria. Another set of experimental factors was realized by a mixed design within the WBT group1. As mentioned above, one half of the WBT group was automatically assigned to a learning team resulting in the between factor 'team vs. individual learning'. In every second WBT module, the learners were instructed to read the overview and to take the self test prior to the access of the module and then to decide - based on the test results - whether they want to go through all pages or only parts of the module. This instruction resulted in the within factor 'complete vs. selective learning'. Each module contained several audio and video files and a simple text version of the same content. The system automatically assigned the WBT users to different groups which had access to the audio-visual media in every second module. This resulted in the within factor 'text vs. audio-visual media'. All factors were counterbalanced by a Latin square procedure among the subjects. In summary, the experimental set-up of the WBT system resulted in a mixed design with the between factor 'team vs. individual learning' and the within factors 'complete vs. selective learning', and 'text vs. audio-visual (av) media'. 1 The design of an experiment refers to the set of variables and the method of their measurement: A between factor measures differences between groups of subjects receiving different treatments (e.g. one group attending a seminar vs. another group using a WBT); a within factor measures differences within a group (e. g. different orders of modules of the same WBT).
113
Collaborative Learning
Additionally, qualitative data were collected by interviews with the participating experts and by team discussions after the training program. Results of Experiment 1 All survey ratings are ranging from 1 ('very good' or 'I totally agree') to 5 ('very poor' or 'I totally disagree'). As the scores of the multiple choice items are differing according to their difficulty, all test scores are expressed as percentage of the maximum score (ranging from 0% to 100%). Due to the variable cell frequencies of the design and some missing data, the General Linear Model (GLM) procedure of the SPSS statistical software package was used to analyze the data. A GLM is comparable to mixed, multivariate analyses of variance (MANOVA).
Table 8. Test results in % of the pre-test (16 items) and the post-test (16+8 items) as a function of learning group (seminar vs. WBT), gender and age.
Pre-test Post-test Pre-test Post-test
Total n=94 56.7 76.5 43.2 72.9
Gender Female n=10 59.2 75.0 * *
Male n=84 55.9 73.3 43.2 72.9
Age in years 36-45 20-35 n=29 n=39 54.2 59.8 68.2 80.1 52.2 45.1 70.4 70.1
46-55 n=17 51.7 65.3 32.9 66.1
56-65 n=9 61.1 75.0 11.1 37.5
* no female participants in the seminar
Comparison of WBT and Seminar The study time of the WBT and the seminar differed significantly (8.5 vs. 12.0 h; F[l,92]=319,9; p<.001)2. The statistical analyses showed main effects of the learning group in the pre-test (F[l,80]=9,3; p<0.01) and the post-test for the 16 old items (F[l,80]=5,l; p<.05), but not for the 8 new items (F[l,80]=1.7; not significant, n.s.): The WBT group started with higher test scores and showed better performance for the items from 2
p indicates the probability of an experimental finding. All p<0.05 are considered to be significant, that is, the resulting data show a statistical significant difference.
114
Learning Support Systems
the pre-test. But there was no significant difference concerning the new items. Taking into account the pre-test scores as a covariant, the main effect of the learning group is reduced to a weak tendency for the posttest results (F[l,80]=1.7; p<.2) and the covariant is highly significant (F[2,90]=29.1;p<.001).
30 20
-O-WBT —•—Seminar
20-35 yrs
36-45 yrs
46-55 yrs
55-65 yrs
Figure 29. The age effect of face-to-face training (cf. table 8): Test results (in %) are lower for elder persons who attended traditional seminars instead of WBT.
The test results showed no significant differences for female and male participants due to the small number of women, although they did slightly better than men (75.0 vs. 73.3% in the final test). The participants were grouped into four categories according to their age: 20 to 35, 36 to 45, 46 to 55, and 56 to 65 years. There is a tendency that young participants performed better in the final test than older ones (F[194,3]=2,4; p<0.1), but there was no significant difference in the pretest results (cf. figure 29). Most importantly, there was an interaction of learning group and age group: Young seminar participants learned more than older ones, but there was no such difference within the WBT group
Collaborative Learning
115
(F[194,3]=3,2; p<.05). Table 8 summarizes the results of the pre- and post-test as a function of gender and age. All learners judged their prior knowledge to be on a medium level (WBT 3.5; seminar 3.6). After the training, the judgement of WBT participants concerning their knowledge was improved, but the judgement of the seminar learners was significantly better (2,7 vs. 1.3; F[l,80])=39.0; p<.001), although their test results were lower than those of the WBT learners. There were no more significant differences in the individual judgements of the WBT and the seminar group. Table 9 summarizes the scores of the individual judgements. Factors of Online Learning Team vs. Individual Learning. There was no significant difference of the pre- and post-test scores between team and individual learning. There are only two tendencies: Individual learning leads to slightly better acceptance of the WBT than team learning (F[l,57]=1.7; p<.2), and to a better judgement of the knowledge acquired during the training (F[l,57]=2.8; p<.l). In general, the two covariats affect the post-test results, but not the pre-test results: A high judgement of prior knowledge and experiences using a CBT lead to better post-test results (F[l,59]=13.9; p<.001 and F[l,59]=6.2; p<.05, respectively). Experienced users of the corporate Intranet, however, did not show significantly better test results (cf. figure 30). Complete vs. Selective Learning. Once again, complete and selective learning strategies did not lead to significant differences in test results and acceptance ratings. Therefore, we checked the number of page and function calls as a function of the different learning conditions. On average, 35 notes were read, seven taken and the forum was accessed 31 times per module. Each user took an average of five notes per module and additionally wrote two messages to the forum. Most of the notes were public. Selective team learners tend to use the note function more frequently than the other learners (F[l,60]=2.1; p<.2). Complete learners accessed 398 pages of the WBT and selective learners 411, group learning led to 395 page accesses and individual learning to 412 page accesses. There were no significant differences in the number of function
116
Learning Support Systems
calls and page accesses in all groups. Table 10 summarizes the data for the factors team vs. individual and complete vs. selective learning.
Table 9. Individual judgements and acceptance ratings in the pre- and the post-test as a function of learning group (seminar vs. WBT); scores are ranging from 1 (very good) to 5 (very bad).
WBT Seminar
Pre Post Pre Post
Individual Judgement or Acceptance Rating Prior/Gained Intranet/ CBT WBT/Seminar Knowledge* training Experience Comparison Experience** *** 3.5 3.3 4.7 *** 2.7 3.2 2.9 *** 3.6 3.5 4.6 *** 1.3 1.0 3.7
*: self estimation of prior knowledge and knowledge after the training, respectively; **: judgement of Intranet experience (pre) and training experience (post); ***: data were collected in the pre- or post-test phase, only
Table 10. Acceptance ratings (1 to 5) and system calls per module as a function of learning strategies within the WBT (team vs. individual learning and complete vs. selective learning). System Calls Readg. Writing Notes Notes Team Learning
Acceptance Ratings Prior Knowldg. Recmd Know- after WBT* ledge Training 31 3.6 2.9 3.2
Access Forum
Complete 31 Learning Selective 43 Learning Individual Complete 35 Learning Learning Selective 33 Learning : The participants were asked whether they would training delivery
30
3.4
2.7
3.4
29
3.6
2.6
3.2
31
3.3
2.7
2.7
recommend WBT as a major source of
Effects of Audio-Visual Media. There is a tendency that learners with audio-visual media did better in the post-test with 24 items than learners
117
Collaborative Learning
without (77.3% vs. 74.7%; F[l,63]=2.7; p<.l). And there is a tendency for better acceptance of the modules with audio-visual media than those without (F[l,41]=3.2; p<.l). Table 11 summarizes the test results and acceptance data of the modules with and without audio-visual media.
90 80 70 60
S^^^^^^_ 5=*^——=8 \
50
\
40
n
—•—Team Complete —•—Team Selective —O— Individual Complete —D— Individual Selective
30 20 10 0 Pre Test
Post Test
Figure 30. The impact of learning strategies: Test results (pre and post tests in %) as a function of selective and team learning strategies.
There are some interesting additional results concerning the module surveys: Module 3 and 4 contained many calculations as interactive exercises while module 2 and 5 did not. Thus, module 3 and 4 got worse acceptance ratings than module 2 and 5, especially concerning their user friendliness and their job-relatedness (F[l,41]=5.2; p<.05). Furthermore, only half of the module surveys contained a direct feedback summarizing all user inputs in simple bar charts. These surveys with direct feedback collected 372 user inputs while the surveys without direct feedback collected only 312 inputs. Thus, it seems to be an easy way to improve compliance to provide direct feedback to the users of surveys.
Learning Support Systems
118
Expert Participation The experts received about ten e-mails per day during the learning phase. The questions ranged from serious questions to complaints, e.g. about the number of calculations which had to be performed by the learners. The experts wrote only 20 answers reviewing the questions they had got via e-mail and via the forum of the WBT. The average length of such an answer was about two to four sentences (40 to 80 words). In professional virtual seminars the input of experts is much higher, e.g. in a virtual seminar at the University of Maryland with 15 sessions the experts wrote about 8,000 words and the participants about 2,750 words (cf. Bernath & Rubin, 1998). However, all experts claimed to have given strong support to co-operation and team learning based on electronic discussion forum or e-mail messaging. However, there was no clear organizational procedure that enabled the experts to withdraw from their normal duties and work on the WBT, instead. Table 11. Test results in % and acceptance ratings (1 to 5) as a function of media use (text vs. audio-visual media) in the WBT.
Text only Audio-Visual Media
Test Results Acceptance Ratings Pre-Test Post-Test Correctness JobUser(16 Items) (24 Items) Relatedness Friendliness 57.7 74.7 2.6 3.3 3.2 56.3 77.3 2.4 3.1 3.2
Team Interviews and Discussion In the team sessions after the WBT training positive and negative aspects of the WBT were collected and discussed. All participants wrote down on a black board whether they considered the WBT to be a very negative, negative, neutral, positive or very positive means for training. As in the surveys, the individual judgements summed up to a neutral attitude towards the WBT. Positive aspects discussed by the participants were: * self paced and self directed learning, « free choice and access of information,
Collaborative Learning
* direct feedback for tests and inputs, » fast and efficient learning, and * opportunity for distant communication. Negative aspects were: « too many and too difficult calculations, * too much content not directly targeted to the departments of the company, * difficult handling of the calculation forms, * too many overviews and indices, * learning time was too restricted , and * not much input from experts.
119
different
Learning Culture One of the major conclusions of this study is that success does not come simply by using the latest online techniques: The learning culture of the participants and the experts involved in the WBT clearly did not support the success factors of online learning. Although the participants used navigational and communicational features quite frequently, they did not receive much input from the experts, and they did not pick up new learning strategies, such as team and selective learning. This line of reasoning is supported by strong impact of self esteem (judgement of prior knowledge) and of CBT experience on test results. The age effect indicates that WBT is offering a more equal opportunity for learning than seminars. Additionally, seminar participants considered their learning results to be better than WBT learners did, although objectively it was not. A future study will examine a similar WBT environment. However, there will be a variable learning time which is not going to be restricted to a single day, clear instructions for the corporate departments how experts should be involved, and the introductory module will not only give a brief overview of the WBT features. The introduction will actively train communicative skills and the selection of information from comprehensive online learning environments (cf. Hasebrook, 1999). The WBT at least reached the performance of face-to-face seminars within a
120
Learning Support Systems
shorter period of time. Thus, WBT is an effective means of training but additional features, such as expert involvement and new learning strategies, have to be trained and motivated carefully. We shall discuss this issue in more detail in the following chapter.
Chapter 5
Implementing Online Curricula: New and Emerging Media in Distance Education
Learning Support Systems Van den Berg and Watt (1991, pg. 119) compared multimedia in competition to a classroom lecture, multimedia supplementing a lecture and multimedia replacing a lecture. They drew the conclusion: 'Objectively the academic performance of (multimedia) users was not different from those attending classroom lectures [...] Although, positive about (multimedia) technology, they indicated that they would prefer to use it as a supplement to lectures and books' Many other studies have confirmed that multimedia applications enhance learning, only if the individual skills and abilities match the demands of the learning task and the functionality of the multimedia system (e.g. Barba, 1993; Mayer & Sims, 1994). Therefore, it is necessary to teach users strategies and concepts to use multimedia applications. Additionally, it is necessary to adapt the system to individual abilities and the overall learning environment (Schulmeister, 1996; Larkin & Chabay, 1992). In what follows, we want focus on the description and discussion of an online course entitled 'New and Emerging Media in Distance Education' (NEMDE) which was developed as a mandatory course module within University of Maryland University College's (UMUC) Master of Distance Education (MdE). This master program is jointly offered by UMUC and the Center for 121
122
Learning Support Systems
Distance Education of the University of Oldenburg, Germany, and it has won several awards including two Sloan Awards for Distance Education. Before we go into the description and discussion of the NEMDE course we would like to explain which system we are using in order to implement those courses and curricula. The event and learning management system we are going to describe is based on the following ideas: 1. Learning management should be part of a company's competence management, and competence management should be part of the enterprise resource planning (ERP). Thus, learning management systems should not replace but extend existing ERP system using existing data repositories and workflows in order to enhance educational planning and controlling. 2. Corporate learning processes are a self-directed and goal-oriented projects as a part of the overall business processes; e-learning should combine the greatest possible flexibility and the best possible coaching or tutoring of the learning process of groups and individuals. 3. Technology solves technological problems in order to achieve more efficiency. Learning is a social process which cannot be directly supported or influenced by technology. E-learning systems, therefore, should aim to organize learning resources and business processes in a way that self-directed learning processes, group learning, and tutoring are enabled and supported 'logistically' and 'organizationally'. In conclusion, a learning management system (LMS) should not be a ERP system to administer tangible learning resources, such as rooms, blackboards, and books. We refer to LMS which focus on resource administration as learning administration systems (LAS). Learning management should be centered around the support of learning processes, such as learning objectives and learning steps, individual and group time schedules, peer communication and collaboration, learning control, and active expert involvement and tutoring. Those learning support systems (LSS) should take into account that the roles of the persons involved, that is, students, tutors, administrators, authors, experts and the like, will be flexible and overlapping: Tutors will be experts for one course, students in another and administrator for further courses. Therefore, we need flexibility in order to support the learning process,
New and Emerging Media
123
and we need flexibility in order to continuously adapt to the changing role models and job profiles in a modern enterprise. This results in the technological demands to provide at least flexible roles and access rights, and open interfaces to general ERP systems, such as SAP or PeopleSoft. Event and Learning Management Germany's leading training institute in continuing finance education is the Bank Academy in Frankfurt (Germany). This non-profit institution was founded more than forty years ago by the leading German banks and has expanded considerably, both in terms of geographic reach and scope of educational products. By the end of 2000, the Bank Academy launched its own incorporated e-learning firm, efiport, Inc. 'efiport' is an abbreviation for "educational financial portal"; strictly speaking, efiport, Inc. is a start-up firm. However, because of its close affiliation with the Bank Academy and support from the Frankfurt banking community, efiport takes advantage of established marketing channels and the synergy between a private education provider, a four-year college, and an international consulting firm. The shareholders of efiport, Inc. are the major German banks in Frankfurt and the Bank Academy. Consequently, efiport's main customers are also its shareholders. Efiport's event management and learning support system is an integrated system which supports administrative as well as learning processes. Learners and trainers will have access to catalogue and booking components, as well as various components for communicating and structuring their learning processes. The main system comprises three parts: * Event Management ® Learning Support ® Communication & Knowledge Management. These main components are complemented by a skills management system which has been described in chapter one of this book. All components of the system are interconnected and are mainly directed by the Event Management System (EMS). Single components of the system can be expanded by those permitted access to the system, that is, trainers, program coordinators, community managers, and administrators. EMS is
124
Learning Support Systems
the main 'operating system' for all people and event-related data: Planing and creating events as well as managing catalogue entries and bookings. Following are the several components of EMS which house basic data: * Event Locations: information concerning hotels, rooms and other locations. » Participants: overview of the participant's basic data imported from the EWD system. This data also contains information detailing where the participant is located within the organization. ® Trainers & Institutions: information regarding trainers and training institutions. * Event Equipment & Supplies: information regarding equipment and supplies such as projectors, user surveys (ironically called 'happy sheets'), or basic seminar supplies (such as pads and pencils). * Events: information for a range of events - from on-site seminars to communities - as well as make catalogue entries, complete booking forms and oversee other relevant information. * Curricula & Programs: This special component helps to merge the data from single events into curricula and programs. « Reporting: The reporting module helps you to analyze different basic data. Different types and categories of events are supported: « On-site: An event with one or more on-site meetings which are scheduled for a certain date and location. * Seminar: On-site training over one or several days for a fixed group of participants on a specific subject with clearly defined learning objectives. « Conference: An on-site event over one or several days in which experts and interested parties come together in order to talk about and present specific topics and developments. « Workshop: A fixed-time on-site event for a fixed group of participants discussing a concrete topic with the goal of identifying further steps and delegating further responsibilities. «• Counseling: An on-site Counseling or Coaching event. * Individual Coaching: Helps an employee or manager to identify processes and methods used for finding concrete task-
New and Emerging Media
125
management solutions or provides consultation on specific questions. ® Group Coaching: Consults groups, teams or organizations in the process management of project, strategy or structure development. This event can be offered as a single meeting or a series of meetings. * Online: An net-based event offered on a learning platform with various components such as forums, a learning plan and WBT offered to a fixed group of participants. * WBT (Web Based Training): Single online learning units which can be worked on individually via computer and at the learner's pace (Intranet and Internet). * Online Seminar: A virtual seminar conducted by a trainer, offered on a learning platform with various components such as a learning plan, knowledge acquisition and interaction through virtual classrooms or forums. This event is offered for a fixed group of participants and has clearly defined learning objectives. * Community: A virtual platform in which people with similar interests exchange knowledge through forums and databases. They communicate with each other in order to develop concrete topics or to work on tasks and projects. It is a closed community which can be extended as needed. « Library: A virtual platform providing information on a special field of interest (e.g., as hyperlinks or data files). ® Two types of 'Blended Learning': Blended learning refers to an event with online segments and at least one on-site meeting. * A "Blended Type 1" event includes at least one on-site meeting attended by a fixed user group. On-site meetings are complemented by online support. *> A "Blended Type 2" event provides online training for participants of different on-site meetings who may vary from date to date. This option is used to provide online information for participants in different regional conferences. Web based learning management systems or so-called learning platforms guide the participants through a learning process (useradministration), manage the contents, the tests and exams. For this
126
Learning Support Systems
reason, web based learning platforms are mainly distinguished by three types of tools that constitute the online learner environment: The information and presentation tools present the contents of a course. Information tools fulfill the task of spreading information throughout the duration of a course (event management), e.g. via a schedule, the description of learning goals or a welcome page with announcements by the trainer. The presentation tools present the course materials and resources. A wide variety of media can be integrated here, starting from simple texts with images or continuing links to the internet via sound and video sequences to entire web-based training (WBTs) with interactive exercises and complex simulations. The communication tools support the asynchronous (forums, computer conferences) and the synchronous communication (chat, shared white board, video conferencing) between participants of an event and their trainers, as well as the communication among learners. Last but not least, the assessment tools are needed for filing, grading and managing tests. The trainers can use the test and survey administration to file answers to problems or self evaluation tests or graded exams. Experience with media-supported long-distance learning and elearning systems has proven that a continuous participation in the learning process, and ultimately its success, are mainly dependant on how much support the learners get. For this reason, the web based learning environment of efiport is aimed at giving learners consistent active and passive support. It is a Learning Support System (LSS). A separate tool that is integrated in the learning environment has been designed to create, manage and supply self-tests and exams. The LSS is based on a role concept, in other words different user groups with different corresponding functions and authorizations can be shown: Learners, trainers or tutor, and administrator. Please note that the role concept is a technical definition of administrative authorizations, and does not imply categories of educational tasks. During a course a trainer could theoretically assume the role of administrator, for instance if he or she were to write up a test or create a forum. In this context a combination of methods of on-site learning and long-distance learning (blended learning) are advantageous, whenever possible. Apart from
127
New and Emerging Media
online-learning programs and tutoring by trainers all of the elements of classical training can be utilized as well, e.g. seminars, reference books and tele-coaching. ^ a i , : - : ^ * ^ : - - : - : ^ " M B ^ : : •'$&$$$• y f'&%&:•:''& '• £ < '.•• X ' ' % \ •
Academy Of!!:::":
DAIMLF.ROIRYSLHU
DaimlerChrysler Services p Chief AdministidBon i Publishing
;V••.•••:•'.• "•".•;"_'*.•'
«aiaiog
•
!> RPSUlfc » Maintain communities * Event Management »> Ovewie*'ol e^nts
• s w a t h e c^aK>0«« | : scMchierm^-
» My events
: i
•> M> Hhr<)rie~
j 1 .
> EwantiCDasheabyme
. f
» Contact
i
fttle
~B^
] Title
™Sa
JTtfle
.i*P
Ud wicsvl Training JBanking pending Jcoechlng |E-LeaM-iing J~: Restnct to cu no
Language of E\ art On-Ste - Semuist pn-S>le - Confai cncs jOn-Srte-VVbr&shop d lirre -Web Based Trdning <\M3T) |On-iins - Onlms Training re - Community
M-r Search ~| [
Alphabetical List
1
Set to start .. reboot
Figure 31. Integrated event management and learning support system developed by efiport in charge of DaimlerChrysler Services: Integrated search facilities for documents, links, forums, catalogs (in display) and people.
LSS can display single training or study programs that cover several semesters. The courses are divided into several learning steps. These learning steps are marked by start and due dates. Depending on the preliminary setup of the system, the learner can set the dates or they are pre-set. Either way, the personal learning schedule for the participant is developed. A learning step can contain several documents: from the complete WBT to an online schoolbook, or from an Excel-sheet to a PDF-document. The learning schedule structures the course. It supports self-organization and learner motivation and helps learners to better review their progress. The LSS actively supports the participants of a
128
Learning Support Systems
course to make the learning process easier and more rewarding. It's not so much the technical possibilities that are the core of the LSS design, what is more important are the individual demands of the learners and the didactic options that online learning has to offer. Active support begins on the start page where fast and efficient orientation within a course is given. This page is always personalized and displays new and important information that is relevant for the learner: ® New messages from the learner's trainer « New messages in the forum * New tasks and announcements * New and updated documents and links. LSS lets both, the learner and the tutor, determine their learning process by themselves. Self-determined learning gives the learner much more responsibility for designing his or her own learning process. For this reason it is important to establish a balance between maximum flexibility and a pre-defined structure that makes orientation easier. The learning schedule tool is one of the components of LSS that is a basic element of each event. The learning schedule supports the selforganization and time management. A course can consist of any given number of learning steps. A learning step can contain any given number of documents. These can be made up of different kinds of document formats. The learner's schedule contains an overview of the required learning steps of a course, complete with begin and due date. The dates can be adjusted to your personal schedule, provided the respective preset structure of the system allows this function: Dates can be changed, and the entire time frame newly calculated. In other words, the structure that is suggested by the learning schedule tool is not a fixed frame, rather it allows an individualized and flexible scheduling of the learning process. The trainer or tutor gives the learners the information concerning how strict the course structure is and where your boundaries for flexing the structure and adjusting the schedule to your personal needs end. In addition, the learning schedule displays the estimated time needed for each learning step, and an icon shows the current status. For each learning step there is a short description which enables the learner to see at a glance the goals she or he is expected to have reached at the end of the respective learning step. This enables her or him to evaluate
New and Emerging Media
129
the importance of the individual learning step within the broader scope of the complete learning process. From a trainer's perspective the learning schedule supports his or her task of tutoring the participant of a course. Trainers command a view of the personal learning schedule of each learner. They can immediately see if the dates for the learning steps have been changed, and whether the deadlines have been met. This enables trainers to accompany the scheduling process and give learners advice and active support if they need help in keeping up with their schedule. The progress control function is an instrument for checking the learning schedule. Self-control is motivating and helps to keep up selfdiscipline. The quantitative progress control is a chart containing the learning steps, which shows the due date for each learning step and the required study time. The qualitative progress control shows the finished documents and new documents in order of learning goals. In other words, one can see exactly which documents still need to be worked through before having reached one or more learning objectives. The learners can take notes and make annotations to all learning materials. Each document of the current learning schedule shows the progress status and a short description. Conducting the NEMDE Course Since efiport is engaged in the business of implementing online-learning platforms in large companies, we know that many potential advantages of multimedia get lost by ignoring the learner's needs during the design and the implementation process. When designing the NEMDE course we tried to prepare our students to be able to avoid these mistakes. Knowledge about how multimedia effects the learning process combined with the knowledge of learner's needs should be an appropriate basis to reach this goal. Knowledge is better achieved by experience than by reading so we decided to give our students the possibility to reflect their own needs in multimedia and distance-learning. The first step to do this was to give the students access to a web-based training about learning with new media. We hoped that the students would use the WBT as a source of knowledge about multimedia learning as well as an experience
130
Learning Support Systems
of multimedia learning itself. For the final assignment we designed a case study about implementing distance learning with multimedia content in a large company to encourage the students not to neglect the environment' of multimedia which consists of the learners as well as the instructors or the promoters of educational programs.
Figure 32. A learning support system, displaying a flexible time schedule (left) and a table of contents (right) developed together with the management center st. gallen (mzsg, Switzerland) in order to conduct the online course 'managing & performing effectively5.
Computer Mediated Expert Communication As has been mentioned before, experts play an important role in online discussions (Ogata & Yano, 1998) supporting direct participation but also to avoid higher drop-out rates. While comparing different learning strategies Boiling and Robinson (1999) found co-operative learning to be
New and Emerging Media
131
the most effective training method as compared to individual and multimedia team learning which did not differ significantly. A transparent and democratic communication style as proposed in LeaderMember-Exchange (LMX) produces higher follower perfomance as compared to other forms of leadership in face-to-face as well as online settings (Howell & Hall-Meranda, 1999). Now let's have a closer look to the expert's role in an online course: The perfect online-course instructor should be an expert in the training field itself, an expert of teaching in general and also an expert of distance-learning via the internet. This combination is not yet found very often within the labor market. And we should not forget that distance-education is not only offered by universities or other educational institutes with full-time-trainers but as a training instrument in large companies as well. A common practice is to hire specialists to participate in online-courses as instructors without caring about their pedagogical skills. This model also implies that in most cases the expert has not much time to care about asynchronous online-conferences - with the result that the individual student quickly gets the feeling of not being looked after. Our experience in different courses shows that far more than 50% of the student's questions do not require the answer of a subject matter expert (SME) but they concern organisational or technicalpoints. Online tutoring could easily be shared between the SME and a tutor to guide and control online discussions and group activity. Knowledge acquisition is done by self study based on the numerous selected mandatory and recommended readings supported by the SME's participation in the conferences focussing on the subject matter itself. Our online tutors have practical experience in face-to-face seminars as well as in distance education. They are engaged to take part in the discussions very actively and to request the SME's expertise whenever necessary. Learning Strategies In general, effects of adaptation and system-controlled tutoring have been small or medium sized, yet (e.g. Schulmeister, 1996). Many Webbased training modules provide not more than some simple help pages or
132
Learning Support Systems
hypertext facilities with keyword indices. Help pages mostly provide information how to access functions but not how to apply them in different learning contexts - and why to apply them. Effective help systems should support the user to overcome his or her weaknesses and take advantage of her or his strength. We therefore designed a generic Web-based training entitled 'Learning in electronic media' as a introductory course module, case study and information repository for the entire course (cf. figure 33 and chapter 3 of this book). The WBTR provides four modules, a comprehensive glossary, a keyword and a fulltext index as well as a graphical overview with brief summaries of all modules. In our course, the WBT serves as a source for information about online learning and - at the same time - as an example for real life online learning in a company; all examples are adapted for international use. The international table of contents comprises the following topics: ® Learning with multimedia: Advantages and disadvantages of computer-based training - Appropriateness of multiple media Learning strategies for multimedia - Combining dynamic and static media - Self test 'Multimedia expert'. » Information from the Internet: Basics about the Internet Addresses in the Internet - Search engines and search strategies Self test 'Internet expert'. « Email and Computer Conferences: Basics about email - Writing emails - Mail and list server - Asynchronous and synchronous computer conferences - Video conferencing - Self test 'Email expert'. ® Learning strategies for CBT: Browsing hypertext and multimedia - Using navigational tools - Using bookmarks and note taking Graphical browsers, maps and overviews - Strategies for learning and re-learning - Self test 'CBT expert'.
133
New and Emerging Media
B-lBlM
HmilihOIIIUilrllBi'.HItiWIilUlil-WlHI• • I •^ .
«< ,
$
a
^
U
ji
$
A±*
&'
£dieS. a MS] I f|J "IpcctpoililHAvbt/'I'ii r=n Mim-s_t_sta I Uml
"33
J t o r i XJ v\rif
Overview (Glossary! Key words
^Westeebi.N
- ^ w , McM.c
jd Ultrashort-term memory for sen;ory impressions
'^Learning with neiru media p
=i
SApplying the five senses tor usef>
Short-term memory 1
Images
•SWh at does multi-median
2
» .
S Computers and Learning
4
^Media suitability
5
'i&The psychology of multimedj Texts
6 7
:>Th
s control 1
i>lnt rpreting the
P„Mp«.
^Discovering new perspect tom otter
&lm ges are easi Mo retai helpful giams
Click the "Next" button to go to the next image in the sequence.
S V i eos and Ani nation 0 Mult in edia expert <8-lnf imjti n from the In emet I5.E-ET
,il
Pievious topics
h f :hu 03 ' of t a ' i n < irn '..e- Q3 i m "11'
tr ^
:i
mj
t g j n t ^
Figure 33. A content screen of the Web-based training course 'Learning in electronic media' provides brief explanations, interactive exercises and psychological self assessment.
In order to motivate users to apply learning strategies we integrated about fifty so called brain tests: Each brain test consists of short psychological experiments which can be easily conducted within a few seconds and illustrate important features of human perception and human memory. Additionally, self tests for each module help the users to check their expertise on each topic and enables them to review directly the parts of the course which are linked to the questions.
134
Learning Support Systems
Course Structure and Objectives
Course Objectives The goals of the course are to provide students with a sound foundation of knowledge, skills and hands-on experiences which is needed to identify and successfully apply psychological factors in online and multimedia learning. Students explore the psychological factors, concepts and findings identified in the multimedia and distance education literature. They critically examine models, theories and observations of the field. Students will: » identify the basic psychological processes involved in multimedia learning. * outline the basics of sensation and perception in processing multiple media. * explore the application of psychological principles in multimedia and online learning. » experience the actual use of Web-Based Training (WBT) by reallife examples. * assess the usage principles and basic technologies of online learning. * explore the use and design of online learning and information systems. * analyze the basic skills and principles of Computer-MediatedCommunication (CMC). * analyze synchronous and asynchronous ways of collaborative online learning environments. The course consists of the following elements: Introduction * to the course set-up » of the participants to each other * to the New and Emerging Media course * to the sample Web-Based Training (WBT) course
New and Emerging Media
135
Module 1: Basic Processes « Students will: ® define multimedia in technical and psychological terms « explain interference caused by multiple media in learning « list at least three advantages and three disadvantages of multimedia learning » explain at least one media taxonomy * Topics: * Definitions of multimedia * Support and interference of learning processes using multimedia » Limitations of electronic media * Applications of electronic media « Foundations of instructional design ® Media taxonomies Module 2: Perception and Processing of Media * Students will: * List the elements of visual perception * Explain the terms 'picture superiority' and 'dual coding' * Explain how these concepts can be applied in multimedia learning ® Analyze at least one additional factor that is influencing individual multimedia learning * Topics: * Visualization of information * Ways of seeing * Basic processes in perception » Advances processes in perception ® Picture superiority effect * Assignment: * At the end of Module 2 students are to submit an essay with the title: 'Multimedia: Worth the effort?'. This essay has to cover at least the following topics:
Learning Support Systems
136
*
*
Advantages and disadvantages of multimedia learning, basic guidelines for the use of multimedia in Distance Education, general conclusion, and must not exceed 10 pages in print (double spaced). The grade for this essay will contribute 15 % to the final grade.
Module 3: Application of Multimedia * Students will: * Know about the discussion concerning media influence over learning ® Explain and analyze aspects that distinguish well formed and ill formed examples of educational graphics * Explain motivational factors in multimedia learning « Explain how educational video and animation can be applied in multimedia learning « Topics: » Applying visual information ® Special formats in visualization * Instructional video and animation * Applying animation and simulation * Self Test: ® Are you a multimedia expert? (Part of the basic Web-Based Training) Module 4: Using on-line Media Section A ® Students will: * Explain the basic components of individual online learning « Analyze learner needs and abilities underlying successful online learning * List and explain at least one technique which is not based on the Internet * Topics: » Basic skills for the Internet ® The global village ® Processing data
New and Emerging Media
137
» Formats of online information * Structuring online information Section B: * Students will: ® Analyze recent limitations of self-regulated online learning and information systems « List and analyze at least three major misconceptions of recent online learning systems « Explain how a lack of media competency and computer literacy can be compensated ® List at least three major trends in future online learning * Topics: * Searching and finding * Search engines and catalogues * Search strategies * Recognizing structures of information * Self Test: ® Are you an Internet expert? (Part of the basic Web-Based Training) « Assignment: * At the end of Module 4 students must have joined a learning team which has to submit an essay with the title: 'The impact of the Internet on Distance Education'. * This essay has to cover at least the following topics: Advantages and disadvantages of online learning in Distance Education, individual and organizational requirements for successful online Distance Education, general conclusion, and must not exceed 15 pages in print (double spaced). The grade for this essay will contribute 35 % to the final grade. Module 5: Computer-Mediated Communication (CMC) * Students will: <* Explain and define the terms 'mailing list', 'newsgroup', and 'nettiquette' » Analyze how the concept 'nettiquette' can be applied in online learning
138
Learning Support Systems
»
»
Explain the concepts of 'educational controlling' and 'educational planning' ® Explain and analyze at least one method how educational outputs can be measured in economic terms * Define the terms 'peer (supported) learning' and 'learning teams' ® Explain and analyze at least one example for the application of learning teams in online learning * Explain and analyze at least one critical factor of expert involvement in online learning « Explain how Web-Based Training (WBT) can be introduced as a means for Distance Education in a company Topics: 9 Basics of emails: Addressing the electronic mail « Using email software * Nettiquette * Communication skills for email * Educational controlling and planning * Asynchronous and synchronous communication » Chats « Newsgroups « Mailing lists ® Mailing archives » Expert involvement in online learning
Module 6 The Final Project * The last two weeks of the course allow the students to focus on the final assignment / project that contributes 50 % to the final grade. « At the end of Module 6 students must submit an individual essay with the title: 'Project proposal for the introduction of WebBased Training at the Forinstance Inc.'. * This essay has to cover at least the following topics: * Advantages and disadvantages of Web-Based Training (WBT) in companies
New and Emerging Media
139
® basic requirements for staff and organization » critical evaluation of the Web-Based Training 'Learning in Electronic Media' used in the course, suggest improvements of theWBT « draft of a roll-out plan of the WBT in a company Course Readings ® For each module we have provided two required and two to five optional readings. The required readings include most parts of the WBT 'Learning with Electronic Media'.
Course Development We want to point out some interesting developments within our NEMDE courses. We started with two parallel courses in 2000. The first author of this book and his colleague, Michael Romeis, tried to combine tutoring, mentoring, and the role of subject matter experts (SME). We offered a plain 1:1 translation of the WBT module 'Learning with Electronic Media' which had been used in German banks. Furthermore, we provided readings mainly based in our own papers and articles and a small case study of the 'Forinstance Inc.' seeking to introduce online learning as a final assignment. We did not do very well, and we got ratings from students below average mainly because we did not answer timely, and because we did not make clear the purpose of all the different materials. Additionally, the student had to make some research on their own in the Internet because the scope of the readings provided was too small. Some students did not understand why we went with a mixture of two group and one individual assignment. Thus, the first modifications we introduced were: The WBT was adapted for international use, some technical problems were fixed. At the same time, we stated clearly that the WBT is an example adapted from learning modules used in German banks. We extended the scope of the reading - mainly by replacing our own papers with more international sources. We added several optional readings and links to relevant sources in the Internet.
140
Learning Support Systems
The most important modification was the separation of tutoring (that is, moderating, timely responses to organizational and technical questions) and SMEs (that is, discussing and motivating the content issues). Michael and I took over the role of SME who are actively engaged in the forums but we left the mentoring and tutoring part to two experienced online tutors. These modifications resulted in much better ratings. We had, however, still some problems in motivating the different assignments. Therefore, we introduced the following extensions: In addition to a general text about group discussion, team work, and writing group essays we provided a short summary of real life group discussion from our last courses. In what follows, we would like to quote a very interesting 'letter to the instructors' which was very helpful to design and adapt our work to the students' demands. The conclusion of our students was that experiencing and overcoming difficulties in virtual teams was one of their strongest and most important learning experiences. This is the paper we provide to our students. Example why transparent communication is so essential for Computer-Mediated Communication (CMC) Dear students, the following incident took place in the last group: Two members joined the group quite late in the second exercise. Efforts of one group member to contact them and to integrate them -inside as well as outside the group - were not effective. Then one member asked the tutors for assistance to 'get the two new people on board with the project'. This caused a big deal of bad feeling within the group, because the two students mentioned felt insulted. What followed, was an intensive communication between the group unfortunately going on by e-mail and not via conference [note: the electronic bulletin board of the course] which would have enabled all students to have the chance to contribute to the problem equally. However, doing this communication by e-mail left many members unsatisfied —always having in mind not to know what exactly the group was discussing. Communication which is not transparent creates a feeling of insecurity within a group which does not help to establish a 'common ground', that you need for working on a project as a team.
New and Emerging Media
141
Especially in this highly sensitive situation, conferencing as a transparent activity might have calmed down people rather than using email. Team communication structures, as aimed to be achieved by the tutors, did not establish in this group. CMC demands a high level of discipline, personal commitment and flexibility. In contrast to F2F, where communication is a flow of verbal and non-verbal information, CMC is a asynchronous stream of information. Participants of the group have to establish social regulations by themselves which demands a certain amount of self-discipline [note: CMC = Computer-Mediated Communication; F2F = Face-to-Face or on site training]. The following messages have been written exactly as they are published here. We have only changed the name of the student. Dear [tutors' names], / had promised to you feedback on the group project after the course was completed. I hope you don't mind, and I thank you for indulging my response. The group project, while difficult, was very rewarding. I learned a great deal, and since I believe this was one of the objectives, it was a great success for me. My personal belief is that our group got off on a 'bad' start, and that our communications were almost doomed from the beginning. If I can refresh your memory, two of the four members checked into the process very late in the second exercise, causing me a certain amount of anxiety. My efforts to contact them were both inside and outside of class, and were not effective. I then asked for assistance in contacting these people and getting them on-board with the project. That request proved to start a great deal of badfeeling within my group. This was the 'flaming' you spoke of in conference. There was a certain amount of mail back and forth between myself and another member that I have not shared with the tutors. I don't think it is necessary. The communications were professional but uncomfortable. As I look back on the exercise I wonder how things might have gone differently. I had read the writing guide that you posted in our class, andfound it a very helpful document - it guided some on my initial group
142
Learning Support Systems
communications with regard to my potential contributions for the group work. I also believe I would have asked for assistance again, if met with the same circumstances. The one change I would suggest is making the class conferencing a required communication element. Now, I know that this is the ideal method, and folks like to use e-mail because it is easier, but if conferencing was the standard, then students might have different or clearer understanding/expectations. If I can explain. Our group's first project was simple, but also a but problematic with regard to communications (multi-media definition). A lot of folks wanted to communicate via e-mail rather than conferencing. We were asked to try and do the latter, as part of the course activity, and to encourage conferencing communication as a group and transparent activity. I remember this being met with a great deal of resistance in my group. The final decision, since the activity was so short was that both methods were used. This method applied to the longer-term project did not foster either transparency in the class or better or more effective communications within our group. It did the opposite. Use of e-mail by the students turned into a power struggle with the flaming you spoke of. What would happen if we, as students, had been required to communicate only through conferencing activities? I don't know, and I want to accept my full responsibility for being part of the e-mail problem. But I do believe people in my group felt as though it was 'OK' to e-mail rather than conference. Would this change anything or make a difference? I don't know. And the reality is, that the distasteful communications within my group did serve to teach me something. However at what expense? I have not spoken with the other group members about this, so have no idea how they have processed the experience (or even if they care to do so). I want to thank all the tutors and assistants for a very interesting and worthwhile class. I started in this class slowly (the break was too long, and I lost a bit of momentum). The content was a bit difficult for me to get a firm grasp of at first, but found I came to it through the readings provided.
New and Emerging Media
143
/ thank you for your instruction and assistance throughout the semester. My best wishes for a warm and relaxing summer, [Student's name] Our response to this message was: Dear [student's name], thank you wishing us a warm and relaxing summer. However, it will be your words that are more heart warming ('though the German spring is quite sunny, right now). Moreover, your letter provides us with very helpful and valuable feedback how to re-design and optimize our course. Thank you very much for that, again! We refrained from making online conferences mandatory because students like to use the more common email - and they like to 'sneak around the tutor' (besides: No one can be forced to 'publish' his or her thoughts). While this might be a good idea in 'normal' Distance Education and F2F training settings, it causes a lot of problems in CMC. From our point of view, it is hard to provide reasonable evidence for that simple but far reaching insight. With other words: Information can be shared among people, experience can only be obtained personally. Our little start-up assignment (definition of multimedia) was meant to spread that insight and to set up preliminary team communication structures. This does not always work, however. I am so sorry that it was your group that sufferedfrom that severe communication problems. And this is a lesson that we learned again: CMC demands much more self discipline, flexibility, and personal commitment than F2F communication. Why? Because onsite communication is a continuous flow of verbal and non-verbal information loaded with context. Therefore, (mostly) we observe an establishment of 'common ground' within a quite short period of time. But in such rather optimal contexts there are lots of communication problems and subject experts tend to speak different languages (e.g. ICT and training experts). The situation is worse as far as CMC is concerned because it is a deprived and asynchronous stream of information. Thus, the participants have to set up their social regulations by themselves and to stick to them by self discipline. This is harder to achieve the more participants think that their
144
Learning Support Systems
'seniority' or 'high management level' does not demand such self discipline [many students in the course were training / education professionals]. What is our conclusion? We shall stress the point that the conferences are mandatory, that email is not forbidden but also not encouraged. And we shall use your input anonymously for the next courses (ifyou allow us to do so) in order to give students a clear example of why and how transparent communication is so essential for CMC. We will try to develop a small 'case study' about that topic. If you are interested to receive it (as soon as it is ready), pis. give us some feedback. Best regards from [Tutors' names] We received the following response: Dear [tutors names] Please feel free to use any information that originated from me in your case study. I remember discussing the transparency issue with [student's name] early in the semester. It is an idea I had to get used to quickly in this program, the 'publishing of one's thoughts' using CMC. However, as we are all aware, e-mail is just as an effective 'publisher' of thoughts as CMC. Personally, I don't mind transparency because: 1) I don't mind being wrong - this always leads to additional information or knowledge. 2) It facilitates accountability. 3) It makes one have to be 'poetic' about their thoughts, or helps provide focus. You were so very right about the differences between information assimilation and experiential knowledge. What an experience this work was! I have to admit, my initial anxiety was just not knowing if 50% of the group was even out there ! Nothing ! I didn't have the non-verbal context you speak of that exists in flf communication. When communications came in, they were difficult, but folks were at least present. I needed a social structure to work within. When the structure finally came in the form of communication, the rest followed.
New and Emerging Media
145
We don't have any bodies in this context, so a non-participating student within a group assignment is VERY visible. A difficult or hurtful conversation doesn't need to be remembered, because it can be re-read. I think it is possible to maintain transparency in communications in CMC, but it does, as you indicate require sincere efforts. I think the rewards in the form of a social and collegia! networks, and shared information/experiences are worth the effort. Again, thank you for indulging my remarks. Regards, [student's name]
Assignments We extended the case study with more background information about the fictitious company including economic data, training schedules and the like. We made clearer than before that the aim of the final assignment is the (more or less) practical application of academic and theoretic knowledge to given context. We think that this can help to situate declarative knowledge acquired in our course. Moreover, we extended the scope of the optional readings and links; some of the required readings were updated or replaced. The following sections are brief fragments taken from the case study underlying the final assignment. This is how we introduced the case study of the final assignment: Dear class members, consider yourself to be a trainee at 'Forinstance Inc.'. You are supposed to support the personnel department to built-up an 'e-learning' environment. The company does not employ online learning, yet. Your boss, Patricia Churchland, is very enthusiastic about online learning. [•••]
As a trainee of 'Forinstance Inc.' you are in the lucky (and somewhat exciting) condition to be the only person who is an expert for onlinelearning. All you have got are the following papers and memos: (1) fact sheet about Forinstance Inc.' (2) memo about a meeting with Patricia and her boss (the CEO of the company)
146
Learning Support Systems
(3) overview of training delivered at Forinstance Inc. [...] Given facts about the fictitious company were: Forinstance Inc. is the premier software provider of Internet-based service process optimization solutions for the service economy. It delivers the only proven solution that enables real-time, world-wide collaboration ofpeople, projects, knowledge, and process throughout an organization's virtual service community. Every service organization within industry is under constant pressure to maximize revenue, minimize project labor cost, and increase customer satisfaction. With shortages of qualified people, rapidly changing technology, and the distributed nature of today's virtual workforces, the pressure becomes more acute. Few service organizations have the necessary tools to optimize their processes and collaborate effectively. Forinstance delivers the solution to these business problems through its two offerings: eServices and ServerNet. Together, the eServices application and ServerNet infrastructure provide an enterprise-wide solution that not only optimizes the business processes of service organizations, but also enables collaboration throughout an organization's virtual service community. [...] The fictitious memorandum about e-learning in the fictitious company stated: Topic: E-Learning at Forinstance From: Patricia Churchland (Vice President Admin.) To: Bruce Springsteen (CEO) We have got a number of applications for the position 'Educational Consultant'. The first project of the successful candidate will be an essay about the introduction of Web-Based Training and e-learning at Forinstance. We expect the following: Description where and how to acquire teachers (or online-tutors) and online learning material.
New and Emerging Media
147
What target groups in our company should be addressed first (sales, administration, executives etc.) Web-based teaching is recommended. Because of tight work schedules, it is impossible for our members of staff to come together for on site lectures, regularly. In addition, there will be a major loss of income or allowance if they are not working. Support for learning how to use the Web is needed. Our employees mostly have little or no experience in computer-based learning in general. A training program is required to familiarize them with the functions of computer and knowledge of the Web. (Can we employ the WBT 'Learning with New Media'? Our new consultant should give us an evaluation and a recommendation.) [•••]
Training data given for the fictitious company: Training at Forinstance Inc. The company employs workshop and computer-based training. The following training modules have been booked by the company, last year [in rectangular brackets: number of trainees/attendees]: (i) Personal development (on site Workshops) 1. It's about time (time management) [41]. 2. Lose the meeting blues (effective team meetings) [32]. 3. Preventing sexual harassment at the workplace [97]. (ii) Presentation skills (on site Workshops) 1. Are you really listening? (effective team discussions and conferences) [44]. 2. PowerSpeaking (effective presentation with PowerPoint slides) [12]. (Hi) Windows (Computer-Based Training) 1. Beginning Windows 98/2000: File management (Explorer) [128]. 2. Beginning Windows 98/2000: Modifying the environment [79]. 3. Beginning Windows 98/2000: Multi-tasking, Object linking & embedding (OLE) [33]. [...]'
148
Learning Support Systems
We felt that there was a need for even more motivation and guidance writing the essays - and at the same time students asked for more freedom and less strict time schedules. Therefore, we introduced the following items: After the first week (and after the pre-week) we asked all students to join their learning teams and write a very short 'test assignment' defining the term 'multimedia'. The assignment is not graded but all groups get feedback. That gives the students to opportunity to adapt to group assignments. Students tend to start asking for information about the final assignment (50% of the final grade) very early. We noticed that the last module about CMC was not looked after very well because the final assignment grabs almost all attention. Therefore, we distributed information and material concerning the final assignment throughout the entire course, and we made clear that CMC is an important topic to be able to do the final essay properly. We separated work, studying, and discussion by introducing 'focused discussions'. Our tutors supported reading the papers, understanding the WBT, writing the essays. We answered all questions concerning the content and tried to elicit discussions by statement, further material etc. But, we did not force all students to be active participants in all discussion forums or threads. There were two weeks announced to be 'focused discussions' where all students had to be active attendees of a discussion with a clear focus, aim, and scope.
Focused Discussions The discussion was opened by a brief paper written by our colleague, Caroline Haythornthwaite, Graduate School of Library and Information Science of the University of Illinois at Urbana-Champaign, entitled 'A social network study of the growth of community among distance learners': These preliminary results suggest that familiarity with the technology and with other class members requires an initial introductory phase, but that choices about partnerships made by the middle of the course lock
New and Emerging Media
149
individuals into their network position. This in turn affects who they communicate with and from whom they receive information. While it is necessary for students to work together, the restriction of the set of individuals with whom a class member interacts may adversely affect their exposure to different information and viewpoints. An early orientation to a subgroup may restrict their interactions, reducing their 'class' from a set of 14 to the size of the project team. While the same may be said of traditional face-to-face classes, on-campus students have more incidental opportunities for interaction, e.g., unscheduled meetings precipitated by co-location during classes, encounters in the hallways or at occasional lectures, and encounters and project work in other courses or on committees. Our questions were: Dear class members, as announced, here is the first open question. I am referring to the few lines from Haythornthwaite. She states that a study group 'in turn affects who they (the students) communicate with and from whom they receive information. While it is necessary for students to work together, the restriction of the set of individuals with whom a class member interacts may adversely affect their exposure to different information and viewpoints'. So what do you think: Has the work in your study groups positively or negatively influenced your work? Has the sub-group in fact restricted your interaction to the class? Please feel free to let your classmates know what you think about this. Many of the students' remarks have been very enlightening. We would like to quote three examples: (1) To start with, I usually prefer working on my own. However I have worked well in the groups I have been assigned to both in this course as well as in two previous courses. In this course, all the other members were new to me but I feel that we really worked well together and co-operated. Besides the conferences in the study groups area and
150
Learning Support Systems
the collaborative docs we had chat sessions and the amount of e-mails we sent to each other overfilled my mailbox. Regarding the question whether a study group keeps the discussion away from the whole class I must say that usually it does not, but in this course there wasn't too much interaction and so I think yes, in this case, the discussion was concentrated among the individual study groups. To be honest, while admitting that I learn a lot from other people's discussions I used to feel relieved not seeing any red stars (meaning that I had read everything). When there is a lot of interaction it is a bit tiring having to read all the contributions and having to answer. Especially when participation is graded, I feel that participation becomes a sort of competition, with students competing to have the greatest number of contributions. Do you think I'm somewhat strange??:) (2) I have been intrigued to read your comments, classmates, on how you each view this learning experience. I am currently enrolled in three online courses, and have to tell you that honestly, if more than one of the three had the level of discussion that I have in only one, I'd be drowning! While I appreciate the contributions of my learning cohort in each class, I am cognizant of the burden our posts put on each other, as well as the potential for additional enlightenment they offer us. In terms of our collaboration: this was the first course I've taken where the bulk of the interaction with my group was outside the learning system interface. While it served a specific purpose, it also made it difficult for the course tutors to support us, as they couldn't 'see' how/what we were doing. We chatted through the learning system once, and though there were difficulties, it was fun (for me, anyway) to experience a little of the humor that is sometimes missing from smaller, shorter email correspondence. I have enjoyed the opportunities in this class for different styles of communication than in other online courses. It's been interesting to see how it can all fit together in so many different ways!
New and Emerging Media
151
(3) I'm glad our tutors asked these questions because they bring up several issues that have been of concern. Some members of the class have already commented that this class is different from others they have taken online, particularly in that there is almost no discussion of the material we are studying taking place in the class conference. As I see it, until now there seems to have been no serious expectation set that individual members of the class should initiate discussion or share their thoughts. Since everyone leads busy lives and many members of this class are also taking other courses, it has simply not been apriority (or a necessity) to take the time to post comments in our course [...] The group I was in interacted a great deal during the module. We really came together as a team around the set task. In many respects our skills complemented one another's. We learned to rely on one another. As trust was built, information and thoughts flowed freely. This is the only online group I have participated in that has achieved this level of sharing. But it is also the only time in this program a group project in which I have participated has had consequences. There was a desire to do well and earn a good grade. In short, there was expectation and incentive. We were all motivated by the assignment and that made a crucial difference in the group's experience [...] The question I would like to raise is this: What can we as students of distance education learn from the experience of this course about what stimulates or does not stimulate student-to-student interaction at a distance? Here's my take on the silence in the class outside of the group activity. It has nothing to do with being locked into a 'network position.' It has to do on the one hand with lack of motivation to post and on the other hand with lack of expectation, encouragement, and structure for interaction in our virtual classroom. When there is only a topic space provided for commenting on readings with little or no theoretical framework or thesis presented as guidance to help us focus as a class on the relationship of the readings to the learning objectives for the module, there is no real context set for discussion. We're out there on our own. Add to that there is no consequence either for initiating discussion or not and the members of the class are busy with matters that have consequence, there is no reason to expect much student-to-student interaction. Now, thanks to the faculty's initiative in raising a
152
Learning Support Systems
provocative question and setting up a period for focused discussion, there is a much greater level of participation. Effective Tutoring The following graph (cf. figure 34) shows that the last course from fall 2001 shows up the best results in pedagogy, interaction and tutor overall. Regarding the changes we made to the course and the feedback we got from the students we support the following two theses. First, tutoring is done best by sharing the work between SMEs and welltrained online-tutors as long as there is no SME with all necessary tutoring skills and a huge amount of time to participate daily in the conferences. Program Avg. -Spring 2000 -Fall 2000 -Spring 2001 -Fall 2001
Figure 34. Comparative overview of course evaluations from spring 2000 to fall 2001 (ratings are ranging from 1 'poor' to 5 'very good') as compared to the overall average.
New and Emerging Media
153
Second, within the context of the online course curricula, a WBT can very well serve as a content provider and also as a medium to reflect personal needs in distance education. In our experience some students have to be encouraged very actively to reflect their own interaction with the learning platform and the WBT. Since we focus very clearly on this point when providing the material for the final assignment, the results of the students' increase significantly concerning the discussion of CMC. The conferences are a very important part of the online-course. But some students prefer focusing on the content more than on the thoughts of their classmates. For those students distance education could be a very convenient way to avoid the 'tyranny of the classroom'.
'i - F i a i i k f i i i t M i m l R o f l i t
/ Stwstei |
orr ^0 10
1 10 2003 •>4 10 CT3 A l j 'or d in
on
1 10 0C3
Danie! Fendrich Stefan Geppert Tanja Goldschmidt Thomas Hartzsch
Christian Hdibrijck
!•
!>•
r
i-
UnMDHt
\
n a
taas
D
CD
cJ -CD
DDDDQ Qiigl siisi
aaaaa acsQg
21103 acassa
3 >CD a
•CD
acnaa QDfleDHaaccDDa a£ODD a a a c a a a o e c a • jacaaa £] a a c u a s s t. L U a£3sEL EJ ta a liSDDDa'JDGDDa aaaaa mi
Figure 35. Tutoring interface with 'alert system' listing all events the tutors has to care for (deadlines, open questions etc.; left) and general activity map sorted by learning groups and group members (right).
154
Learning Support Systems
The technological problem which has to be solved in order to support tutoring processes is the collection and ordering of data concerning the tutoring process. Process-oriented systems, such as the LSS presented at the beginning of this chapter, enable system designers to retrieve those process data from the system which are best suited to support the work of tutors and subject matter experts alike: Tutors have to be aware of certain events which could hamper or even stop the learning process, such as deadlines, open questions or poor participation. Furthermore, tutors need special statistics and overviews such as number and length of forum messages or notes listed by topic, date, and student, and overviews of learning activities, such as individual time schedules and test results. Moreover, tutor oriented system support should provide fast and easy communication on a group or student level. For instance, efiport's LSS provides all data collections, lists and graphical overviews mentioned so far. In addition, the LSS provides communication features such as individual and group home page messages, (push) news channels, feedback on test items and exercises as well as one-to-one dialogue features which are exclusively used by tutors.
Part 3
Decision Support: Introduction
Many researchers aim to make their multimedia systems more adaptive and, therefore, more pedagogical. Expert systems and intelligent tutoring systems (ITS) adapt to the learner's demands, abilities, and knowledge, particularly in subjects conforming to formal logic. A growing number of adaptive computer programs contain media such as videos and photographs. Though there are no clear borders between expert systems, ITS, and other adaptive multimedia systems. Expert systems are mainly distinguished from ITS by two key characteristics: The knowledge base of an ITS attempts to model human knowledge, the knowledge base of an expert system does not; and Expert systems are not designed to support the learning process, because they do not explain their rules, knowledge or inferences. So far, artificial intelligence (AI) has been a disappointment. A recent academic study of AI startup companies comes to the conclusion: 'If it works, it's not AI.' The revenues of AI corporations over the last decade support this view. There are basically two types of AI. The original goal was to create machines that are intelligent in a human sense. A compromise solution has emerged that merely refers to programs that perform tasks once the exclusive domain of humans. This type of AI is referred to as computational intelligence (CI). As some researchers point out, the original AI vision was to mimic human intelligence in the same way the first flying machines mimicked birds. Just as today's airplanes and helicopters don't flap their wings to achieve lift, there could be other ways to achieve intelligence than simply
155
156
Learning Support Systems
mimicking human intelligence. Two basic assumptions support the use of artificial intelligence in education: The key concepts of learning and intelligence are not fully understood and clearly defined. A theory of learning must be developed that will lead to a scheme for measuring a systems' adaptability and learning capabilities. The critical lift for CI to fly will not come from the systems' intelligence as much as its ability to communicate with humans. Basic AI- and Cl-based technologies include: ® rules-based expert systems, « neural networks, * genetic algorithms, * fuzzy logic, and <» advanced statistical systems. Each has its advantages and disadvantages for education. The following chapters are presenting a system supporting career counseling and vocational guidance from a more technical and a more pedagogical point of view. Today, a broad spectrum of techniques and philosophies impedes progress in intelligent learning systems. There are promising results, however, supporting use of such systems in teaching mathematics and programming. There are few adaptive testing systems for career counseling employing AI or CI available. But other software products are available, such as the roughly 200 counseling software programs for self assessment, job hunting, and job retention. About 100 titles incorporate self-tests that give the user feedback regarding career alternatives and prospects for career satisfaction. The first chapter of this section, chapter 6, describes a multimedia program, which combines a vocational encyclopedia and a testing facility to foster adequate career decisions. The testing facility is designed to suggest the same careers which a given number of experts would have suggested, if presented with the same user's input. The vocational database includes imprecise data like expert ratings enabling the calculation of suggestions of career options. The most important group of users of software for vocational guidance are young adults, who are about to leave schools. The results of cluster analyses show that the
Decision Support
157
interests of students are poorly structured and are not compatible with experts' ratings. The test facility has been implemented on several CDROMs, a short quiz to identify occupational fields, and a wide range of surveys which are answered by letters. Young students participated in an experiment to investigate the understanding and acceptance of the information provided by the system. The results show that students are able to judge, whether careers match their individual interests or not. Furthermore, it was explored whether the system is able to reconstruct experts' ratings. The system shows a good performance in reconstructing the experts' data - except in the case of one academic career which was not described very clearly. In another study, the influence of the testing facility on recall of information and individual acceptance was tested. Acceptance and recall of information about career options is clearly enhanced when studying individualized materials compared to more general information. The next chapter describes two studies about the use of adaptive multimedia in vocational training and consulting. Multimedia could potentially facilitate learning processes: A great number of studies address specific effects of media like video and photography. It has been argued that multiple media does not influence learning, learning gains are due to instructional methods, and pictorial superiority has not been supported. Therefore, many researchers aim to make their multimedia systems more effective using 'intelligent' software technologies to adapt to the learner's demands, abilities and knowledge. A group of students participated in two concurrent experiments. The influence of video, photography and individualized testing on acceptance and recall of information provided by a multimedia encyclopedia about professions and educational programs was examined: Video has no measurable influence in both experiments. Photography supports recall of all facts related to the illustrated professions. Individually calculated career options enhance acceptance and facilitate recall of all facts related to the suggested jobs. In a field study, the individual validation of four media for vocational guidance, two multimedia applications and two products of printed matter were compared. Data analyses reveal that the students enjoyed using electronic media, but they rely on printed matter.
This page is intentionally left blank
Chapter 6
Implementing Expert Guidance: Expert Advisor for Vocational Guidance
Career Decision Making Many people feel a lack of competence in career decision making. A lot of programs provide help but most of them cover selected topics and are judged to be of poor quality (Katz, 1993; Bridges, 1989). We developed a multimedia program which combines a vocational encyclopedia and a testing facility to foster adequate career decisions. The testing facility is designed to maximize the probability to suggest careers which a given number of experts - based upon the same user input - would have suggested. The testing facility and the vocational encyclopedia have been implemented on microcomputers using various Windows operating systems, and on UNIX / Linux machines. Although knowledge based systems and intelligent tutoring system (ITS) are well established tools, there are hardly any implementations of systems for vocational diagnosing and counseling (Ueckert, 1995). Psychological testing procedures including computer-supported diagnosis are used to conduct aptitude tests in personnel selection (e.g. Ghiselli, 1973; Sweetland & Keyer, 1984; Funke, 1993), adaptive testing optimizing economy and performance of personality, aptitude, and ability tests (e.g. Cronbach & Gleser, 1965; Park & Tennyson, 1983; Weiss & Vale, 1987; Bennett, 1993), and decision analysis applied to management diagnostics (Nagel, 1993; Sonnenberg, 1993). Although there are numerous tests which check for individual interests (e.g. Todt, 1967; Irle & Allehoff, 1984), ability (ITB, 1988a+b, 159
160
Learning Support Systems
Deidesheimer Kreis, 1993), and aptitude (Fock & Engelbrecht, 1986), there are hardly any computer-based, psychological testing procedures addressing vocational guidance. This does not mean, however, that there are no software products available: Counseling software guides list 200 programs, approximately, that are designed to support self assessment, job finding, and job keeping (Walz, Bleuer & Maze, 1989; Katz, 1993). About 70 titles incorporate surveys or self-testing facilities in order to provide the user with information about his or her career alternatives and career satisfaction. Most of these program address restricted topics like nursing professions or writing successful job applications. Additionally, many programs are judged to be of a poor quality in terms of completeness, accuracy and counseling efficacy (Bridges, 1989). Despite several shortcomings career counseling software is considered to be useful: The software can offer both instruction and practice (e.g. interviewing for jobs). Some programs have quizzes or exercises allowing an evaluation of career maturity or job hunting skills. The program can access information from large databases of occupational fields or careers. A large group of software allows users to answer questions about likes and dislikes and match them to careers. Some of this programs take the repetitive work and tedious jobs away from the counselor, hence, enabling him or her to focus on interactive career counseling. The most common complaints regarding these programs are (Katz, 1993; Bridges, 1989): « Poor usability in terms of screen layout and functional design, « long response and searching times, • a lack of adaptability to the audience, ® too few or inadequate assessment questions, » and a lack of complete and up-to-date information. Therefore, we decided to develop a program to support career decision making and to provide a comprehensive collection of relevant and up-to-date vocational information.
Expert Guidance
161
Models for Vocational Guidance Recent tendencies of the job market force young adults to acquire a high qualification in order to get adequate jobs. Additionally, the transition from school to work life gets more difficult and protracted. Therefore, improvements in educational and psychological counseling are required (Buchtemann, Schupp & Soloff, 1993). A core concept of career counseling is the assessment of vocational interests and abilities. Interests play an important role in career decision processes. They are considered to be that part of the self concept that matches personal traits and job structures (Allehoff, 1985). Therefore, vocational interests are effects of cognitive decision processes when judging about objects and actions which are relevant for jobs. Bergmann (1992) found reliable correlation between the dominating personal traits and the profession. These findings are theoretically explained by person-environment models which suppose that people tend to reduce dissonance between personal and environmental conditions (Holland, 1985). Vocational Interests of Young Adults The most important group of users of software for vocational guidance are young adults, who are about to leave schools. An important factor of career decision making is the knowledge young adults can access to in order to come up with valid decisions. 426 pupils participated in two studies designed to identify structures within job related interests and aptitudes (Hasebrook & Gremm, 1996). The results of a cluster analysis of 156 items show that the interests of pupils from secondary schools are poor structured and do not fit the situation on the job market: Production, technical equipment, and the notion of 'dirty work' form one cluster, 'clean' and social professions form another cluster. Working in offices, e.g. commercial professions, are supposed to be between 'dirty' and 'clean' work. Students of high schools have even simpler images in mind when judging about jobs in crafts and industries: They simply divide in 'dirty' and 'clean' jobs, and they are much more fond of getting a clean job than students of secondary schools are. Academical jobs, however, were described in complex and highly interrelated patterns. Figure 36
162
Learning Support Systems
displays the results of the cluster analyses conducted with the data from high school students judging non-academical jobs. This structure is compared to the structure derived from the experts' ratings. (a) Students
All items
"dirty" (tec -inical)
technical
equip™.
"clean" (social)
Trading, selling, and organizing
Usin
9
m
bi
9
™
Working in an office
working with animals or plants
Care and counseling
(b) Experts Physically demanding work
Conctruction, Mining
Commercial and organizational work I
technical and electrical equipment
Office related work 1
office related work 2
Crafts and arts
Care and Counseling
Laboratory
Plants and animals
Figure 36. Results from two cluster analyses: (a) high school students judging about nonacademic professions (n=157), (b) experts' ratings of non-academic professions and educational programs (n=l 18).
Knowledge and System Engineering for Vocational Guidance Students are novices with respect to career decisions, because they have no experiences which help them to evaluate their own vocational orientation. Experts tend to be over-specific and cannot take a 'naive' point of view. Thus, an ideal system supporting vocational guidance should reduce and transform inputs from experts and students into profiles which can be compared automatically. Based upon this considerations a system to support career decision making can be
Expert Guidance
163
constituted by an interactive test or quiz to match interests and preferences with job characteristics, and an encyclopedia or database to inform the user about relevant job characteristics like tasks, educational programs, work load, income, and so on. Our decision support system is not designed to mimic the full range of an interpersonal counseling process. Rather, it shall provide the user with a variety of career options which match his or her personal interests and preferences. The knowledge engineering process covered five project phases: The relevant factors of the experts' reasoning about careers were identified using interviews and surveys. More than 800 German careers and educational programs, about 50 items like 'Is this job physically demanding?', and 118 interviews with experts were included in this first step. The data revealed from this investigations were reduced to dimensions that discriminate between careers by means of statistical measurements. We calculated Principal Component Analyses (PCA) which are standard procedures of multivariate statistics in order to identify main sources of variances in multi-dimensional data sets (Stevens, 1992). We identified seven factors for academical career and nine factors for non-academical careers. Thus, 16 factors are sufficient to calculate discriminating profiles for each career in our expert system. We wrote short statements like 'I would like to explain new products to customers' which were dedicated to describe one of our 16 factors. 12 Items per factor were written and reviewed by a team of eight vocational experts resulting in an item pool of 192 statements. 426 students filled in surveys about their state of information, their vocational interests and responded to our item pool. By means of a multiple regression procedure (Stevens, 1992) we identified those items which described a single factor most clearly. This step resulted in a reduced item pool of 80 statements, 5 items per factor. In the last step, we implemented a database containing relevant information about 800 German careers and educational programs: Task descriptions, special demands, agenda of educational program, income wages, prognostic data, and the job profile relying on the reduced expert data.
164
Learning Support Systems
The database serves a the basis of our expert advisory system. In particular, we aimed to meet the following conditions: 1. The probability should be maximized that our system generates the same list of careers which our 118 experts would have suggested, if presented with identical input data. 2. The implemented module should be small and fast - that is, the program should be small enough to be stored on a single floppy disc including basic data about 600 German professions and educational programs; additionally, system response times should not exceed 2 seconds. 3. The model should be able to incorporate information from our vocational databases like average income, prognostication of the job market etc. 4. The data of the model should be updated and maintained easily. This is a crucial point for all systems supporting career decisions, because vocational databases have to be updated and modified quite frequently. Calculating the Goodness of Fit The core component of all career guidance systems is to inform the user about the goodness of fit of his or her vocational interests and a variety of career options. Katz (1993) tries to give a feedback about goodness of fit by performing a multiplication of self-ratings and expert ratings (as shown in formula (1)).
Y
=
2-4
Self ' ^Expert
^
Self-ratings (Rsdj) refer to the variations of the importance attached to each dimension by the user, e.g. how much she or he likes team-oriented working. Expert ratings (RExpen) are judgments of experts about how important each dimension is in every-day work life, e.g. the proportion of time spent by team-oriented work. More sophisticated models rely on a measurement of the distance between self rating and expert rating. These
165
Expert Guidance
model mostly use the City-Block distance or other forms of Minkowski metrics as shown in formula (2).
8=
(a
**Self
C ^Expert
(2)
Parameter c has to be greater than or equal to 1; with c=\ the term calculates the City-Block distance and with c=2 the Euclidean distance is calculated. Simple distance models, such as described in (1) and (2), have several shortcomings: The functions are not monotone and the distances measurements are far from being equi-distant. There are no information about positive or negative matches but only about how much self and expert ratings fit. It is hard to introduce preference models, e.g. to take into account whether a certain item or dimension is judged to be more or less important compared to another item or dimension. The scores expressing distances between self ratings and expert ratings cannot be compared directly - that is, scores derived from multiplication like in (1) and Minkowki metrics like in (2) are probably not contained in comparable scales. One of the paradox consequences of these simple distance models is that the more conflicting interests the user has got the more careers are fitting the user's profile. Consider a user who states that she or he likes to work in an office and at the same time she or he likes to work open-air. According to Minkowski metrics all careers are appropriate that can be carried out either in an office or open-air. A more adequate response would be to suggest careers that fulfill both conditions: partly to work in an office and partly to work open-air, e.g. security staff in big companies. Matching Careers to Individual Interests As mentioned before, there were no systems at hand based on classical AI methods to guide or implementation process (cf. Ueckert, 1995). We reviewed rule-based expert systems, neural networks, genetic algorithms,
166
Learning Support Systems
fuzzy logic, and advanced statistical systems in order to check their appropriateness for our purposes. Neural networks are superior to many statistical methods because they provide an easy way for non-linear forecast. They are effective in recognizing patterns in noisy or incomplete data. Therefore, neural networks are suited for career counseling where clear rules cannot be formulated (Weiss & Kulikowski, 1991). Unfortunately, it is impossible to explain the reasoning of neural networks to users. Explanation of the underlying reasoning, however, is a crucial point in career counseling. Genetic algorithms are successful in searching huge databases and large optimization problems including time and job-shop scheduling, and data-mining. Moreover, genetic algorithms can provide explanations of the decisions they produce (Davis, 1991). The performance of genetic algorithms, however, is strongly affected by the representations schemes employed. Additionally, setting the parameters such as mutation rate and crossover need extensive experimentation. This contradicts our goal to develop a algorithm which is simple to use and simple to maintain. One of the obvious advantages of fuzzy systems is their capability to deal with imprecise data using a rule-based knowledge base which is easy to understand and explain. This advantage, however, turns out to be a problem if clear rules can not be elicited (Cox, 1994). While interviewing 118 experts for career counseling, we learned from the interviews that the experts use rules to guide the counseling process, but they do not rely on any detectable rules when matching career options and personal traits. Thus, we had difficulties to translate their expertise into simple If-then-rules. Our statistical analysis of the expert data revealed that only very few statistically significant factors are discriminating hundreds of career options. We aimed to take advantage of this fact and reviewed statistical methods to match multi-dimensional data sets. We looked for an algorithm which allows easy explanation of reasoning, easy updating and maintenance, and incorporation of precise and imprecise data. Above all, no paradox results should be obtained by the system.
Expert Guidance
167
Implementation of the Expert Advisor We developed a system which embedded the following components: 1. An easy quiz or test about vocational interests and experiences. 2. Rating data and dimensions derived from a PCA (Principal Components Analysis; Stevens, 1992) of the expert ratings. 3. An algorithm based on a GLM (General Linear Model) to compare user and expert ratings. 4. A component to enter additional preferences and to modify the results of the GLM. 5. Finally, a report module generates an assorted list of suggested professions or careers. Although there are more comprehensive databases about educational programs in Germany than ours (e.g. KURS direkt, 1995), our database is the only one which includes imprecise data like expert ratings enabling the calculation of commendable career options. There are two ways to incorporate additional information into the calculation: Precise data like student fees and income wages are considered by looking up the appropriate data tables. Imprecise inputs like individual preferences are used to modify the distances measurements. Figure 37 gives an overview of all components of the testing facility. We decided to rely on distance measurement models from multivariate statistics providing information about strength and direction of matches, enabling easy modifications and transformations into appropriate scales. The most simple form to express a directed goodness of fit is to calculate a linear regression. We choose the covariance (formula (3) with u. being the arithmetical mean of the ratings R). P =~ Z
= " S
( O W - t*w)
(^Expert ~
i^Self • RExperl)~ PW
^xper^j
• /^Expert
P
r =
CTself
(3)
CJExpert
168
Learning Support Systems
USER'S INPUT: Yes/No Quiz
EXPERTS' RATINGS: Job Survey
Calculate Dimensions
Calculate Dimensions
Z
Actual User PROFILE
DATABASE Experts' Ratings
Compare Distances
Additional Preferences Income Security
S
DATABASE Income
Modify Distances
DATABASE Prognostication
Figure 37. Components of the testing facility supporting career decision making by matching individual interests and job characteristics.
This score expresses the distances of the ratings by means of the squared distances of individual ratings and the arithmetic mean and can be transformed into statistically comparable values. The covariance P can be adapted to the standard deviations of the ratings (aSeif and aExpert) - resulting in the linear correlation coefficient r - and can be fitted to the normal distribution by a Z-transformation (4) - resulting in the value Z (Stevens, 1992).
Expert Guidance
169
1 fl+r Z = --l nl in ( -
2
M_r
(4)
Z expresses an directed and normalized measurement of the linear fit of self ratings and expert ratings within a range from -1 (negative fit) to +1 (positive fit). As the Z-value is fitted to a normal curve, it can be used even in sophisticated statistical procedures (Stevens, 1992; Bennett, 1993). Profiles of single careers can be included in or excluded from the database without affecting the comparison algorithm. The system does not produce paradox results because conflicting interest ratings lead to a general decrease of the correlation coefficients. Furthermore, the coefficients can be easily adapted to individual importance ratings attached to different dimensions: In one implementation of our expert advisor the users could move sliders from 0 (low preference) to 100 (high preference) to express the importance of their ratings. Our system represents variations of the importance attached to each dimensions by an modified covariance formula (5) with A, being the importance rating (e.g. raging from 0 to 100): ( D
P="I
3
^k
^
• [R^er, ~ MExpert)
(5)
After having been modified the covariance can be used in statistical procedures like normal covariance coefficients. Finally, Z-values (see (4)) allow for an exact definition of the cut-off point separating career options that match individual interests from those that do not: We defined the peak of the Z-value distribution (the 'bell curve') to be the cut-off criterion. The peak can easily be calculated by comparing the differences between the Z-values of two neighboring careers: The peak of the distribution is reached when the preceding difference is smaller than it's successor.
170
Learning Support Systems
System Implementation and Product Development The model described in figure 37 has been implemented on UNIXcomputers using the programming language C because of its portability. This C-program was modified, extended, and incorporated in several products: A series of CD-ROMs for Windows-PC, a short quiz to identify occupational fields, and a wide range of surveys which are responded by carefully commented letters containing 10 to 15 pages each. We produced a series of CD-ROMs for microcomputers equipped with Windows operating systems. Each CD-ROM applied to a certain vocational sector, such as 'Economics and Law', 'Natural Sciences', and so on, and contains a vocational encyclopedia combined with the testing facility. In the beginning of a session the user enters his or her personal qualifications and conditions in order to allow the program to come up with reasonable suggestions and information. The user decides whether he or she wants to examine occupational fields, to explore the index of the encyclopedia, or to take a job quiz. The quiz provides the user with up to 80 simple yes/no assessments. After having responded to 20 items the system signals to the user that it is ready to suggest jobs and educational programs. If the user accesses the list of suggested items he or she may jump into the encyclopedia by a simple mouse click and retrieve information about qualifications, tasks, work load, income, prognoses, and so on. All occupational fields and many jobs are illustrated by videos and photos in order to elaborate the text and give an realistic impression of work life (Hasebrook & GraJM, 1995). Figure 38 displays a block diagram and a screen shot of the graphical user interface of the CD-ROM. A multimedia program containing our testing facility has been produced and published in charge of an international bank. The program contained a wide range of information: In the final version the program informed the user in short about 300 jobs and educational programs, psychological testing procedures, preparing job applications, studying in Europe and North America, how to improve decision making, and about 1500 relevant addresses. Furthermore, the program was equipped with a testing facility combining preferences concerning interests, income, and
Expert Guidance
171
job security: The user can take a quiz about his or her interests responding up to 60 questions. Two sliders enable the user to compare his or her interests (1) to the preference to get a high income and (2) the preference to get a high job security. The slider can be put into any position between '0' (less important than interests) to '100' (more important than interests). The user can choose between lists of suggested job fields, single jobs, or university studies. A short description can be accessed for each suggestion. A third version of our system had been implemented for a international chemical trust. The company wanted to present its educational programs and career opportunities in schools motivating students to start their career in the chemical industry. The program was used as a point-of-information with a sophisticated graphical user interface including maps, charts, photographs, and sounds. As there were hardly any menus and buttons, the screens contained lots of interactive elements: clicking on a map calls for information about regions and cities, clicking on a photo illustrating a certain profession accesses further information concerning the depicted profession. The testing facility consists of 28 questions focusing on the educational programs of the company. Each questions can be answered by moving sliders between T totally disagree' to T completely agree'. The system calculates up to five suggestions based on the direction and the strength of the user's valuation of the questions. Figure 38 displays a block diagram of the program. Evaluation of the Expert Advisor There are several psychological inventories dedicated to measure career development and vocational maturity. One goal was to make valid prognostications of the career development and the vocational success. Most of the studies, however, show essential validity for short-term predictions, only. Therefore, most investigations focus on enhancing construct validity. High construct validity, however, gives little support to career decision making and vocational guidance (Seifert, 1994). Therefore, we evaluated our testing facility for its practical validity - that
Learning Support Systems
172
is, how ...are the system's responses accepted by students and counseling experts. Forty-three students participated in an experiment to investigate the understanding and acceptance of the information provided by the system. Furthermore, we explored whether the system is able to reconstruct 38 experts9 ratings. ^"tg^fmmmmmm
Chemielaboianten arbeiten meist in Laboratonen def Induce. Sie fuhren Analyses, chemisuhe und phyeikalisch-chemische Uriteisuchungen duich. Ihte Tatiykeits&chwerpunkte smd das Vorbereiten und Dun-hfiihren vetschiedenster Analysen und Untetsuchurtgsarbeiten sowe das Protokollieien der durchgefuhrten Atbeiten und det errnittelten Ergebnisse.
:^^^^^^^^^^^RS Introduction
Encyclopedia (entries with texts, charts and photos)
Testing facility
Suggested professions and educational programs
Keyword search, alphabetical list etc.
Overviews and videos about work life etc.
tools, book marks, additional information etc.
Figure 38. Block diagram and screen shot of the CD-ROM "Career Counselor5 with Yes/No quiz, list of suggested professions, and information from a vocational encyclopedia (screen displays an entry of the encyclopedia).
Expert Guidance
173
Figure 39. Block diagram of a point-of-information (POI) about the educational programs of a chemical company.
We asked the students to rank four different job lists according to their judgment, whether the lists match their vocational interests or not. The students were told that the four lists were generated by four different computer programs. In fact, only one list was calculated by our computer system, three list contained random selections: A short random list with 6 suggestions, a long random list with 20 suggestions, and a list with a random selection of 12 popular jobs. The results show that students are able to judge, whether careers match their individual interests or not (cf. figure 40): They preferred individually calculated job lists compared to random lists (F[3,122]=l 1.8; p<.001; Eta2=.23). But they were not able to tell them apart from the list which consists of popular jobs which do
Learning Support Systems
174
not match their interests. Therefore, the students' judgments rely on weak criteria (sometimes) leading them to wrong conclusions. This assumption is confirmed by the data displayed in table 12: There is a positive correlation between the students' judgments about: how well the suggested jobs match their interests, how well they know the suggested jobs (based upon sound and valuable information), and how well they can imagine what typical professionals are doing. However, there is a negative correlation between all these variables and the actual state of information: The more information the students have got, the less they are willing to accept suggestions - and the less they have got a notion of knowing. Therefore, information leads to more skepticism and criticism. And skepticism may help to guide the career decision making process more carefully.
• Individually calculated Dwell known jobs m Random (short) m Random (long)
Rankl
Rank 2
Rank 3
Rank 4
Figure 40. Students prefer individually calculated job lists - but they cannot tell them apart from well known, generally preferred jobs, which do not match their interests (n=29).
More information makes students more critical and keen to get more information. This is the result of a study in which 156 high school students participated. The students filled in a survey about their actual
175
Expert Guidance
state of information and their need for information. Figure 41 displays the main results: Those students who are not engaged in information seeking (14%) are not very likely to change their attitude, only 32% of this small group want to collect more information. On the contrary, nearly 100% of the best informed group (19%) is fond of receiving more information about their career options.
Table 12. Correlation between questions concerning acceptance of the suggested jobs & information about jobs (n=43). Content of question: Match Know Imagine Information
Match
Know
Imagine
+ 0,56 + 0,42 -0,49
+ 0,58 -0,36
-0,34
Thirty-eight experts were instructed to respond to the yes/no assessments of our testing facility like ideal professionals from their point of view would do. The system shows a good performance in reconstructing the experts' data - except of one academical career which was not described very clearly: The system correctly identified between 85% and 100% of the experts' ratings (academical career 42%). Adjusted goodness of fit scores (AGF) of a LISREL model were between 0.81 and 0.92; AGF scores greater than 0.8 indicate a high goodness of fit (Joreskog & Sorbom, 1989). In a recent study, we tested the influence of the testing facility on recall of information and individual acceptance. The testing facility enabled the participants to enter their vocational interests and to receive a list of suggested jobs and educational programs from a multimedia encyclopedia. Additionally, all participants received the same list of jobs. The study has two parts: During the learning phase the participants read information about jobs and educational programs; during the testing phase, all participants completed two surveys.
176
Learning Support Systems
Figure 41. Students who have gathered more information are fond of getting even more: 100% of the 19% best informed students want to receive more information. However, only 32% of the students, who are currently not engaged in information seeking (14%), plan to do so in the future (n=156).
First, they rated the overall acceptance of the program, its functions, and its information; these results are summarized in table 13. Second, they completed a cued-recall task which consists of five questions about the job's description, income, distribution of age groups, usability indices, and unemployment rates; these results are displayed in table 14. The results confirm that individualized information effects acceptances ratings positively (F[l,73]=8.38; p<.01): The students considered all information about recommended jobs to be more interesting and valuable. Additionally, recall is clearly enhanced when studying individualized materials compared to general information (F[l,73]=13.9; p<.01; Eta2=.16): Students tend to recall more information about careers that match their individual interests.
Expert Guidance
177
Table 13. Acceptance ratings (O=complete acceptance, 25=complete rejection) as a function of list ofjobs (previously fixed list vs. individually generated list).
Tasks, motivating Tasks, valuable Income, motivating Income, valuable Prognoses, motivating Prognoses, valuable Summarized score
Job list Previously fixed Individually generated 2.2 2.5 2.8 2.9 2.5 2.6 2.6 2.7 2.5 2.7 2.7 2.8 15.3 16.2
Table 14. Cued recall scores (0=no recall; 25=complete recall) as a function of list of jobs (previously fixed list vs. individually generated list).
Recall job title Recall income Recall age groups Recall usability index Recall unemployment Summarized score
Job list Previously fixed Individually generated 2.2 2.7 1.4 2.7 1.4 1.8 1.2 2.0 1.4 2.3 7.6 11.5
Career and System Options Multimedia have potentials to enhance and facilitate career decision making. Most of the recent multimedia systems, however, show small positive effects or none at all. The effective use of multimedia is influenced by many internal and external factors, like motivation, knowledge, mode and contents of media, learning strategies, features of the task, etc. (Hasebrook, 1995). Making multimedia applications effective means to start from the user's perspective: Mostly, this implies to conduct a study about needs and abilities of the intended users (Hasebrook & Gremm, 1996; Glowalla & Hasebrook, 1995). Expert advises provided by the system clearly increases acceptance and performance of the system: Users pick up more information about career options that match their interests and they consider this
178
Learning Support Systems
information to be more valuable. The more information they have gathered and elaborated the more the loose their notion of knowing and develop a critical approach to expert advisory. Multimedia applications should not be designed to provide 'something for everyone', but they should provide exactly that piece of information which is needed in a particular state of the decision making process. Multivariate statistical methods can complement (or partly replace) AI methods where they are not fully applicable. Our system does not address the students' inability to make reasonable judgments as a whole but tries to motivate careful consideration of as many career options as possible. In order to achieve this goal system responses have to be valid in terms of the students' understanding and the experts' valuation. Three steps of the development of our testing facility ascertain its validity: » A comprehensive database containing precise and imprecise data, « examination of the items and testing procedures with the intended users, and ® applying an adequate algorithm which refers to expert knowledge. Recently, a German bank charged us to develop a new testing facility addressing career decision making amongst non-students, namely professionals who want to make a shift within their company (e.g. Goodwin & Hoppin, 1988). In the future, we will test whether our approach can be extended and applied to different target groups and career options.
Chapter 7
Implementing Adaptive Multimedia: Effects of Individualized Testing, Videos, and Photography on Acceptance and Recall
Media Effects There is great number of studies which address specific effects of media like video, photography and audio. Levie and Lentz (1982) compared 55 studies in which texts with and without illustrations were examined. No study showed negative effects. Most studies ascertained that there are positive effects on learning rates with an average increase of about 30%. Learning rates were increased, however, for the illustrated parts of the text, only. Pictures used for decorative purposes did not show any positive effects. Special images like statistical graphs were hard to understand without additional instruction or training. Levin, Anglin and Carney (1987) summarized 187 studies in their meta-analysis. They found that analogies, mental images and mnemo techniques can increase retention rates up to 50%. Once again, these results stand for the illustrated parts of the text, only. And: The texts must not be understood very easily without the help of pictures. Levin, Anglin and Carney (1987) did not find any positive effects of decorative or organizing pictures, either. Recent studies with electronic textbooks by Mayer and Anderson (1991, 1992), Drewniak (1993), and Rinck and Glowalla (1995) confirm the findings of the meta-analyses: Illustrations can potentially facilitate learning but the actual effect is focused on the illustrated parts of the text, and the effect very much depends on the
179
180
Learning Support Systems
content of the text and the picture, the user's ability and his or her motivation to understand the picture. Video has been considered to foster learning processes, because dual encoding of pictures and texts are supported (Paivio, 1986; Mayer, 1989), and because video provides a vividness which illustrated texts often lack (Livingstone, 1990; Brosius & Mundorf, 1990; Rolandelli, 1989). But Salomon (1984) showed that audio-visual media does not lead to better retention automatically: Children considered television to be 'easy' and printed matter to be 'tough'; therefore, they learned from television, only if they were instructed to prove how much they could learn watching television. Further research came up with additional conditions to make audio-visual media more effective: Motion pictures enhance comprehension, if they match the explanatory texts; there is no positive effect, if the pictures elicit strong emotions - e.g. showing violence and illness; using video presentations mostly does not facilitate learning, but switching presentation modes and media does (Brosius & Kayer, 1991; Brosius & Mundorf, 1990). Plowman (1994) points out that digital video can be used to incorporate easy to use explanations and guidelines into multimedia applications. Dynamic media like video and animation are often used to explain and visualize technical equipment or to support science instruction (e.g. Mayer & Anderson, 1991, 1992; Rieber, 1990, 1991). Recent literature does not indicate whether video works best as an advance organizer or a review; both approaches have been used successfully and no significant differences have been found so far (Calvert, Huston & Wright, 1987; Rice, Huston & Wright, 1986). In conclusion, video is not overly effective, but it can potentially enhance learning while visualizing technical and abstract systems, and while supporting vividness and elaboration of information (Escalada, Grabhorn & Zollman, 1996). Many researchers aim to make their multimedia systems more adaptive - and therefore more 'pedagogical' (e.g. Cox & Brna, 1995). Expert systems and Intelligent Tutoring Systems (ITS) adapt to the learner's demands, abilities and knowledge - especially in subjects which can be described in formal logic (Bastien, 1992). There is an increasing number of adaptive computer programs which are equipped with media like videos and photographs. Although there are no clear cut
Adaptive Multimedia
181
borders between expert systems, ITS, and other adaptive multimedia systems, Clancey (1987) divides expert system from ITS by two distinct features: 1. The knowledge base of an ITS tries to model human knowledge, an expert system does not. 2. Expert systems are not equipped to support learning processes, because they do not explain their rule or knowledge bases and the inferences drawn from that bases. As of today, a diverse spectrum of techniques, approaches and philosophies impede the progress in intelligent learning environments (Self, 1992). There are promising results, however, supporting positive effects of intelligent learning environments teaching mathematics and programming (e.g. Weber, 1995; McGraw, 1994). In general, effects of adaptation and system-controlled tutoring have been small or medium sized, yet (e.g. Rosenberg, 1990; Kelley, 1988). Career Counseling An important aim of career counseling is to broaden the spectrum and to elicit awareness of the many possible individual choices. For instance, more than 40% of the German students, who are about to leave school, focus on only ten professional careers, although the German dual vocational educational system provides nearly 300 vocational training programs and job profiles. Although knowledge based systems are well established tools, there are hardly any implementations of systems for vocational diagnosing and counseling (Ueckert, 1995). Psychological testing procedures including computer-supported diagnosis are used to conduct aptitude tests in personnel selection (e.g. Ghiselli, 1973; Sweetland & Keyer, 1984; Funke, 1993), to implement adaptive testing optimizing economy and performance of personality, aptitude, and ability tests (e.g. Cronbach & Gleser, 1965; Park & Tennyson, 1983; Bennett, 1993), and to perform decision analysis applied to management diagnostics (Nagel, 1993). There are numerous tests which check for individual interests (e.g. Todt, 1967; Irle & Allehoff, 1984), ability (ITB, 1988a+b), and aptitude (Fock & Engelbrecht, 1986). But there are hardly
182
Learning Support Systems
any computer-based, psychological testing procedures addressing vocational guidance. This does not mean, however, that there are no software products available: Counseling software guides list 200 programs, approximately, that are designed to support self assessment, job finding, and job keeping (Walz, Bleuer & Maze, 1989; Katz, 1993). About 100 titles incorporate surveys or self-testing facilities in order to provide the user with information about his or her career alternatives and prospects of career satisfaction. Based upon these considerations we have developed an adaptive testing facility to support career decision making by matching interests and preferences with job characteristics. This system has been described in full detail in the last chapter, which is chapter 6 of this book. The results of the testing facility do not restrict the users' problem space, but they focus on relevant information and help to work out the decision making processes - that is to use 'the computer as a tool for learning through reflection', as Collins and Brown (1988) put it. Additionally, a vocational encyclopedia consisting of eight CD-ROM was produced to inform the user about relevant job characteristics like tasks, work load, income, and so on. Each CD-ROM applies to a certain vocational training sector, such as 'Economics and Law', 'Natural Sciences', and 'Construction and Mining' (Hasebrook & Nathusius, 1997). While interviewing 118 German experts for career counseling, we learned from the interviews that the experts use rules to guide the counseling process, but they do not rely on any detectable rules when matching career options and personal traits. Thus, we had difficulties to translate their expertise into simple If-then-rules. Our statistical analysis of the expert data revealed that only very few statistically significant factors are discriminating hundreds of career options. We aimed to take advantage of this fact and reviewed statistical methods to match multidimensional data sets. We developed a system which embedded the following components (Hasebrook & Nathusius, 1997): « An easy quiz or test about vocational interests and experiences. « Rating data and dimensions of the expert ratings. * An algorithm to compare user and expert ratings. * A component to enter additional preferences. ® A report module generates an assorted list of suggestions.
Adaptive Multimedia
183
The testing facility provides the user with up to 100 simple yes/no assessments. After having responded to at least 20 items the system signals to the user that it is ready to suggest jobs and educational programs. If the user accesses the list of suggested items he or she may retrieve appropriate entries of the encyclopedia. All occupational fields and many jobs are illustrated by videos and photos in order to elaborate the text and give a realistic impression of work life. Figure 42 displays a screen shot from the CD-ROM which has been used to conduct the experiments reported in this article.
Figure 42. Screen shot of the CD-ROM 'Career Counselor5 with quiz and list of suggested professions (screen displays the quiz on the right and list of suggested items with photo on the left).
184
Learning Support Systems
Pilot Study Forty-three students participated in an experiment to investigate the understanding and acceptance of the information provided by the testing facility. We asked the students to rank four different job lists according to their judgment, whether the lists match their vocational interests or not. The students were told that the four lists were generated by four different computer programs. In fact, only one list was calculated by our computer system, three lists consisted of random selection. The results show that students are able to judge whether careers match their individual interests or not. They preferred individually calculated job lists compared to random lists (F[3,122]=ll,8; p<.001; Eta2=.23). But they were not able to tell them apart from the list which consists of popular jobs which do not match their interests. This data is displayed in figure 39 in chapter 6 of this book. There is a positive correlation between the students' judgments about how well the suggested jobs match their interests, how well they know the suggested jobs (based upon valid information), and how well they can imagine what typical professionals are doing. However, there is a negative correlation between all these variables and the actual state of information (cf. table 12 in chapter 6 of this book). The more information the students have got, the less they are willing to accept suggestions and the less they have got a notion of knowing. Experiment 2: Photo and Video
Participants of Experiment 2 Seventy-five male and female students from different secondary schools and high schools participated in this study. Ages ranged from 15 to 20 years (Mean =17). All students worked individually with the computer equipment. They were paid for their participation.
Adaptive Multimedia
185
Design of Experiment 2 We tested the influence of the multimedia components, like videos and photographs, on recall of information and individual acceptance by means of a 2x2 factorial design. The factor 'Video' was a betweensubjects variable (watching video before vs. after reading). The factor 'Photo' was a within-subjects variable (text with illustration vs. without). The subjects were randomly assigned to the two video conditions. The factor 'Photo' was counterbalanced by a Latin square procedure. The study had two parts: a learning phase and a testing phase. During the learning phase the participants read information about jobs and educational programs. During the testing phase, all participants completed three surveys: 1. They rated the overall acceptance of the program, it's functions, and it's information. 2. They completed a cued-recall task. 3. They filled in a survey about personal data, like age, school, and individual use of a Personal Computer. Materials and Procedure of Experiment 2 In the learning phase the CD-ROM 'Career Counselor: Construction and Mining' (cf. figure 42) was used to provide videos, photographs, statistical graphics and texts. The video lasted 2.5 minutes and displayed tasks, locations, and tools, which are typically used in the occupational field of mining. An explanatory text provides precise information about professions, tools and work places shown in the film. This video was shown either before reading or after reading any text. Each particular career was introduced by three texts. The first text described tasks, half of them were illustrated by a photograph depicting workers with adequate tools in a typical location. The second text described the income wages in the course of the career illustrated by a bar chart. The third text described prognostic data like unemployment rates, structure of age groups, and usability indices of educational programs on the job market - all rates and indices were illustrated by line drawings.
186
Learning Support Systems
It took the students about five minutes to read each text. All participants read descriptions of two careers - one illustrated by a photograph, one without photograph - resulting in a studying time of approximately 30 minutes. After having watched the video and having read the texts all students filled in a survey about their age, school type and their access to and actual use of a Personal Computer. Thereafter, they received a survey to rate the overall acceptance of the CD-ROM, the texts, and the pictures. They needed 10 minutes to complete both surveys. Finally, they were given a third survey which was not previously announced. The survey was a cued recall test consisting of six tasks: 1. Remember the exact name of the first profession. 2. Recall the exact name of the second profession. 3. Remember the income of both professions and indicate whether it was more, less, or near the average income. 4. Recall the unemployment rate of both professions and indicate whether it was a positive, negative or neutral indicator. 5. Recall the structure of age groups of both professions and indicate whether it was interpreted as a positive, negative or neutral sign. 6. Remember the usability index of both professions and indicate whether it was considered to be positive, negative, or neutral. It took the students about 20 minutes to complete the cued recall test. All in all, they needed 60 minutes to complete both the learning and the testing phase. Results of Experiment 2 Mixed MANOVA3 were calculated with the between-subjects factor 'Video' (displayed before reading vs. after reading) and the withinsubjects factor 'Photo' (text illustrated by a photograph vs. no illustration). All MANOVA were performed on summarized scores of
3
MANOVA is the abbreviation of multiple analysis of variance which is a special general linear model to analyze variations of the means of different distributions based on their variances.
187
Adaptive Multimedia
acceptance ratings (ranging from l=rejection to 30=agreement) and recall scores (ranging from 0=no recall to 25=total recall). Acceptance ratings. The nominal data seem to support a slight advantage for the acceptance of information displayed after having watched a video. The results of the MANOVA, however, show that there are no measurable effects of the factor 'Video' (F[l,73 ]=1.27; n.s.) and the factor 'Photo' (F[l,73]=.78; n.s.) on individual acceptance ratings. The results are displayed in table 15.
Table 15. Acceptance ratings as a function of video (displayed before vs. after reading) and photo (text illustrated by a photo vs. no illustration); scores range from 1 (rejection) to 5 (agreement). Photo Watching Video Tasks, motivating Tasks, valuable Income, motivating Income, valuable Prognosis, motivating Prognoses, valuable Summarized score
Yes Before 2.9 2.1 3.0 2.7 2.7 2.3 15.7
After 2.6 2.2 2.6 2.3 2.5 2.2 14.3
No Before 2.8 2.3 2.4 2.3 2.4 2.4 14.6
After 2.7 2.2 2.4 2.3 2.3 2.3 14.2
Cued recall. The data of the cued recall test are summarized in 3. The statistical analysis resulted in no effect of the factor 'Video' (F[l,73]=.54; n.s.) and a strong main effect of the factor 'Photo' (F[l,73]=17.07; p<.001), although the photograph illustrated only one out of three text portions. There was no significant interaction of the factors. We performed paired Wilcoxon tests for the single values of the acceptance ratings and the recall tests in order to identify differences within these variables as a function of the factors of our design. The data about income proved to be remembered better than all other information, regardless of all experimental factors (all comparisons p<.05 after alpha adjustment). Corresponding to this result, information about income and tasks were judged to be more interesting than the other information. In
188
Learning Support Systems
conclusion, the experimental factors 'Video' and 'Photo' did not influence single parameters but the whole set of variables - but the data about income were considered to be the most important and interesting information (all comparisons p<.05 after alpha adjustment). Figure 43 displays the summarized scores for acceptance ratings and cued recall as a function of the factors 'Photo' and 'Video'. Table 16. Cued recall as a function of video (displayed before vs. after reading) and photo (text illustrated by an photo vs. no illustration); scores range from 0 (no recall) to 5 (total recall). Photo Watching Video Recall job title Recall income Recall age groups Recall usability index Recall unemployment Summarized score
Yes Before 2.7 1.8 2.0 1.9 1.9 10.3
After 2.4 1.6 1.7 1.6 1.8 9.1
No Before 2.4 2.2 1.5 1.7 1.9 9.7
After 2.3 2.0 1.5 1.5 1.7 9.0
Co-variates 'usage of P C and 'usability of video'. We introduced the individual usage of a Personal Computer as a covariate to the MANOVA (ranging from 0 = 'no experience' to 5 = 'daily use > 30 min.'). Due to missing data and pair wise exclusion of cases the number of cases range from 69 and 75. We did not find significant influences on acceptance ratings (t[74]= -1.95; n.s.) and recall scores (t[74] =.19; n.s.). Additionally, we asked the students to judge the usability of the video about the occupational field (ranging from l=rejection to 5=agreement) and used these judgments as co-variates to re-calculate the MANOVA. There was no significant influence on recall scores (t[68]=.83; n.s.). But there was a clear influence of usability judgments on acceptance ratings (t[68]=4.12; p<-01) - mainly resulting in a decrease of the main effect of the (non-significant) factor 'Video' (F[l,65]=.07; n.s.).
189
Adaptive Multimedia
Summary of Experiment 2 Video has no measurable influence on acceptance ratings and recall scores in this study. Photographs used to illustrate texts describing the typical tasks of a job or profession showed no significant influence on acceptance ratings. But they clearly support recall of all facts related to the illustrated job. Although the photographs were not directly linked with information about income and prognostication. The re-analyses using co-variates reveal no additional effects - except the correlation between direct judgments about the (non-significant factor) video and overall acceptance ratings. Therefore, we conclude that there is a local effect of photographs on awareness and information processing, but no global effect on motivational and emotional judgments.
• With Photo • Without Photo
> < Figure 43. Acceptance ratings and cued recall as a function of 'Photo' (with illustration vs. no illustration) and 'Video' (displayed before vs. after reading); acceptance scores range from 1 (rejection) to 30 (agreement), recall scores range from 0 (no recall) to 25 (total recall).
190
Learning Support Systems
Experiment 3: Individual Information
Participants of Experiment 3 The same seventy-five male and female students as in experiment 1 participated in this study. All students worked individually with the computer equipment.
Design of Experiment 3 In this experiment we tested the influence of an adaptive testing facility and digital video on recall and acceptance by means of a 2x2 factorial design. The testing facility enabled the participants to enter their interests and to receive a list of suggested jobs and educational programs from the system (cf. figure 41). Additionally, all participants received a fixed list of jobs. The factor 'Job list' was a within-subjects variable (individually generated list vs. fixed list). The factor 'Video' was a between-subjects variable (watching video before learning vs. after learning). The factor 'Photo' could not be manipulated systematically, because only one third of the encyclopedia entries are linked to job descriptions containing photos and two thirds are not, but all entries could potentially be accessed from the individually generated lists. The subjects were randomly assigned to the two video conditions. The factor 'Job List' was counterbalanced by a Latin square procedure. During the learning phase the participants read information about jobs and educational programs; during the testing phase, all participants completed two surveys: 1. They rated the overall acceptance of the program, it's functions and it's information. 2. They completed a cued-recall task.
Adaptive Multimedia
191
Materials and Procedure of Experiment 3 We used the same materials and procedures as in experiment 3. Half of the participants started with the procedure described in experiment 1, the other half began with the procedure of experiment 3. All participants passed both experimental procedures. The survey asking for age, school type etc. was omitted, because it has been filled in already, thus reducing the time spent with experiment 2 to approximately 55 minutes. Results of Experiment 3 Mixed MANOVA were calculated with the between-subjects factor 'Video' (displayed before reading vs. after reading) and the withinsubjects factor 'Job List' (individually generated vs. fixed). All MANOVA were performed on summarized scores of acceptance ratings (ranging from l=rejection to 30=agreement) and recall scores (ranging from 0=no recall to 25=total recall). Acceptance ratings. As in experiment 2 there was no measurable effect of the factor 'Video' on acceptance (F[l,73]=1.80; n.s.). But there was a main effect of the factor 'Job List' (F[l,73]=8.38; p<.01). Table 17 displays all acceptance ratings as a function of both independent variables.
Table 17. Acceptance ratings as a function of list of jobs/educational programs (individually generated list vs. fixed list) and video (displayed before vs. after reading); scores range from 1 (rejection) to 5 (agreement).
Watching Video Tasks, motivating Tasks, valuable Income, motivating Income, valuable Prognosis, motivating Prognosis, valuable Summarized score
Job/Education Previously fixed Individually generated Before After Before 4fter 2.4 1.9 2.4 2.5 2.7 3.0 2.8 2.8 2.5 2.4 2.5 2.6 2.5 2.6 2.6 2.7 2.4 2.5 2.7 2.7 2.6 2.7 2.7 2.9 15.5 164 14.7 15.5
192
Learning Support Systems
Cued recall. Whether the students watched a digital video before or after learning, did not influence their performance in the cued recall test (F[l,73]=.28; n.s.). But the factor 'Job List' clearly influences learning (F[l,73]=13.86; p<.001) resulting in higher recall scores, if individually generated information was provided. The MANOVA reveals no interactions between the both factors. Table 18 summarizes the recall scores as a function of the factors 'Video' and 'Job List'.
Table 18. Cued recall as a function of list of jobs/educational programs (individually generated list vs. fixed list) and video (displayed before vs. after reading); scores range from 0 (no recall) to 5 (total recall).
Previously Watching Video Recall job title Recall income Recall age groups Recall usability index Recall unemployment Summarized score
Before 2.4 1.5 1.5 1.4 1.4 8_12
Job/Education fixed Individually generated After Before After 1.9 2.6 2.8 1.3 2.8 2.5 1.3 1.8 1.8 1.1 2.0 1.9 1.3 2.3 2.2 6J5 11.5 11.2
As in experiment 2, we performed paired Wilcoxon tests for all acceptance ratings and recall scores in order to identify differences within these variables as a function of the experimental factors. Once again, information about income proved to be remembered better than the other information. These findings were independent of the experimental factors. Information about income and tasks were judged to be more interesting than the other information (all comparisons: p<.05 after alpha adjustment). The conclusion is the same as in experiment 2: The experimental manipulations 'Video' and 'List of jobs' did not influence single parameters but the whole set of variables. Figure 44 depicts summarized acceptance ratings and recall scores as a function of the independent variables.
Adaptive Multimedia
193
Co-variates 'usage of P C , 'usability of video, and 'usability of job list'. As in experiment 2, we introduced usage of a Personal Computer as a covariate to the MANOVA. Due to missing data and pairwise exclusion of cases the number of subjects ranges from 62 and 75. We did not find significant influences of the variable on acceptance ratings (t[64]=1.83; n.s.) and recall scores (t[64]= -.12; n.s.). The usability of the video about the occupational field was used as a covariate to re-calculate the MANOVA, too. There was no significant influence on recall scores (t[65]=-1.59; n.s.), but a clear influence of usability judgments on acceptance ratings (t[65]=3.57; p<.01) resulting in a decrease of the main effect of the (non-significant) factor 'Video' (F[l,65]=.02; n.s.). In the first survey of experiment 3 all participants judged the usability of the career options - or jobs - suggested by the testing facility. The scores ranged from 1 (rejection) to 5 (acceptance). We used these judgments as another covariate in the MANOVA procedures. There was no influence of the usability of the testing facility on recall. But we found a significant influence on acceptance ratings (t[61]=2.78; p<.01) resulting in an increase of the main effect 'Job List' (F[l,63]=9.35; p<.01). Summary of Experiment 3 Video has no measurable influence on acceptance ratings and recall scores in both experiments. Individually calculated career options clearly enhance acceptance and support the recall of all facts related to the suggested professions. The re-analyses using covariates reveal no additional effects of usage of a Personal Computer and usability of the video. Taking into consideration the participants' judgments about the usability of the testing facility the influence of individual generated information on acceptance is clearly supported. Our conclusion is, that there is a global effect of individually generated information on acceptance, motivation and information processing.
Learning Support Systems
194
18 16 14 cu 12 oo w 10
CO
O)
8
r-B
•I •I
I• II I
II II I
CO O (0 CD O (n 0) O
-a o
D Fixed Job List • IndividualJob List
^B
o ti TJ O
•
•
J1
•
1 1 1 1 1
•B => •O
O
o a>
T3
•l
—'
> i Figure 44. Acceptance ratings and cued recall as a function of list of jobs/educational programs (individually generated list vs. fixed list) and video (displayed before vs. after reading); acceptance scores range from 1 (rejection) to 30 (agreement); recall scores range from 0 (no recall) to 25 (complete recall).
Field study: Comparing Electronic and Printed Media Electronic media for vocational guidance combine various advantages: access to huge amount of data, up-to-date information, and guidance provided by surveys or quizzes. This does not necessarily mean that students enjoy working with electronic media for providing vocational orientation. Therefore, we compared four media for vocational guidance in a recent study (Hasebrook & Wagner, 1998): two of them are multimedia applications and the other two products are printed matter. A. The multimedia application were: 1. a fully interactive point-of-information (POI) with fancy graphics, animation and sound and
195
Adaptive Multimedia
2. a more restricted computer-based training (CBT) containing guided tours, texts and digital video. B. The printed material was 1. a quiz with a graphical layout based on 26 exercises (Quiz) and 2. a comprehensive survey consisting of more than 100 yes/no items (Survey) which was responded to by an individually generated letter. We measured individual acceptance ratings after having used the four different products with 75 students participating in this study (between 15 and 18 years, Mean = 16). The results show that printed matter are preferred. This result is statistically independent of sex, education, and experience in using a computer. Thus, the students enjoyed using electronic media, but they rely on printed matter. Table 19 summarizes the mean acceptance and validation scores of the four compared products.
Table 19. Means of the subjective validation of two electronic media (1. Point-OfInformation, 2. Computer-Based Training) and printed matter (3. Quiz identifying occupational fields, 4. survey with response letter), scores range from 1 (rejection) to 5 (agreement).
l.POI 2. CBT 3. Quiz 4. Survey
like to work like 'look with and feel' medium 2.24 2.13 2.24 2.19 2.95 2.79 2.97 2M
Statement like recommendations 1.96 1.78 2.89 3M
product is would buy valuable product 2.02 1.96 2.87 3.04
1.99 2.03 2.75 3.05
The scores were compared by means of paired Wilcoxon tests; all reported results are significant at least on a level of p<.05 after having corrected the alpha error for repeated testing. Moreover, all linear correlation were calculated and transformed into Fisher-Z values in order to perform pair wise comparisons. The results show that the subjects of our study considered the printed survey to be more usable than electronic media (POI: z=4.5; CBT: z=4.8), they liked it (POI: z=3.9; CBT: z=3.5), they preferred to work with the survey (POI: z=3.4; CBT: z=3.4) and
196
Learning Support Systems
they recommended to buy it (POL z=4.4; CBT: z=4.6). The printed quiz was judged to be useful, but it was not rated as good as the survey: The students liked the quiz better than the POI program (z=3.7) and recommended it (z=3.7). The pair wise comparisons show that printed media score better than electronic media. In general, the participants of our study rated the survey to be the most valuable and the POI application to be the least valuable product. Summary of Field Study Teenagers are considered to have a general preference for electronic media. This study, however, shows that they are 'critical' users of computer-based programs: They enjoy watching videos and photos but they rely on comprehensive printed material which gives them the opportunity to discuss their career options with their parents and friends. It is important to note that we found the clear differences of the acceptance ratings of the four products only in the final judgment: When asked for a rating directly after having worked with the respective product, the students rated all products to be ',medium' or 'good'. Thus, subjects need clearly defined alternatives in order to come up with reasonable acceptance ratings. Mental Integration of Multiple Media It is possible that we were not able to catch subtle changes in acceptance using simple rating procedures. Therefore, we performed interviews with 15 individuals who had participated in both experiments: The students enjoyed watching the digital video, but they considered it not to be very useful in giving them an impression of work life: The average usability rating of the video was 2.1 (with l=no usability and 5=high usability). Our data supports other research about audio-visual media which reports only weak effects on learning, if the video is not precisely linked to particular pieces of information (e.g. Livingstone, 1990; Escalada, Grabhorn & Zollman, 1996). Video may provide advantages as an
Adaptive Multimedia
197
educational medium, if used as an advances organizer. But acceptance and recall clearly depends on the individual usability ratings of the video. We were somewhat surprised to find a positive influence of photographs on the recall of all information related to a career, although only one part was illustrated by a photograph. Furthermore, we did not find significant effects of videos and photographs on acceptance, nor any interactions of the experimental factors. Comparing our research findings with other findings reported in the literature we would recommend to use more 'simple and cheap media' like photographs which are definitely connected with particular pieces of information provided in the multimedia system (e.g. Levie, Carney & Anglin, 1987; Mayer & Anderson, 1991, 1992). Individually generated information clearly supported both acceptance and recall of vocational information. The effects were enhanced, if individual differences in usability ratings were taken into consideration. Our findings match other research results indicating that individually adapted information enhance motivational and learning processes within computer-supported learning environments (Bastien, 1992; Weber, 1995; Cox & Brna, 1995). The weak effects of audio-visual media, the strong emphasis on a close connection of pictorial and verbal information and the strong effects of individual relevant information were independent of subject variables like computer experience and usability judgments although the students generally show great interest in data about income. These findings can be seen in correspondence to the use of computer media by young adults: Young people tend to favor information with a high personal relevance and entertaining values (cf. Hasebrook & Wagner, 1997). The distribution of access to and actual use of computers in our sample suggests that it can be generalized for young Germans, at least: Our results match the data derived from other German studies: 14% used a computer more than 30 minutes a week, 21% worked with a computer daily, 53% once a week, and 12% never used a computer similar distribution has been found in other studies (e.g. Hoelscher, 1994; Lukesch, 1989). Future research should focus on the interaction of individually generated information and well designed packages of visual and verbal media. The interaction of dynamic media like videos and static media
198
Learning Support Systems
like photographs has to be examined more carefully in terms of internal and external resources spent to produce and to use them. Furthermore, the complex interactions of individual judgments about relevance and usability of information with acceptance and performance measures in testing and multimedia environments provide a rich research field which stands at its beginnings, yet.
Learning Environment Learner
First Recognition: Selection and Encoding
World Knowledge Text Base
Image Base
Recognition and Long Term Encoding
Meta-Cognition: Application of Elaboration and Inferencing
Mental Models Summarizing Situation Models
Dynamic Process Model
Figure 45. A model for the mental integration of multiple media (cf. Hasebrook, 1998).
Adaptive Multimedia
199
We would like to suggest a model for the integration of the empirical findings reported in the literature and found in our experiments: Initially, Multimedia information is encoded in simple text and image bases; using more sophisticated elaborating and inferencing processes mental and situational models can be generated based on the information in the text and image bases (Hasebrook, 1998). Information selection and encoding from short term memory leads to separated encoding of verbal and pictorial information in the long term memory (Baddely, 1990; Paivio, 1986). There is a tendency to understand pictures 'at a glance' resulting in a simple representation that is not linked to verbal information (Weidenmann, 1988, 1994). Deeper processing of images can be elicited by teaching appropriate learning techniques (Drewniak, 1992) and obvious links between pictures and verbal explanations. These higher levels of processing can help to generate appropriate static and dynamic mental models (Hegarty, 1992; Johnson-Laird, 1983). Many authors suggest that deeper understanding means that sequential verbal information is highly interconnected with analog pictorial information (e.g. Mayer & Anderson, 1991, 1992). Supporting understanding, then, demands the construction of semantically connected pieces of text and pictures, activating appropriate pre-knowledge, providing learning strategies for multimedia, and changes of media and learning perspectives to support the construction of comprehensive mental models (Albrecht & O'Brian, 1993). Research like our experiments and others (e.g. Mayer & Sims, 1994) support the consideration of individual differences in abilities and interests in order to enhance the understanding processes.
This page is intentionally left blank
Part 4
Self Learning Systems: Introduction
In 1950, Alan Turing's paper about 'computing machinery and intelligence' appeared in which he used the term 'learning machine' proposing to develop a 'child machine' filled with hereditary material, on which training is employed and 'natural selection' is simulated by the experimenters' choices. Two years later, first compilers were built and first programming languages began to shape, such as FORTRAN and COBOL. Although Turing's help to break German military code during the Second World War may well be decisive for the defense of the United Kingdom, he was arrested for homosexuality. Turing was unable to see the rise of Artificial Intelligence (AI) and programming languages, because he committed suicide two years after his arrest (Wardrip-Fruin & Montfort, 2003). Until the mid-1980s, AI researchers assumed that an intelligent system doing high-level reasoning was necessary to couple perception and action. In this traditional model, cognition mediates between perception and plans of action. Rodney Brooks (1999) proposed the opposite view by suggesting a behavior-based approach: The main assumption of the behavior-based approach of 'intelligent robotics' is the realization that the coupling of perception and action gives rise to intelligence and that cognition is only in the eye of the beholder. This approach formed the basis of considerable advances in the development of autonomous mobile robots and more humanoid robots such as Brook's 'Cog'. Classical machine learning is based on the assumption that human high-level reasoning can be and has to be artificially re-built in a 201
202
Learning Support Systems
bodiless, virtual environment. Therefore, learning has been defined as: Learning is a process which builds an accessible representation of interactions with its environment (Scott, 1983) - or quite similar as: Learning is the construction or modification of representations of experiences (Michalski, 1986). Consequently, aggregation (or clustering), characterization and symbol grounding, and classification and perception are the core elements of this way machine learning. The typical learning process is 'induction', that is, drawing a general inference based on specific cases such as: If all human beings I know have to die, all humans are mortal. Learning, then, is the search for predefined solutions using various information retrieval and detection techniques such as top-down and bottom-up search methods and reinforcement learning (Morik, 1993). These techniques have been introduced into Web-based environments mostly by using 'rational agents'which use the belief-desire-intention (BDI) model of rational action-defining agency where agents are autonomous (and mostly virtually mobile) software entities which apply different forms of reinforcement learning (Woolridge, 2000). Brookes (1999) pointed out that the connection of perception and action gives rise to intelligence. More general, we could state that there is no mind without a body, because pure symbol (or representation) transformation without novel experiences caused by external events in an external environment does not lead to any new information - and therefore to no learning (Keil-Slawik, 2003). However, simple machine learning based on direct coupling of perception and action cannot cope with complex environments where not all state changes can be foreseen or observed directly. All multi-agent platforms are 'complex environments' in this sense. Therefore, special team-partioned reinforcement learning techniques have been proposed (Stone, 2000). In what follows, we do not adhere to the classical approach of (machine) learning. Instead, we stick to the learning definition used in cognitive science which is based on observable behavior and has been given in chapter of this book: 'Learning is the acquisition of a relatively lasting change of behavior or the potential for it'. In the first chapter of this section we would like to show how computer science is in search for user interfaces that are human-like or
Self Learning Systems
203
may even outperform the cognitive skills of human beings. Up to now, the Web is far from being an ideal interface: Search engines deal only with about 30% of all Web pages, providing information at various and uncertain levels of correctness, broken links and up to 75% irrelevant information. Object-oriented multimedia databases which generate Web pages may technically control information in the Web. It is essential, however, that the users have the ability to select and to understand the information and to make proper use of the navigational tools. Scientific research has shown that multiple media are hardly influencing learning, but instructional methods do. Additionally, hypertexts mostly do not help to convey structural knowledge. Web technologies must provide interaction, meaningful settings and actively engage the user in order to enhance the learning process. Innovative user interfaces, such as real world simulations, Web ontologies and social engineering of the learning process help to support learning. Knowledge robots which mimic functions of the human communication and are based on recent results of brain studies are organizing knowledge according to human learning processes. Therefore, they provide a promising framework to build flexible and adaptive knowledge networks. Future research should focus on the complex interaction patterns involved in on-line learning and on the practical use of Web-based learning. Then, rules and systems for real men-computers dialogues will emerge in the near future which will be discussed in the second chapter of this section. Humans are not able to cope with the exponential growth of information and the increasing speed of information and business processes fostered by information and communication technologies. Technical support not only for information storage and retrieval but also for information selection, process planning, and decision support is needed. Moreover, the use of a (desktop) computer is restricted in many ways. In this paper, it is predicted that smart and mobile computing units embedded in a variety of things, such as TV sets and cars, will bring computing power close to their users. It is also predicted that users will get closer to computing power by using natural language and by using their social skills in computer mediated communication. A holistic architecture of knowledge robots (know bots) is described based on multi-agent platforms and distributed computational intelligence. Know
204
Learning Support Systems
bots consist of a self-learning artificial brain, speech recognition and synthesis, direct access to other software agents and computer programs, and direct connections to networks of human users. It is pointed out that a newly defined partnership between men and machine is a possible way to keep control of the exploding 'info verse'. Web-based multi-agent platforms mostly operate on semantically enriched online data which have been called the 'semantic Web'. The idea of the semantic Web was conceived by Tim Berners Lee (1999, 2001) proposing how agents are able to effectively conduct different operations from coordinating activities to keeping schedules based on the semantic Web. This generation of the Web is a collection of XML documents, of semi-structured databases, and of objects on the Web with rich semantics (Thuraisingham, 2003). The following two chapters examine Internet-based technical standards as well as the use of XMLbased technologies in the main application areas of learning support systems, namely content generation, delivery and searching, as well as self learning and adaptation capabilities. International, XML-based standards for describing e-learning products are being developed. Perhaps the most important standard is learning object meta-data (LOM). XML is an ideal standard for elearning. It is powerful, highly flexible, and supports modularization of data. Authors using XML can design courses incorporating a variety of multimedia and interactive elements. Most important, XML is powerful because it separates content from presentation style and, therefore, course content based on XML can be reused for a variety of purposes. Thanks to its structured meta-data, XML also facilitates searching for learning resources. Thus, XML learning objects can be used in any XML-enabled learning management system regardless of authoring tool. E-learning lacks industry wide standards. As a result, the industry is highly fragmented and tends to rely on vertical integration (such as mergers, acquisitions, and alliances among content providers, technology developers, and learning portals) as a strategy for creating de facto standards. Such standards often go through long periods of uncertainty. The Internet demonstrates the benefits of open, industry-wide standards. Ironically, it also demonstrates that specific interest groups can
Self Learning Systems
205
sometimes create standards that others end up embracing, while more grandiose, industry-wide efforts fall by the wayside (e.g., OSI standards). In general, there are three major components to most e-learning solutions: content, delivery vehicle, and student testing. Standards would permit any e-learning content to be delivered via any e-learning delivery vehicle, with testing performed using solutions from yet a third vendor. Thus, vendors would be able to reuse products (lowering development costs) and customers would be able to buy compatible products from competing vendors. Several groups are working on comprehensive learning specifications. These include: 1. the Aviation Industry CBT Committee (AICC), 2. EDUCAUSE's Instructional Management System (IMS) Project, 3. the U.S. government's Advanced Distributed Learning (ADL) initiative featuring the Shareable Courseware Object Reference Model (SCORM), 4. the European Commission's Alliance for Remote Instructional Authoring and Distribution Networks for Europe (ARIADNE), 5. the Institute of Electrical and Electronics Engineers (IEEE) Learning Technology Standards Committee (LTSC), 6. the HR XML Consortium defining a Human Resource Standard Exchange Protocol (HR-SEP) which has been mentioned in chapter 1 of this book, and 7. the Foundation for Intelligent Physical Agents (FIPA) defining a multi-agent platform. These organizations have been working for several years and it is likely that open, industry-wide standards will soon be proposed. It is also hoped that various standards addressing specific problems will be merged into the most comprehensive standard which is the shareable courseware object model (SCORM). There are several components to e-learning standards. One component will standardize the metadata that describes learning objects. Metadata is essentially information about information. E-learning metadata could include information such as the author, subject, last update, length, time required to study, prerequisites, relation to other learning objects, etc. The standardization of metadata is necessary for indexing learning objects for rapid search and retrieval. E-learning
206
Learning Support Systems
content could also benefit from standardization. A comparison of elearning content from various vendors reveals vast differences in quality, depth, degree of interactivity, and amount of communication with the instructor. Content standards would make it easier for students and instructors to identify and purchase the e-learning materials that best suit their needs. Extensible markup language (XML) is the next-generation language of the Web which will be a 'semantic Web'. In 1999, the HR-XML Consortium was founded to create first XML based human resource standards. Meta-data standards based on domain specific ontologies and expressed in XML standard schemas play a dominant role to built the semantic Web. The following two chapters, chapter 8 and chapter 9, are contributing to this discussion.
Chapter 8
Implementing Knowledge Structures: Searching the Web Without Losing One's Mind
Visionary Terabytes Nicholas Negroponte described a vision in his book 'Being digital' (1995), his dream of the ideal interface: 'Computers' that are human-like. Such a human-like interface would have the ability to communicate with men in a human manner. Negroponte is well aware that there is no computer that understands language as we do, up to now. Although in some restricted areas of language recognition some success has been achieved, no computer system incorporates something like 'meaning' or 'understanding'. Vilem Flusser, the European philosopher of the Electronic Age, envisioned a world where no human being is working but all men are engaged in solemn cogitation, being part of a networked humancomputer brain (see Flusser, 1985). Hans Moravec, director of the Robotics Institute of the Carnegie Mellon University, states that in the year 2050 the evolution of computer systems will outperform the natural evolution having the capability of more than thirty million instructions per second (MIPS). Therefore, the human race will only survive as hybrid creatures, something between human being and computer virtuality (cf. Moravec, 1998). This technical evolution has already begun in the field of Artificial Intelligence and robotics. Another 'time
207
208
Learning Support Systems
line' of human-computer evolution has been proposed by Ray Kurzweil (1999) which will be discussed in the next chapter of this book. Reality Bites Each user of web search engines, such as Google, Hotbot, Alta Vista, Yahoo and Excite, experiences a reality that is far away from these far reaching visions: Relevant information is buried under thousands of irrelevant links - many of the links are not valid any more and information in the web tends to be of low quality. The situation is even worse, as far as content providers are concerned: As most content can only be found and accessed by world-wide search engines, the content provider can hardly influence at what time, to what extent and in what indexing mode his or her content will be provided by the search engines, that is, the broker governs the provider. Statistics from different sources like 'searchenginewatch.com' indicate that even leading search engines, such as Google and Alta Vista, do not index more than 30 to 40 percent of all web pages - at various and uncertain levels of correctness and not always up-to-date. Whereas the Hypertext Transfer Protocol (http) based on TCP/IP indicates important progress in computer networking and routing, the Hypertext Markup Language (HTML) stands for the stone age of structured data modeling. Fixed links and HTML references elicit inconsistent and incomplete document and link structures. Even basics of data storage, such as normalization of data including data about links, are an error '404': They simply do not exist! Some advances, such as new (Dynamic) HTML versions have been achieved but the Hypertext Markup Language (HTML) is still far behind is ancestor SGML (Structured General Markup Language) and its general online format XML (extensible Markup Language). Therefore, it is not too surprising that maintaining large web severs leads to serious problems. A first solution is: Do not use HTML, use a database, which generates HMTL. Up-to-date web servers, such as the Hyperwave server, which was developed at the Graz University of Technology, help to maintain huge amounts of linked pieces of information easily. Basic concepts are: The use of object oriented
Knowledge Structures
209
databases in a multi-user environment, automatic link generation and management, separation of structure, content and layout, a variety of navigational and managerial tools for the web master and tools to engage the user, such as annotations and flexible content views (Maurer, 1996; 1998). See www.hyperwave.com. We do not want to address other important problems, such as the storage and linkage of audio-visual media, which are facing important progresses through the definition of MPEG-4 and MPEG-7 (documents can be retrieved via Motion Picture Expert Group: http://www.chiariglione.org/mpeg/standards/mpeg-4/mpeg-4.htm). An adequate database and technical concepts for audio-visual media form the technological platforms for the Electronic Age. But: Technology solves only technological problems. The accessibility of information depends on the ability of the users to navigate through the knowledge space, to understand texts, images, videos, and to understand how to use the navigational tools. Recent trends in Web data management according to Thuraisingham (2003; pp. 229) are: 1. Web databases including applications of scalability, query languages, transaction models, issues of security and integrity, indexing and distribution. 2. Web architectures comprising scaleable infrastructures, components, integration of online transaction processing (OLTP) and message-oriented middleware (MOM, three-tier architecture). 3. Web Mining applied as mining for e-commerce regarding privacy, scalability, and interoperationability. 4. Web Metadata referring to XML extensions, such as SHOE (described in the next paragraphs), ontologies as well as foundations and policies for them.
210
Learning Support Systems
Figure 46 displays the connection of metadata, ontologies, and XML and RDF schemas (RDF = Resource Description Framework). Ontologies are defined as a specification of concepts to be used for expressing knowledge (Frikes, 1996; see wwwdb.stanford.edu/LIC/HPKBtalk/ for more information): Ontologies are an agreed upon way to express knowledge, that is, they are not distinguished by their format but by the role they play in building the semantic Web.
Meta-Data Concepts: - Learning objects - Skills - Competencies
Ontologies: - Learning objectives - Learning topics - Competence profiles
XML Schemas: -LOM -SCORM -HRXML
RDF Schemas: - Dublin Core
Figure 46. Connection of meta-data, ontologies, and XML and RDF schemas (cf. Thuraisingham, 2003; pg. 149).
The Myths of Multimedia and Hypermedia Most of the hypertext systems are implemented by computer scientists and technical staff. Therefore, they focus on the technological side of the site and not so much on the psychological and pedagogical aspects. This may be the reason, why several multimedia and hypermedia myths
Knowledge Structures
211
managed to survive more than fifty years of psychological and pedagogical media research. I consider the start of this kind of research to be the publication date of Edgar Dales' book 'Audiovisual methods in teaching' in 1946. Here are the three most popular miss-understandings: Myth 1: More Media Leads to More Learning As of today, empirical research has not been able to support the enthusiastic visions of multimedia. In one of their meta-analyses, ChenLin and James Kulik (1991) examined 248 research studies about computer-supported learning. 150 studies failed to show any significant effects. The other studies showed only a slight advantage of multimedia over textbooks or lectures: Error rates of simple retention tests were reduced between 5% to 15%, problem solving was hardly enhanced, and study time was reduced between 20% to 70%, with an average reduction of time about 30%. Considering all studies included into the metaanalysis, multimedia produced only a small effect (Hasebrook, 1995a). Although, multimedia seems to save some time and reduce simple learning errors, it has not been found to be very effective as a problem solving tool. Clark and Craig (1992) reviewed several meta-analysis about the efficacy of multimedia supported learning. They draw the following conclusions: 1. Multiple media are not the factors that influence learning, 2. the measured learning gains are most likely due to instructional methods, 3. the aspects of picture superiority and dual coding of texts and images have not been supported. Fortunately, however, there are also some promising studies showing that multimedia could potentially facilitate the learning processes. The Software Publishers Association (1995) reviewed the effect of instructional technologies in 133 school studies from 1990 to 1994. They stated that there were better test results, an increase in self-reliance, and a closer interaction between students and teachers. Similarly, Boettcher (1993) collected 101 success stories in higher education in his book. Many case studies support this general impression: For instance, we conducted a study comparing a digital TV broadcast with live interaction
212
Learning Support Systems
between students in four German cities and experts in the TV studio to classroom teaching. The learning groups supported by electronic media clearly outperformed classroom teaching (Hasebrook & Steffens, 1997). Thus, multimedia can help people to enhance communication, motivation, and self-efficacy. This, however, does not necessarily lead to better learning rates but it could potentially facilitate the every-day life in schools and universities. Myth 2: Hypertexts Convey Structural Knowledge Picking (1994) observed users of a hypertext stack about Jazz music while solving different tasks: To get a brief overview users stick to the paging facilities and the subject index; to perform a goal directed search they rely on key words and indices; only if the users are free to get an impression of the system, they use hypertext links more frequently (cf. figure 47).
Get insight
Get overview
Targeted search
Figure 47. Frequencies of access of program tools of a hypertext system as a function of learning task (Picking, 1994).
Knowledge Structures
213
Retterer (1991) tested whether the use of hypertext features leads to better understanding. He compared three conditions: The first group studied a written text, the second group red the same text on a computer screen, the third group studied with a hypertext, which contained links between that parts of the text where names and cities are mentioned and that parts where they were explained. Retterer (1991) found that learning with hypertext leads to the best results. Crain (1994) compared lectures, video, and hypertext in a course about public relations. She found video to be the worst learning condition when tested immediately after having finished the course. She found no differences, however, four weeks later. Many authors claim that hypertext studies convey different or contrary results, because study setting and user skills are not sufficiently regarded. Glowalla and Hasebrook (1995) conducted studies about the effect of user skills and study setting while using hypermedia: Navigational tools, such as hypertext links and maps, were used most frequently by skilled re-learners, but skilled learners preferred informational tools, such as a glossary and a table of contents. Many other studies have confirmed that hypertext enhances learning, only if the individual skills and - especially verbal - abilities match the demands of the learning task and the hypertext system (Reynolds & Danserau, 1990; Barba & Armstrong, 1992; Barba, 1993; Mayer & Anderson, 1992). Additionally, multimedia features, such as animations with audio, rely on the visual ability of the user: Visual literate persons profit from animations, illiterate learners do not (Mayer & Sims, 1994). In conclusion, it is necessary to teach users strategies and concepts to use a hypertext. Additionally, it is necessary to adapt the system to individual abilities and the overall learning environment (Lajoie & Deny, 1993). Myth 3: Web is Easy, Print is Tough As web-based training refers to multimedia and hypertext, it is clear from what has been said before that appropriate learning results will not be achieved easily. Salomon (1984) showed that audio-visual media does not lead to better retention automatically: Children considered television to be easy and printed matter to be tough; therefore, they learned from
214
Learning Support Systems
television, only if they were instructed to prove how much they could learn watching television. Therefore, it is important to activate and engage the learner into the knowledge building process. The learner, who is not engaged, does not learn: This is the lesson that Jonassen (1993) learned when testing several hypertext indexes and maps. Although he provided well structured hypertext links, maps and tables, the users were not able to grasp the main concepts and to transfer them to related fields. Only one group was superior to all other groups: They had used a hypertext generation tool, called LearningTool, that allowed them to develop their own hypertext map. How can the effectiveness of multimedia over any other form of learning be improved? There are three important factors: ® Interactivity, * communication, and * individualization (or adaptability). Therefore, learning is (almost) not facilitated by use of certain media or multimedia techniques. But multimedia can help the student to be selfmotivated and become an active learner. An enormous amount of information can be stored and accessed easily. Interactive systems can support the responsible use of electronic media and international communication, such as language learning when students from different countries communicate via e-mail or computer conferences. Computer applications can adapt to preferences, knowledge, and abilities of single students (Hasebrook, 1995). On-line databases can provide up-to-date information while books tend to be out-dated as soon as they are printed. Homework assignments, such as 'Read the next 50 pages until Monday', do not make a lot of sense anymore. Instead, students may be more motivated to measure the air pollution in their hometown to find out that it is higher in the center of the town than in the periphery. Carefully designed animation, feedback facilities, and simulations can help teachers overcome the weaknesses of study materials and to focus more on the learning and communication processes. The learning places of the future won't be dim places filled with computers and isolated students in front of the machines. There will be an intensive interaction and communication between teachers, students, and other learning places from all over the world. Newly designed seminar rooms and lecture halls
Knowledge Structures
215
underline this development (Keil-Slawik & Selke, 1998; Hiltz & Benbunan-Fich, 1997). Complexity of Models and Reality Most graphical user interfaces do not support the capabilities of our perceptual and cognitive systems. For instance, the popular three dimensional statistical charts cannot be inspected very quickly, because the human visual system focuses on small details or gives an overall impression: Certain details pop-out of the scene, a general figure, a Gestalt, is projected into a scene. But three dimensional statistical graphs, diagrams and hypertext maps, like moving maps and fisheye views, need a continuous flow between details and general figures which have to be linked to certain meanings. But our mind did not evolve in rectangular block worlds and meaningless linkage maps. Natural scenes support understanding, if the scene is appropriate for the meaning of the data displayed in the scene. Therefore, the popular three-dimensional views of web pages do not support selection and interpretation of information. One of the most comprehensive efforts was the development of dynamic, three-dimensional computer graphs which could be perceived like reality. In 1989, Jaron Lanier, CEO of VPL, coined the term 'Virtual Reality' (VR) to bring all of the virtual projects under a single rubric. The term therefore typically refers to three-dimensional realities implemented with stereo viewing goggles and reality gloves. The term refers to threedimensional realities implemented with stereo viewing goggles and reality gloves. VR is electronic simulations of environments experienced via head mounted eye goggles and wired clothing enabling the end user to interact in realistic three-dimensional situations.VR is an alternate world filled with computer-generated images that respond to human movements. These simulated environments are usually visited with the aid of an expensive data suit which features stereophonic video goggles and fiber-optic gloves (cf. Myers, 1998). In summary, VR can be described as the potential to produce three-dimensional, artificial environments based on computer generated graphics which fills more
216
Learning Support Systems
than 60% of the visual field in order to produce the impression of being part of that environment, and which differs from reality by being more selective and more restricted (cf. Hasebrook, 1998). Whereas VR is locking its user into an artificial world, 'Ubiquitous Computing' is embedding the computer into everyday life by enabling 'smart things' such as computerized shoes and suits, so called 'Wearables' (cf. Gershenfeld, 1999). Mark Weiser is the father of ubiquitous computing (1991). Ubiquitous computing is just now beginning: First were mainframes, each shared by lots of people; now we are in the personal computing era, person and machine staring uneasily at each other across the desktop. Next comes ubiquitous computing, or the age of calm technology, when technology recedes into the background of our lives. We shall discuss issues of ubiquitous computing in more detail in the epilogue of this book. The term 'Augmented Reality' (AR) was coined at Boeing in 1990 by the researcher Tom Caudell. AR is often seen as the opposite of VR because AR is not replacing the natural reality by an artificial one but 'augmenting' or extending it. Caudell and a colleague, David Mizell, were asked to come up with an alternative to the expensive diagrams and marking devices then used to guide workers on the factory floor. They proposed replacing the large plywood boards, which contained individually designed wiring instructions for each plane, with a headmounted apparatus that would display a plane's specific schematics through high-tech eyeware and project them onto multipurpose, reusable boards. Instead of reconfiguring each plywood board manually in each step of the manufacturing process, the customized wiring instructions would essentially be worn by the worker and altered quickly and efficiently through a computer system. We envision a 'Augmented Virtual Reality' (AVR) where VR models of hidden or lost environments, such internal parts of the human body or archaeological sites, are augmented by data from the semantic Web. A first draft of a possible AVR was the development of a 'hypertecture': The German Institute of New Media (INM) and the authors of this book were co-operating on the development of innovative web-based learning and presentation tools. One project of the INM was SkyLink, a model of the city of Frankfurt which contained the
Knowledge Structures
217
skyscrapers of the major banks. An avatar allowed to float around in the scene and to choose different content views, such as the cash flow between the banks. While a real building is always more complex than the architects plan, the usage of virtual models tends to be more complex than the plan. We aimed to connect VRML models and the interactivity of MUDs (Multi-user dungeons) in order to provide our students with a deeper insight of real (but partly invisible) processes (cf. figure 48 and SkyLink browser at: www.inm.de/people/bernhard/skylink.html). This hypertecture differs from the desktop metaphor by its immersion. Additionally, there have to be immediate system responses: If it takes more then 10 seconds to understand the basics features, the user moves on to the next site. Known architectural elements can be used and transformed to semantic symbols of information: For example, the skywalk is not meant to stroll on but is used as a semantic element promising some interesting events to go along with. A door like the sky station entrance has to be visible from far away and has to be big enough to float through it - even if the visitor is not well trained in using VRML environments. In this way, architectural elements become icons, and natural environments become virtual learning environments. SHOEs for Web Walkers Obviously, we have to use a proper technological base and to engage the user if learning, understanding and problem solving is the objective of information sharing in the Web. But we have to understand what 'understanding' means in order to come up with tools to support selection and interpretation of information. Understanding certainly refers to internal activities like to grasp the meaning of something. How has meaning brought to the web, so far? There are three prominent ways: ® Keyword subject indices. ® Catalogs painstakingly built by hand. * Private robots using ad-hoc methods to gather limited semantic information about pages, such as: 'Everyone with links to me'. It is easy to see the disadvantages of all three techniques. All recent searching and indexing techniques come with about 75% irrelevant items
218
Learning Support Systems
responding to a retrieval query. Projects from the Leland Stanford Junior University and the University of Maryland, for example, try to overcome the semantic barrier in the web information technology.
Figure 48. The sereenshot of the SkyLink project of the German Institute for New Media (INM) presents the architecture of the city of Frankfort/Germany as a graphical user interface, a hypertecture (dark spot in the center: SkyStation; floating dark strings: SkyWalk).
The University of Maryland developed the Knowledge Query and Manipulation Language (KQML), a language and protocol for exchanging information and knowledge. It is part of the ARPA Knowledge Sharing Effort which aims to develop techniques and methodology for building large-scale knowledge bases. KQML is both a message format and a message-handling protocol to support run-time knowledge sharing among robots. "A related approach is Simple HTML Ontology Extension (SHOE) which gives authors the ability to embed knowledge directly into HTML
Knowledge Structures
219
pages, making it simple for user-agents and robots to retrieve and store this knowledge. SHOE is a set of HTML tags that adds a knowledge markup syntax; that is, to enable the publisher to use HTML to directly classify the web pages and state the Web pages' semantic attributes in machine-readable form. A similar project but much more demanding approach is the ontology server of the University of Stanford (Gruber, 1993; http://www-ksl-svc.stanford.edu). Further projects are run at several institutes, such as the Massachusetts Institute of Technology (MIT). Following arguments from Thuraisingham (2003; pp. 233) we would like to argue that... 1. the Web will be the integration platform for all types of multimedia and knowledge management technologies which will be (mainly) based on XML schemas. 2. the data types and technologies have to be connected to online databases - once again using XML and XSL schemas and XMLQL as a query language. 3. the various applications of the semantic Web, such as knowledge management, e-commerce and e-business, will thrive only when the Web becomes a real integration platform. 4. much research is still needed to enable the Web to be this platform of integration - and, thus, all research about information and knowledge management technologies should be related to the Web. 5. ontologies and the semantic Web will be the driving forces to develop XML schemas and technologies - and these technologies will play the major role in the integration of the different technologies needed to built the semantic Web. Traveling Agents: Knowledge Robots Systems whose information processing structures are fully programmed are difficult to design for all but the simplest applications. Real-world environments call for systems that are able to modify their behavior by changing their information processing structures. Cognitive structures and processes, embodied in living systems, display many effective
220
Learning Support Systems
designs for biological intelligent networks. They are also a source of ideas for designing artificial intelligent agents. The INM and the authors of this book were implementing a knowbot scenario to enhance computer-supported learning and working in corporations. To fulfill the requirement that knowledge (explained in a certain language) can be learned by an intelligent user's interface, two strategies are near at hand: 1. Robots are built, which are like human beings (the vision of Hans Moravec) or 2. virtual agents are constructed within appropriate environments which will meet aspects of the function of human communication based on the establishing of verbal meaning. Strategy one is followed in the Real World Computing Program of the Japanese MITI (Ohsuga, 1992). Strategy two forms the basis for the Knowbotic-Interface-Project (KIP) of Gerd Doeben-Henisch from the INM and the first author of this book. We think, that the hypothesis is valid that sufficient isomorphy of the data-structure and the functions are enough to achieve interesting results (Doeben-Henisch, 1998). Within the context of the Knowbotic-Interface-Project the virtual agents are called knowledge robots or 'knowbots'. This term has been coined by Doeben-Henisch (1998) in order to establish a distinction between knowbots and the robots of the Real World Computing program, and to avoid the still diverse use of the term 'agent' within the context of AI. Crucial for the knowbots within the project's context is their ability to learn knowledge about the virtual world they live in. They must also be able to learn an intelligible language in relation to this knowledge. Since the exact functionality of human ability to learn and human usage of language is still impenetrable, all varieties of modeling experiences have a hypothetical characteristic. Neural networks used in computer linguistics and computational sciences lack many of the properties of biological brains, like permanent learning, forgetting, global and local flexibility of the neural structure and the neuro transmitter exchange. The design goal of the knowbots' neural networks was to incorporate all essential features of biological neurons known today. The basis of this neuron model are formed by an abstract model of the chemical processes that elicit the electrical potentials of the neuron membranes. In a first implementation we tested a well studied biological network: the conditioning of the eye blink reflex
Knowledge Structures
221
of a rabbit. The keys to classical conditioning are a general dynamic neural structure and automatic short-term changes in the structure of the network and the neurons' behavior. All these key features are needed to detect simultaneous signals and transform them into a learning rule. We shall describe this experiment and the implementation of knowbots in the next chapter. Educating Knowbots for Education The research on constructivism and situated cognition support the idea of new roles for teachers and interactive learning systems. The main ideas of constructivism and situated learning are: Learning is an active construction of knowledge instead of passive absorption of knowledge. Additionally, physical and social aspects of the learning situation have to be considered. Central theories of situated cognition led to corresponding instructional models such as 'anchored instruction' (overviews are provided in CTGV, 1990; Collins, Brown & Newman, 1989; Duffy & Jonassen, 1992; Gerstenmaier & Mandl, 1994). State of the art in pedagogical and cognitive research provides a number of key concepts: * Constructivism, * situated cognition and situated learning, Cognitive apprenticeship - especially in vocational training, » knowledge transposition, that is the transformation of expert knowledge in practical knowledge and in practical competencies, * enhancing the learners' motivations (although the role of intrinsic motivation seems to be overestimated in the recent literature), * definition of the teaching process as an active coaching and guidance process; » immediate feedback based on adequate evaluation criteria. When designing adaptive multimedia learning systems as described in chapter 6 and chapter 7 of this book, we reviewed different AI models, such as neural networks, genetic algorithms, fuzzy systems, and multivariant statistical analysis. In a study with seventy-five students, which was described in chapter 6 of this book, we were able to show that individualized information lead to better retention of relevant
222
Learning Support Systems
information than multimedia: Our subjects recalled 45% of the information after having used the testing facility but only 33% after having watched videos and photographs accompanying the explanatory texts. Statistical methods represent a static set of factors. Thus, it is impossible to set up a dialogue between human users and the computer system. Moreover, static systems do not learn from users interactions and do not adapt to the communicational needs of the users (Hasebrook & Nathusius, 1997). Knowbots may help to overcome such obstacles by representing relevant aspects of the users needs and knowledge. They may be able to provide individualized communication with the user and by communicating with each other - to optimize counseling and learning systems. We claim that only those elements and rules, which apply to recent results of brain studies, are highly plausible candidates for the modeling of flexible learning behavior. Long-term learning implies the coordination of many neurons and the establishment of inter-neural information exchange. The neuronal networks of the knowbots can possibly handle all these demands for successful learning. Knowbots, therefore, could lead - among other promising technologies - to a better understanding of the human learning process and a better support of human learners. Future research has to focus on the interaction of key aspects of learning. This implies that simple experimental settings which aim to study main effects, such as comparing two types of media, are not too promising. Furthermore, study settings have to cover a broad range of criteria reaching from basic brain research to practical needs in schools, universities and business. Relevant advances in networked multimedia computer environments will emerge, if recent results of basic research are applied explicitly and the evaluation criteria meet the demands of the practical use of computer environments (cf. Glowalla & Hasebrook, 1995). We want to conclude with two examples illustrating the kind of research we are focusing on. The first aim is to identify the structure of neural networks that enable knowbots to perform operant conditioning. Up to now, neurobiological studies have not been able to point out which network underlies operant behavior, although some candidates have been
Knowledge Structures
223
identified (Sheperd, 1994; 1998). Artificial brains which show the same behavior as biological brains help us to pinpoint promising neural structures and to examine - in a computer laboratory - how the brain manages to learn in an operant conditioning style. We are sure that more than recent visions will come true: Computers won't be like humans, but they will be partners in an on-going communication and learning process. Rules and systems for a real mencomputers dialogue will emerge in the near future. This will not result in a simple extension of the anthropomorphism that can be observed, if tools like hypermedia, virtual reality or Al-based dialogue system are used. Inter-connected computer systems will learn to support understanding in the human users, and they will bring meaning to the World Wide Web, if we understand, that meaning and representation are not separated but two sides of one coin (cf. Cummins, 1991).
This page is intentionally left blank
Chapter 9
Implementing Knowledge Robots: Knowledge Robots for Knowledge Workers
Entering the Infoverse Reasoning and simulation mechanisms of currently unthinkable complexity will take over the control of process planning and information exchange. Fourth generation robots with the capability of performing more than 30 million instructions per seconds (MIPS) will be the heart of a company's knowledge base. This is the vision propagated by Hans Moravec, Principal Research Scientist at the Robotics Institute and Director of the Mobile Robot Laboratory of Carnegie Mellon University, Pittsburgh (U.S.A.). The global economy gets accustomed to the idea of the 'new economy' where the knowledge workers' creativity and skills are the companies' most important capital and competitive advantage. If only parts of Moravec's vision come true, however, it will certainly mean that the relevance of human expertise and experience will diminish. Current developments seem to support this point of view: A supplier of computer storage systems reports that especially banks are consuming more storage space within six months than has been used during the last twenty years; the increasing speed of product innovation and life cycles depreciate technological knowledge and skills within one to three years. The 'infoverse' stored in the worldwide Internet starts to exceed the amount of information that has been stored in more than 60,000 years of human culture before: It has been estimated that in the years 1972 to 1980 more information has been collected than in the 2000 years before. 225
226
Learning Support Systems
Fifty years after the publication of the first Gutenberg bible about two million books had been published; today, more than 3000 books are published per day, more than one million per year. Some authors, therefore, are discussing the advent of the 'age of knowledge'. Others, however, argue that the Internet is not more than a gigantic heap of information garbage. Recent studies show that we are not able to remember more than one to two percent of all the information we perceive in the mass media, such as radio, TV, or newspapers (Kroeber-Riel, 1996). A single search engine covers not more than about twenty to thirty percent of the World Wide Web pages, meta-search services using more than one search engine comprise about fifty to sixty percent of the WWW pages. Even the best text searching and indexing techniques do not come up with more than 30 percent of relevant links or search results, that is, an optimal search process accesses a quarter of a half of the information in the Internet - and one or two percent of this information can be remembered. Thus, we have to state that we have lost control over all the information gathered in technical systems. Exponential growth of information, information access at light speed and the increasing speed of business processes and the decreasing value of human knowledge force to re-focus the development of ICT. Information accessibility is no longer the main concern, but navigation, orientation and selection of relevant information. As computers and robots provide us with incredible capabilities to process increasing amounts of data within decreasing periods of time, it seems clear that we can only master the self-made 'information overload' if we manage to enhance our skills: 'Automation is over', that is, there are no more gains in productivity simply by investing in ICT instead of people (cf. chapter 1 of this book). We have now ask for more 'augmentation' - the enhancement and support of cognitive abilities while using ICT. The key topics of this new level of CMC (computer mediated communication) is a mobile, ubiquitous and selective information access enabled by smart software agents based on multi-agent platforms using distributed computational intelligence. We are now at a turning point in the development of ICT usage where sustainable progresses can only be made if we are able to delegate basic functions of information retrieval,
Knowledge Robots
227
process planning and decision support to technical systems. We have to decide whether we want to become garbage collectors within heaps of information - or the human masters of smart agent systems which we do no longer fully understand. If It Works, It's not AI Up to now, the progresses of the so-called Artificial Intelligence (AI) have been disappointing. A recent study about the commercial success of AI startup companies comes to the conclusion: 'If it works, it's not AI'. This assumption has been reflected in the revenues of AI corporations during the last decades (cf. figure 49). The strong position of AI is to develop machines that are intelligent in a human way. The weak position of AI is to implement programs that can be viewed as 'partly intelligent' because they are able to perform actions that used to be dedicated to human workers. This mode of AI is now referred to as 'Computational Intelligence' (CI). Patricia Churchland (1996) pointed out that we are at a stage where the strong AI position tries to mimic human intelligence in the same way the first pioneers of flight tried to mimic the birds' way of flying. As no modern airplane or helicopter is flapping its wings, it is clear that solutions enabling flight are not relying on flapping wings but on a proper lift. So, what might be a way to lift the weak position of AI to a higher level? At the beginning of the year 2000, Knowbotic Systems Inc. Ltd. was founded by the German Institute of New Media (INM) and Bank Academy, a non-profit organization for vocational training carried by the German bank associations. The purpose of this company was to develop and to examine knowledge robots or 'knowbots' which help to exploit the knowledge capital of a company by facilitating information selection, planning and decision making. The mission of Knowbotic Systems relied on two basic assumptions: As long as key concepts, such as 'learning' and 'intelligence', are not fully understood and clearly defined, computers won't be intelligent learners. Therefore, a formal learning theory has to be deduced from recent theories and empirical studies in order to set up a virtual testing
228
Learning Support Systems
environment for knowbots which helps to measure their adaptability and to extend their learning capabilities. The critical lift of CI will not come if a system is intelligent in itself, but it comes from the human capability to communicate with such a system in a intelligent and social way. Thus, knowbots have to mimic critical and observable components of intelligent communication behavior in order to transfer the results of machine learning and machine reasoning to human users.
o o o
1200 1000 Q W 3 c
o
o
55
^ M
55
1986
1987
1988
1989
800
600 400 200 0
1983
El
1984
1985
1990
Figure 49. Approximate AI revenues (Philipps, MIT, 1999).
Neuroscience Aspect Phil Johnson-Laird introduced the idea of 'Mental Models' as a complex mental representation of reality in scientific psychology (1983). He stated that we cannot distinguish between our world knowledge mostly organized in 'mental models' and reality because our world knowledge is our entire world. Recent scientific psychology has made some progress in explaining the way how reality is constructed from audio-visual perception, how muscular activities are planned, and how attention is shifted from one subject to another (cf. Kosslyn & Koenig, 1995; Mehler & Franck, 1995; Lane, Nadel & Ahem, 2000). There is some evidence
Knowledge Robots
229
that conscious behavior is not some much about planning and reasoning but more about the freedom of choice to stop or to modify 'automatically' triggered behavioral programs. There is also some evidence that the different steps of perception and understanding can be identified in different phases of the 'visual and auditory paths' which are guiding the information through the brain (cf. Challis & Velichkovsky, 1999). Latest findings about the close interrelationship of cognition and emotion as well as the experiments about the understanding of metaphors seems to indicate that we cannot cognitively understand what has not be understood in a more basic and emotional way before (cf. Ortony, 1993). These findings rise some interesting research questions: Significant evidence for the psychological invention of 'mental models' are rare. Among the few supporting findings are experimental results which show that people are able to re-adjust their memory if they get new information which helps them to improve their mental models of a story (see Hasebrook, 1998). Moreover, we know from research that most information about the real world is not organized in 'Euclidian measures' but in symbolic distances or semantic episodes (cf. Wolf, Hasebrook & Rinck, 1999). Can experiments show an symbolic distance effect for emotions? How are emotions integrated into mental models? If different perception times (e.g. of paintings) indicate different levels of processing, can we detect those levels of processing and understanding in visual, auditory and olfactory information alike? Are there any levels of processing for emotions underlying understanding? Understanding poems or jokes demands a good understanding of metaphoric information and is therefore based on cognitive and emotional evaluation to the same extend (cf. Aebli, 1994). What brain activities are related to understanding vs. not understanding poems or jokes in different cultural aspects? Do the brain activities reveal information about the interplay of emotion and cognition? Additionally, it should be noted that brain damages and consequent failure of certain functions are still among the most important source of information for neurologists (cf. Ramachandran & Blakeslee, 1998). Many researchers, such as Michael Paradis, have pointed out that the extend and the form of aphasia strongly depends on the language and its structure (Paradis, 1977; Paradis et al. 1982; Minkowski, 1963).
230
Learning Support Systems
Research results have shown that different languages demand different cognitive resources. For instance, Russian and German have a quite close connection between letters and sound as compared to English and French; it has been reported that multi-lingual patients were able to read Hebrew texts but not German or Russian texts. Ideographical languages, such as Chinese and Japanese, show different problems within the same patient when reading the different token system 'kana' and 'kanji'. These and similar results can lead to the following research questions: Are the different forms of aphasia related to different patterns of brain activity? Can these patterns also be found in differently coded information, such as sequential audio or analogue pictures? Which processes of cognitive and emotional understanding are most vulnerable for degenerative diseases, which are for traumatic brain damages? Which problems to understand languages can be observed in multilingual patients with emotional disorders, such as major depression and anxiety disorders, which have an impact on brain activity - and are there the same problems and activities for different languages? 'Consciousness now!' demanded Francis Crick in his ground breaking book about 'The Scientific Search for the SouV in 1994. Kosslyn has compiled and reviewed most relevant facts revealed by the 'Neuroscience' - and he put it together to a congruent picture of the basic neural and psychological functions of the brain in his book about 'Wet Mind'. However, he decided to leave out the question of the existence of the 'soul'. Neurobiology has unearthed the most reliable and valid facts. However, little is known about their relevance for a phenomenon, such as the origin of consciousness (cf. Churchland & Sejnowski, 1996). Gerd Doeben-Henisch and the first author of this book examined basic learning functions, such as classical and operand conditioning, because some of the underlying neural circuits are well described. We could demonstrate that computer simulations based on the chemical (not the electrical) activities of the neural network can exactly describe and simulate the real behavior (Hasebrook, 1999). Moreover, those 'pulsed neural networks' are very powerful and generic structures to built 'virtual' functions and meta-structures (cf. Doeben-Henisch &
Knowledge Robots
231
Hasebrook, 1999; Hasebrook, Doeben-Henisch & Erasmus, 2001). Based on these findings, some research questions come into mind: If neurobiology is suggesting candidates for neural 'learning circuits', can there function be re-built and tested in computer simulations? If neural networks can be simulated, are we able to detect more plausible candidates for higher order learning circuits, such us operand conditioning, by using these computer models? Can we apply genetic algorithms to automatically breed plausible candidates for leaning circuits by producing artificial (computer simulated) 'brain tissue'? Maass (1997) classifies the existing neural network models according to the following three generations: First Generation: McCulloch-Pitts neurons, or perceptrons, and threshold gates are used to implement multi-layered perceptrons, Hopfield nets, and Boltzmann machines. They can compute boolean functions as well as digital functions. Second Generation: Applying activation functions with a continuous set of possible output values to a weighted sum of the inputs (e.g. sigmoid functions, linear saturated functions, piecewise exponential functions) enabled the implementation of feed forward and recurrent neural networks and networks of radial basics function units. These networks can compute boolean function, digital functions, and continuous functions with a compact domain. Moreover, they support learning algorithms based on gradient descent. Third Generation: Spiking neurons, such as integrate and fire neurons, form networks of spiking neurons in which timing of individual computational steps plays a key-role for the computation. Maass (1997) proves that the third generation model has at least the same computational power as neural nets from the first two generations of a similar size. Only a few neural networks underlying learning have been identified yet. As a first test case we have chosen a classical conditioning circuit and several candidates that might be responsible for active stimulus-response learning (operant conditioning) (Menzel, 1996). In first experiments we implemented a pulsed neural network consisting of spiking neurons which represents the eye blink reflex of a rabbit as shown in figure 50a. The network matches the neuro-psychological data
Learning Support Systems
232
qualitatively (see figure 50b): If the unconditioned stimulus (US = air flow) is given slightly before the conditioned stimulus (CS = sound), then the connection of the US and CS is learned in a few trials,. After several runs presenting CS without US the 'connection' between the stimuli is unlearned. It is re-established very quickly, if the CS and the US are presented simultaneously again (Menzel, 1996). This means, that not only the neuro-biological structure of brain cells can be simulated on a PC, but also basic learning behavior which perfectly matches empirical data. The spiking neurons used in this experiment were defined by a XML schema; here follows an example of such an XML description: <soma id=122 name='s'> <membrane id=123 name='m1'> "^neurotransmitter value='GABA7> <event value='false7> ... ... ... ...
For the first time in scientific history, educational programs and methods can be based on results of brain research. Computer simulations of the activities of certain neural circuits and structures which may be part of learning processes will help to better understand the human brain.
Knowledge Robots
233
Recent research has shown that the brains of infants are far more flexible and adaptable than neuroscientists believed only a few years ago: The human mind is optimized to solve problems in social and natural contexts, and not for fact learning and logical reasoning. Moreover, learning mainly is based on social learning and imitation (Bandura, 162;
Figure 50. a) Neural circuit responsible for the eye blink, reflex of a rabbit (above) and b) test environment for classical and operant conditioning experiments with pulsed neural networks implemented with Java (below; cf. Hasebrook, Erasmus & Doeben-Henisch, 2001; pg. 72).
234
Learning Support Systems
The German brain researcher Gerald Huethig has compiled insights about the critical failures and success factors in education which can be deduced directly from findings of recent neuroscience research (Gebauer & Huether, 2002): 1. Knowledge acquisition and learning has to be a desirable value and behavior within the family and the society, and it may not be replaced by different values such as 'fun and leisure'. 2. Children need many opportunities to actively explore and influence their environment, and they should not become passive consumers of a media-based environment (that is, a 'coach potato'). 3. Learners need sufficient freedom to be creative, innovative and to make their own experiences, thus, learning may not be restricted to directly measurable (economic) and short-term outcomes. 4. Children and adults do not learn if they are flooded with inputs, and if sensation and perception systems are over-stimulated by multiple media; learning needs a stimulating but protective and safe environment. 5. Learning does not occur if learners are prevented from making errors and from solving problems on their own; research has shown that even in ICT training learners did better when encouraged to make errors without receiving any further explanation as compared to participating in a fully enhanced but over-protective training program (Frese, Brodbeck, Zapf&Prumper, 1991). 6. Finally, children need individual care, protection and empathy because learning does not happen in an atmosphere of fear and danger, but only in a supportive and emotionally positive environment.
Information Science Aspect Computer and Information science could possibly play an important role in the archaeology of the human mind - very similar to some new findings in research which have been made possible by virtual reality simulations or re-calculation of astronomical constellations. Within the context of information science, Norbert Wiener (1961) identified the role of information in natural and human systems in a way that had never
Knowledge Robots
235
been recognized before. He developed the field of cybernetics, which deals with the guiding or governing of systems. Wiener suggested that many systems are determined by the feedback of information to a governing element of the system. Wiener demonstrated, for example, that when a person reaches for an object, it is done with continual visual and kinesthetic feedback of information, which is then used to guide the hand further. The hand does not just respond to a single impulse from the brain to 'grab'. Recent research indicates that the 'free will' of a person is the choice to stop an action, such as grabbing, but not to start it. Another work that had an electrifying impact of all was Claude Shannon's information theory (Shannon & Weaver, 1949). Shannon measured the amount of information going through a telephone wire. His theory was abstract, and seemingly applicable to many environments, including not only the technical but also human language and psychology. The limits of Shannon's theory for the human sciences ultimately became evident, but the legacy of a new, abstract sense of information as reducing uncertainty by measurable amounts, remained. Similarly, Noam Chomsky's theory of syntactic structures in language (1971) - common patterns underlying all different languages had an deep impact on several fields, and was inspiring the emerging field of psycholinguistics. Miller, Galanter and Pribram, three wellknown psychologists, wrote Plans and the Structure of Behavior (1960), which posited a common underlying structure to all, or virtually all, human behaviors. Finally, Gregory Bateson identified common underlying structures in learning, as well as meta-structures in communication that reference other communications. He dealt, thus, in many different ways with representations of representations. It is no accident that the cover of the 1972 paperback of his 'Steps to An Ecology of Mind" states: 'The new information sciences can lead to a new understanding of man' (Bateson, 1972). A new view of the founding works of information science provokes new research questions: The so called 'strong Artificial Intelligence' (AT) seeks to mimic human intelligence by reproducing its necessary and sufficient prerequisites. Moderate positions of the 'Computational Intelligence' (CI) are focusing on computational behavior which can be interpreted as 'intelligent' by an intelligent observer (see Hasebrook, 2000). What are
236
Learning Support Systems
the prerequisites for anthropomorphism: What features of hardware and software foster the illusion of being intelligent? Several kinds of so called 'self-learning automates' or algorithms exist; mostly they are only loosely tied to existing neural networks. Can we learn from nature to built more reliable and flexible self learning systems as a basis for a 'bio-engineering' of virtual intelligence? We know from research that even higher cognitive processes cannot be understood without their emotional (and sometimes social) context. Can this context be adequately modeled - and how does it influence 'self learning' systems? Robotics' research has shown that intelligent or adaptive behavior is based on a close interaction with the outside world. Moreover, the measure of learning or intelligence clearly depends on observable behavior corresponding to well defined learning tasks and environments. Our knowbot research, therefore, concentrates technical developments on interface technologies which facilitate the access to knowbots by human users. The most important way to communicate is speech. The knowbots we tested were equipped with a speech recognition and synthesis system: The speaker independent speech recognition was able to identify about fifty words in five different languages at a time. As the word recognition can be adapted according to the actual context, this small amount of words is sufficient to implement small navigational or command systems. The speech recognition unit may also be trained to understand a specific user and it is then capable to handle dictionaries of several hundreds or thousands of words. The speech synthesis can read any text, such as HTML pages, tables or documents. The user can choose between several 'speakers' with different pronunciation or intonation. Much work is still necessary in order to make speech synthesis and speech recognition 'human like' as has been envisioned by Negroponte (1995) in his already mentioned book 'Being digital', although the research field has developed rapidly as can be seen in the overview displayed in figure 51. Moreover, even traditional methods, such as statistical analysis of large bodies of text, have been revived, recently: The German researcher Franz-Josef Och, Information Science Institute (ISI) of the University of South California, could show that classical statistics applied in an innovative way can produce cross-references of
Knowledge Robots
237
any formerly unknown text if a large body of text of more than 3 million words can be analyzed (Och, 2003). Och is now producing statistical translation systems based on cross-referenced text bodies such as the Christian bible. He is in charge of the research organization .of the Pentagon, DARPA, in order to support the National Security Agency (NSA) to better analyze emails and other written messages in foreign, languages. Complexity of speeking style
Lata 19|i$y
,'':''•:•; Natural veBJnversation
Spontaneous speech Fluent speech Read speech
::;$j$^$V, Connected speecj
by vqi^KO
Isolated words
Complexity (and size) of vocabulary
Figure 51. Overview of the development of speech recognition technologies (cf. Rauterberg, 1999).
Human users, the users of information systems, visitors and creators of the infoverse, are the main 'component* of a knowbot's environment. Additionally, other knowbots or standardized software agents may also enrich the knowbot environment. For this purpose, Knowbotic Systems has developed one of few worldwide available multi-agent platforms based on the FIPA standard (FIPA = Foundation of Intelligent Physical Agents). The platform FATE (FIPA Agent Template) comprises templates or suits which allow programmers to convert nearly any
238
Learning Support Systems
computer program into a software agent, that is, the knowbot technology provides easy-to-use ways to introduce a large variety of programs into the virtual learning environment. FATE also allows to run several platforms on different Web sites. This enables knowbots and other agents to communicate, move or replicate themselves all over the World Wide Web. Once again, XML schemas are used to implement the agent communication (cf. Zhong, 2001). The Age of Intelligent Machines ,The age of intelligent machines' is the title of a (thought) provoking book published by Ray Kurzweil in 1990. Nine years later, he published a book entitled ,The age of spiritual machines' (Kurzweil, 1999) predicting a time line when computer will not only match but exceed human intelligence. Knowbots are one of the few holistic visions of a man-machine dialogue in its actual sense, dedicated to support humans where they need help to access and select information - and to learn from them. But knowbots are not the only development in this field. A new level of smart agents and self-learning machines will develop in the near future. Figure 52 summarizes some major developments which are expected in the near future. Among them are software agents, mobile computing, and speech control. Kurzweil (1999) is providing an even more far reaching vision of human-computer interaction. His time line starts in the year 2009 and stretches to the year 2099: 2009: Computers are embedded, translation telephones are commonly used. 2019: Computers are largely invisible and embedded, threedimensional VR displays are embedded in glasses, gesture recognition and two-way natural language comprehension are the common command languages, people begin to have relationships with automated personalities. 2029: Computers have high-bandwidth connection to the human brain, neural implants to enhance cognitive abilities are available, a
Knowledge Robots
239
discussion about the legal rights of computers grows as computers claim to be conscious. 2099: The merger of human and machine intelligence has started, most conscious entities (being humans, computers or both) do not have a permanent physical presence as the number of software-based humans exceeds those using ,wet minds'.
2004 Ubiquitous online learning in universities
2005 Online learning in schools (K12)
2006 Software agents to search and select information
Interactive commun- Interactive ities in the TV for big WWW audiences
2007 Increasing use of speech recognition and synthesis Central remote control station for 'intelligent buildings'
2008 Increasing use of electronic cash Broadband access to information
2009 Mobile computing and eCommerce
2010 Selflearning softw. agents 3D virtual reality
Increasing use of eCommerce Figure 52. Some major developments in interactive media in the next ten years according to a recent study of the Fraunhofer Gesellschaft (Institute for System Technology and Innovation Research).
His ideas to achieve all this astonishing progress, however, proves to be somewhat over-simplistic (Kurzweil, 1999; pg. 295): ,First, carefully state your problem [...] Next, analyze the logical contours of your problem recursively by searching through as many combinations of elements [as possible]. For the terminal leaves of this recursive expansion of possible solutions, evaluate them with a neural net. For the optimal topology ofyour neural net, determine this using an evolutionary [that is, genetic] algorithm.' Of course, this receipt can only applied to first and second generation neural networks - let alone the problem that all real learning experiences, which are involving ,detection of the new', are not captured by Kurzweil's receipt. We envision future developments in networked
240
Learning Support Systems
computing and distributed computational intelligence where the users are no longer forced to adapt to the computer. The computers will adapt to the human capabilities to perceive and process data. The communication between and with computers will adapt to the human way of communication, namely natural language. And computers will be accessible at any time from any point with any device, such as PDAs, laptops, or mobile phones. Computer networks will also become people networks, taking into account specific deficits and potentials of computers and humans. Up to now, many individuals and companies are fascinated by the potentials and the exponential growth of the Internet. We do not think that future generations will be too enthusiastic about slow networks, unstructured information heaps and poorly equipped online shopping malls. Smart computers will be part of our every-day life, will be part of houses, cars, TV sets, refrigerators, bags, and suits. As a matter of fact, many ordinary machines are based on so-called embedded systems, that is, a small specialized computer. So, the things start to become computational things - and they will be smart things in the future. Knowbots and other smart agent technologies will support work, leisure and even cultural or social entertainment. Computers in the form of smart things will make computational intelligence as ordinary as cars or TV sets. But if the computers get nearer to their users, at the same pace the humans will get nearer to the computers: Not individual human beings nor software agent platforms will be the masters of the infoverse, but a new form of man-machine-interaction will emerge.
Epilogue
Future Developments
This book started with a prologue discussing e-learning on a global scale and how digital media can link people to knowledge and wealth. The structure of the book is built around the three layers of digital media in education (Keil-Slawik, 2003): Infrastructure, special software, and selflearning systems; the four main sections of this book are: ® Management support (infrastructure layer), ® performance support (infrastructure and special system layer), « decision support systems (special system layer), and * self-learning systems (self-learning system layer). The four main sections and nine chapters of this book were focussing on single topics such as learning support and corporate organization. The epilogue of this book will lead us back to a broader perspective: The human being, the human culture, and the impact of digital media on both. Let's have glimpses into the future - the future of our personal lives and the learning society as a whole. E-Assisted E-Learning in 2010 Laptops are becoming lighter and lighter, hand-helds are getting more sophisticated, and the mobile phone is gaining increasingly powerful computing resources. Yet we believe this is very much the beginning of an era that will start around 2010. At that point we believe mobile phones will have turned into veritable computer- powerhouses. Let us call them eAssistants. They will be not much bigger than a credit card, with a fast processor, gigabytes of internal memory, a combination of mobilephone, computer, camera (still and video), global positioning system 241
242
Learning Support Systems
(GPS), a variety of sensors and in continuous connection with huge nonvolatile local storage and the then existing equivalence of the internet: continuous, since there will be no charge for connect time, and probably not even for data transfer: an unlimited (time and volume) GPRS access is currently offered in the US for US$ 29,90 per month! Most importantly, the PC of 2010 will not have: 1. It will not have a hard-disk: this fragile energy consuming device with rotating parts will be replaced by a version of the memory stick as we now use them in digital cameras, but with hundreds of Gigabyte capacity; 2. it will have no screen nor keyboard as we now have; and the much reduced energy required by this device will be provided by tiny fuelcells. Of all of the above, we believe that most readers might be startled by only two things: the stated absence of screen and keyboard. Let us first address the issue of screen. Presently, there are half a dozen technologies competing to replace the screen as we now have it. They include flexible screens that can be attached to your sleeves ('wearable screens'), projectors that create images wherever you want (even on uneven surfaces of any color), and specialized eye-glasses that replace the screen, just to mention three alternatives. Which of those technologies will 'win' we do not know, nor does it matter: what matters is that wherever you go you will have a more or less zero-weight high quality display at your disposal, connected to the small computer proper by a modernized version of Bluetooth, and via the computer to a huge archive of information locally and all the servers on the internet. Of all possible technologies we particularly fancy a certain version of eyeglasses: the electronics in the eyeglasses are in contact with the computer via Bluetooth. The computer delivers (if wanted stereo) sound to the side of the glasses, that transmit it directly to the ear-bones (thus, only the wearer can hear the signals); the computer transmits (moving or still) pictures (if wanted 3D) through little mirrors through the pupils of the eyes directly to the retinas; and a tiny camera in the middle of the glasses provides the computer with what the user sees, e.g. for gesture recognition.
Epilogue
243
Indeed, the eAssistant may also have additional sensors and I/O devices and is supported by powerful software (for sophisticated image processing of the pictures obtained by the camera) as was already mentioned in (Maurer, 2003). This will be discussed in more detail in the following sections. Let us now turn to the keyboard. First, alternative input techniques are already starting to emerge. Speech input is one of them, and is particularly attractive if 'speech that is not heard' is used (i.e. utterances with closed mouth), e.g. using microphones near the larynx. Second, techniques that use the movement of fingers, the head, or the body using tiny sensors are becoming realistic; third, by using the glasses with an integrated camera described above a 'virtual keyboard' can be made visible to the user, and the finger movements on that keyboard can be analysed by software that does image-processing of what the camera delivers. We are not trying to suggest in this chapter that any one of the technologies described above will take precedence over others but more to suggest that the screens, hard-disks and keyboards, as we know them today, will be obsolete within ten years, give or take a few years. The Wizard in the Glasses In light of what we have discussed above, it is possible to predict that the eAssistant might look similar to what is shown in figure 53 below. We want to emphasize once more that we do not necessarily believe in the 'eyeglass' version that is shown, but it is a helpful metaphor to convey the functionality we believe will be available. The computer proper O is not much larger than a credit card and has all the functionality describe earlier. It is connected wireless to the internet and to the eyeglasses plus necklace. The computer delivers on the side of the eyeglasses (if wanted stereo) sound © ; it delivers via tiny mirrors © into the eyeglasses (if desired 3D) visual multimedia material of whatever kind, such as text, pictures, animation, 3D models, movies, 3D movies, etc. This may be technically accomplished by projecting images through the pupils
244
Learning Support Systems
directly onto the retinas of the eyes, or by creating a virtual image in front of the eyes. ® represents a camera that has multiple uses: 1. the user can look through it (thus having infrared vision during the night, or macro vision or zoom when this is useful); 2. the user can transmit what is being seen to others (i.e. we have video-telephony, of course); finally, the pictures taken by the (still and movie) camera can be analysed by powerful image processing software.
Figure 53. The wizard in !he glasses: cAssislant and associated devices.
The camera has also a built-in compass, hence the eAssistant is not only aware of where the user is (because of the GPS system), but also in which direction the user is looking. © is a larynx-microphone that can pick up what is spoken by the user (even if done witii closed mouth: this takes a bit of practice on part of the user), and it also has a loudspeaker so that a conversation or audio outputs can be shared with others, even if those do not happen to have an eAssistant at this point in time.(If they
Epilogue
245
had, the audio information could be sent directly to their ears using the devices © mentioned earlier). © symbolizes a device that can detect different states of brain activity. At this point in time a very limited number of states can be detected (typically the intention to move the left arm can be distinguished from the intention to move the right arm) (Purtscheller et al., 2000), but it is foreseeable that a dozen or more states will be achievable, allowing to create input for the eAssistant by thinking only. © will also have integrated further sensors, typically for the detection of head-position and head-movement or speed of movement of user, but potentially also to measure physiological parameters of the user like body-temperature, pulse, skin conductivity, etc. or even environmental parameters like temperature, humidity, air quality, air pressure. If it is not self-evident the next section should convince readers that an eAssistant as described will indeed revolutionize our world. Note that each of the features and sensors described above has been implemented in some way or another. The 'only' thing that is missing is integration of all into one small unit. However, the assumption that this will happen is basic to some of the research we are seeing today, e.g. in (Maurer, Stubenrauch & Camhy, 2003) or (Lennon & Maurer, 2001). E-Assisted E-Learning in the Future
Virtual Keyboards Input of information using the keyboard or the mouse will be replaced to a large extent by other means, such as speech recognition, gesture recognition, employing a 'virtual keyboard' and other methods that still sound unorthodox today. The 'virtual keyboard' is a good example that shows how the various components of the eAssistant interact with each other. By a spoken command 'Create keyboard' the eAssistant creates the image of a keyboard for the eyes of the user. In our eye-glass model the user will now see a keyboard floating in mid-air, and can type on it. Image processing based on what the camera delivers determines what keys have been touched: It may deliver audio feedback (producing a
246
Learning Support Systems
different click for each key that is hit) and video feedback by showing the text being typed, the text floating on a virtual screen above the keyboard. Using gesture recognition, movements of fingers or the hands replaces the mouse, nodding the head can be interpreted as mouse-click; or if two alternatives yes/no are offered or a two-button mouse is to be simulated, nodding or shaking the head could select one of the options. Alternatively, a simple gesture with the finger might also be used. As mentioned, simple inputs are also possible using the measurement of brain activity or other sensors: there is no limit to what one might imagine and it will be one of the interesting tasks to experiment with the combination of various techniques. Global and Culture-Fair Communication Clearly one of the functions of the eAssistant is that of a mobile phone. However, it is not necessary to press a phone against an ear, rather one can either use the loudspeaker mentioned earlier, or feed the audio-signal to the sides of the eye-glass and thus directly (via the ear-bone) to the inner ear, i.e. without other persons hearing or noticing anything. Since one can talk with closed mouth (after some practice) or spell a message by invoking a sequence of brain-states by thinking of designated actions (much like we spell a message we send an SMS) two persons can communicate over arbitrary distance in a way that other persons on the sender's or on the receiver's end do not notice it: thus, we have basically implemented telepathy in a technological manner, a fact much used in e.g. the XPERTEN- novel series (see www.iicm.edu/XPERTEN). Note that while two or more persons are communicating this way, the can also share arbitrary information, from what they currently see to information from local storage or a server, accessed via the net. It is also conceivable that the whole conversation might be recorded and stored for later perusal. Even communication with persons speaking different languages is quite conceivable: persons talks in the language of their choice; a speech recognition- translation- speech synthesizing program is translating what is said into whatever other languages are desired. Note that such
Epilogue
247
'machine' translations will not be perfect in the foreseeable future: 'To understand a language is to understand the world' is a famous statement that indicates clearly that for even near perfect language translations one would need computers at least as 'intelligent' as humans, and aware of all facets of the world and human life: and if this can ever be achieved is still a topic of much discussion. However, good machine translation programs are certainly doing a better job than the average person after having studied a language in school for a few years! Further, misunderstandings due to the translation process can be fairly easily avoided by feedback techniques: the translated material is translated back: if this now differs from the intent of the speaker, appropriate actions can be taken. By the way, the communication between persons with different languages may be made much easier by using dynamic symbolic languages, a main aim of the project MIRACLE (Maurer, Stubenrauch & Camhy, 2003) and its forerunner MUSLI (Lennon & Maurer, 2001). The eAssistant also changes how we discuss things: while someone is telling us something, we have the possibility to check if the information provide is correct, by accessing back ground libraries (Maurer, 2001) on local storage or in the internet. Conversely, we can use information from such background sources in our statements. This is of course assuming that access to desired information is easier and more selective than what we good get today using e.g. search engines in the internet. This is where techniques of knowledge management come in (Maurer & Tochtermann, 2002; Ives, Torrey & Gordon, 1998). That techniques such as similarity recognition and active documents (Heinrich & Maurer, 2000) can make quite a difference is shown in (Maurer & Tochtermann, 2002), that semantic nets and metadata (Meersmann, Tari & Stevens, 1999 allow to produce much better search results is shown by the success of knowledge networks and is the basis of one the currently leading knowledge management systems Hyperwave (Maurer, 1996). The new technologies will create many more uses and applications for the digital libraries and repositories currently being researched and developed (Oliver et al., 2003). Clearly the access and more productive use of information is not restricted to discussions with other persons, but applies universally to all situations when information is of critical importance. This is why
248
Learning Support Systems
knowledge tools as outlined in (Maurer & Tochtermann, 2002) are of such importance. PDAs Revisited It does not require that much imagination to see how the eAssistant is going to change our lives. We may suffice to provide just a few examples: When we meet a person the first time, information is usually exchanged by passing business cards and talking a bit about mutual backgrounds. Of course the exchange of information on business cards together what all that is available on the internet about this person, plus the pictures taken by the camera in the eye-glasses is recorded for later use. When we see the person next time, image processing software identifies who this is (even we have forgotten it), and supplies us with a wealth of information about this person. The eAsisstant is a perfect guide. Of course it can guide us when we drive the car, something already quite common for persons who drive top-of-the-line car models with built in navigation systems. But the eAssistant is equally helpful when we are walking, and not just for routing: when we look at a building the speech command 'Explain building' will be enough for the eAssistant to give us ample information: after all it knows (by GPS) where we are and (because of the compass) in which direction we are looking, so going into a guide book or such to retrieve what we want to know is easy. Clearly, this is not restricted to buildings, rivers, lakes, mountains...but equally well applies to plants or animals. If we look at a plant the speech command 'Explain flower' activates the camera and image processing, identifies what we are looking at and gives us information on the flower, on the berry, on the mushroom, etc. We will be paying with the eAssistant rather than with credit-cards or the like. The eAssistant will be our driving license and passport. It will automatically open those doors that we are authorized to enter. It will us allow with one command to turn on the light, the water, or what have you. And this does not just apply to things near to us: while we drive to our home we can turn on the air-conditioning, or the heating in our skiing cabin. This list can be continued indefinitely, and we intend to prepare a
Epilogue
249
more detailed study at a later point: it is our belief that an extensive list of what the eAssistant is good for will be rather mind-boggling and will be a strong incentive for the fast wide-spread deployment of eAssistants. So let us conclude this subsection with one more example from the realm of medicine: Suppose we have a sore throat. We call our doctor. He asks us to show our tongue. We use the macro mode of the camera in our eyeglasses (the camera can be taken out of its casing for such purposes) to send the picture to the doctor. While viewing the picture the doctor is supported by a computerized diagnostic system that uses image processing to find out what kind of infection we might be suffering from. Having decided what it is, the doctor makes sure that we can pick up required medication from a pharmacy near to us; after all, our current position is known (if we permit it) to the system due to our GPS. Note that sensors that supervise some of our physiological data (like bodytemperature and pulse) and monitor environmental data (like airtemperature and air quality) might alert us to take actions, or even initiate actions such as sending an ambulance to help us! There is even more to the widespread use of sensors and constant monitoring of sensory data. Suppose a person dies of some rare disease: comparing the date that has been monitored concerning this person over a long period with persons in similar circumstances who did not suffer from this disease might well discover the real reason for the disease at issue. Maybe this is the way how we will finally be able to combat mysterious illnesses such as the SDS (sudden death syndrome) in children, or the high rate of some type of cancer in certain groups of the population. What We May Learn Sophisticated learning programs that allow communication with others, with experts or with 'Interactive Knowledge Centers' or 'Active Documents' (Heinrich & Maurer, 2000) (which has been described in chapter 4 of this book) will allow us to pick up knowledge as we require it. The 'in-time' learning will be the natural thing to do, rather than learning just because certain things might be needed at some future time. The fact that we will have continuous access to information in all areas
250
Learning Support Systems
will mean that the learning of facts, today still an important component in myriad of areas from geography to history, from law to medicine, will become significantly less important. Thus, the eAsistant is not just going to continue the trend that we will learn in a different way, but will also deeply influence what we will want to and what we will have to learn in the future: Even activities like handwriting might become unimportant: Why use handwriting, when we have our eAssistant with rather more convenient ways of data- input - or: Why learn a foreign language for simple communication when we have automatic language translation, as explained earlier? Of course, we recognize that to understand a culture, we have to understand the language of this culture at a deep level: but for just travel or business the automatic translation devices will do. Thus, one of the main issues that will have to be investigated more than has happened so far is not HOW we learn in the future, but what we have to learn, when the eAssistant and the internet is more and more turning into an extension of our brain. Also, learning in the workplace will finally become a favored form of learning for many with help from the eAssistant and its supporting technologies and software (Oliver, 2001). Indeed this fact will have a deep influence on all of humanity. We live today in a totally tayloristic society as far as material goods are concerned, i.e. we are completely dependant on thousands of other professions and hundreds of thousands of other people for our daily life - from food, to housing, to clothing, to entertainment etc. - and we have accepted this. We have accepted that we can hardly survive as autonomous individuals any more, but only as part of large group of diverse humans. The point is, what has happened in the area of material products is now about to happen also in the area of the non-material products information and knowledge. We will become, for better or for worse, much more dependent on hundreds of thousands of other humans for what we have to know to function properly. The eAssistant is a big step in this direction: we will be profit from a powerful network sharing knowledge with others, the positive aspect, and we will become more and more dependent on this network, the negative aspect of this development. Thus, a further revolution of e-learning will be upon us if and when versions of eAssistants become widespread.
Epilogue
251
Planets of Learning Gilly Salmon from the ,Center for Innovation, Knowledge & Enterprise' of the Open University Business School located at Milton Keynes (UK) gave a keynote presentation for 2002 ,European Computer Assisted Language Learning', Eurocall, Conference entitled: 'Future learning encounters'. Salmon describes four scenarios which he called ,planets': Scenario 1, Planet of Contentious: Landing on Contentious you find technology as a delivery system. High importance is given to content management systems, integrated learning management systems, multi media, industry standards, DVDs, digital and cable TV. Rivalry between solutions providers is still strong, though two or three market leaders are emerging. The associated pedagogy is that of the transmission model of teaching, where information is transferred from experts to novices. Content is king. E-librarians and e-lecturers have closely linked roles. Lecturers need to captivate big audiences. The Internet and digital TV spawns its own lecturing stars and the most successful assume 'rock star' status. Of course there are still a few lecturers campaigning, to actually be with their students, rather than look at them on monitors [see: www.contentious.com]. Scenario 2, Planet Instantia: IBM estimates that 25% of employees' skills become obsolete every 3 years. With the increasingly global society, language and cultural understanding has become a paramount skill. Instantia meets these requirements through sophisticated learning object approaches, with information technology seen as the basic tools. The pedagogy on this planet is usually called e-learning. The role of ambient intelligence in devices is seen as key on this planet. Every device that is connected to electricity is also connected to the Internet. Hence educational providers are able to think both creatively and in a very integrated way about learning devices. Online trainers support autonomous learning. Real or virtual trainers are available 24 hours a day, both synchronously and asynchronously. Trainers focus on skills development in employees and on ways of fostering the adoption of a strong in-house knowledge culture. Scenario 3, Nomadict Planet: On Nomadict there is less stability, less structure, less fixed time for work and leisure, retirement and education
252
Learning Support Systems
compared to Earth. The sense of physical place is not strong. The Planet Nomadict provides portable learning for mobile lifestyles. Travelling users replace travelling information. Learning on the Planet Nomadict is time independent and individual. The learners are seen as electronic explorers and adventurers. Learning online is now called m-learning (for mobile-learning) instead of e-learning. Teachers, academics and researchers are as mobile as their students are. Many are portfolio teachers- working for several educational institutions and providers, all over the world, at any one time. They have not only a highly developed awareness of the ways in which traditions of learning and expectations vary in different cultures but also the ability to work across discipline and levels of education. Scenario 4, Planet of Cafelattia: On Planet Cafelattia, learning is built around learning communities and interaction, extending access beyond the bounds of time and space, but offering the promise of efficiency and widening access. The key technology is the developed, entertaining, effective Internet to allow immediate and satisfying interaction between students and students, and between teachers and students. The pedagogy is based on notions of a very strong social context for learning with the model of acquisition, argumentation and application. A key activity for learners is finding and interacting with like-minded individuals anywhere. Assessment is based on complex problem solving and knowledge construction skills. Teachers see the technologies as yet another environment for learning rather than as tools. An interesting debate on Cafelattia has been around the drivers towards commonality or difference. Salmon concludes that it is likely that all the ,planets' will have elements of reality and there will be a variety of players and processes. Currently there is not as much innovation and excitement as one may originally have imagined in a global society with good Internet access: Much of the learning looks similar to that in other disciplines and applications. You can be sure that learners will be on the Web; Research in a cluster of schools and kindergartens in late 2001 showed that 50% of the 3 year olds in the group recognized components of computers, were able to turn them on and off and had mouse skills. In 2013 these children will be secondary students. Of course, closing the gap between what we
Epilogue
253
have now in education and building a productive and successful future inevitably involves organizational change. Ecology of Mind The philosopher Vilem Flusser (who has been mentioned in chapter 8 of this book) envisioned a world where all humans are a part of a huge networked human-computer brain (Flusser, 1985). He also predicted that written languages will evolve into one global mathematical and graphical language. Whether or not we trust such predictions, we have to acknowledge that using ICT and acquiring the skills and competence to use it are mandatory to be part of modern societies: Competent usage of digital mass media, strategies of online learning and a strong basis for scientific and technical understanding are no longer ,computer literacy' but a key competence for ,social literacy'. It is interesting to note that the European discussion of this topic refers to the term ,cultural technique' which does make no sense when used in the U.S.A. for two reasons: 1. In the European culture ,technique' is a very general term for crafts and arts - in the US the conotation of technique' is much closer to ,technology'. 2. The meaning of ,culture' and ,civilization' is flipped around: The (US) English term ,culture' refers to ,civilization', e.g. in France and Germany, whereas the European term ,culture' refers to the English term ,civilization'. The ,Clash of Civilizations' (or, for Europeans: ,Clash of Cultures') has been considered to be a difference of religious values and social wealth. We may also think about the ,Digital Divide' and the dramatic ,brain drain' from the low developed regions discussed in the prologue of this book as the major ,clash of civilizations'. We would like to point out that the ,war for talent' (discussed in chapter 1 of this book) and the race to make the profit while supporting or defending the ,clash' might be the biggest divide in the world: The European Union and the European Commission has clearly stated that the growth perspectives of Europe are depending on the ability to start and to maintain a sustainable and global economy based on digital media (cf. Hasebrook, Rudolph &
254
Learning Support Systems
Herrmann, 2003). Another important gap might be the gap between the actual ability of the human brain to be fully ,computer alliterate' - and the fast growing and changing demands of the competence needed to master ICT. Therefore, we suppose to re-think technological developments from a ,human perspective', that is, not ask for possible technological progress but to ask for the fastest possible progress in our societies and cultures (or, for Europeans ,civilization'). Even if we do not share the idea of the near advent of 'spiritual machines' or 'artificial consciousness', we strongly support the idea of a discussion about the technological and social impact of those technologies. Many scientists consider the consciousness to be a 'virtual machine' (e.g. Daniel Bennet, 1986) or a not directly measurable effect of a meta-structure based on the innate brain structures developed in the millenniums of the evolution of the human race. Therefore, the argument goes that intelligence, such as the usage of tools and the development of language as a tool for communication, cannot be understood without the biological and social context, the 'ecology of mind' (cf. Kurzweil, 1999). Virtual Minds The description of an 'Ecology of mind' was a landmark in the development of cognitive psychology which was inspired and initiated when Gregory Bateson published his series of essay and talks in 1972. One of his most intriguing findings going beyond the simple cognitivism of the 70's was the detection of the 'learning of the unlearned': Dolphins were trained by Bateson not just by using 'shaping' - a behavioral approach which reinforces given behavior. Bateson reinforced completely new behavior, only. He found that his trainees were able to learn this complex meta-rule. The ecology of mind is supported by modern approaches of computer science and robotics, such as the works of Hans Moravec (2000), who claims that machine intelligence will automatically evolve when machines will possess an adequate computing power of more than 100 billions instructions per second and a (virtual) ecology to train their abilities. Recent scientific work, especially from genetics and archaeology, as well as some earlier psychological work is challenging this common approach of a smooth and continual
Epilogue
255
development based on pure computing power of the relatively large human brain: Richard Klein and Blake Edgar have published their electrifying book about 'The Dawn of Human Culture', this year. Their still unproven and hotly debated hypothesis is that - following a drastic reduction of the human gene pool due to natural catastrophes some 50'000 years ago - a single or a small set of mutations triggered the development of a conscious self and therefore of language, new tools and art. Julian Jaynes' psychological work about 'The Origin of Consciousness in the Breakdown of the Bicameral Mind' (1993) demonstrates that in most evolutionary aspects - and in daily life - is a unnecessary or even hampering, e.g. while learning new behavior such as swimming or skiing. Moreover, he claims that literature before and after the historic change of the writing direction of old and modern Greek letters between 600 to 400 B.C: indicates the origin of modern consciousness and culture; consciousness then is only a reflection of the boosted activity of the left hemisphere in the right hemisphere of the brain. Surprisingly, this assumption has been supported by scientific work about reading Semitic (right to left) and Latin (left to right) languages which shows that old languages tend to be perceived and understood by both brain hemispheres whereas modern languages tend to be 'left hemispheric'. Additionally, the outstanding experiments and experiences compiled by the psychiatrist V.S. Ramachandran (1998) gave a deep insight in the plasticity of the brain. More detailed analyses of brain activities while reading and understanding different languages have been conducted with psychological and neurological impaired persons (e.g. Paradis, 1977; Paradis et al., 1982) or for teaching reading to children (Furr, 2000). We would like to encourage interdisciplinary and international research activities which foster a comprehensive 'archaeology of mind' rather than an 'ecology of mind': Ecology asks for operational parameters which can be measured in an objective and comparable manner, rules about their interrelationship and reasonable judgements about the error parameters of the models (cf. Press etal., 1989).
256
Learning Support Systems
Ecology is a meta-structure (or meta-science) in itself which evolves as a scientific field of scientific biology well after the fundamental models of modern biology were settled; ecology was considerably transformed when evolution and genetics were discussed together (cf. Wells, 2003). The study of the human mind has - like most other scientific theories and models - a bias towards well measurable, consistent and continual developments. However, many scientific areas begin to acknowledge that more often than expected before chaotic and revolutionary events use to have (and still have) a deep impact on evolution. The study of consciousness is not ripe for an scientific 'ecology'. But it should come into a state where methods of scientific archaeology are applied to different scientific disciplines which can contribute to lay out a first map of the development of the human mind (e.g. Kosslyn & Koenig, 1995). Jean Piaget (1896-1980), the famous Swiss philosopher and amateur biologist, requested in his book about the wisdom and the limitations of philosophy from his colleagues (cf. Piaget, 1974, 1976): Theories and models have to be laid out in a way that they can be tested in experimental procedures collecting reasonable empirical and/or experimental data. The natural sciences and the humanities are now up to go the first steps of Piaget's way applying his ideas not only to the mental development of human children but also to the mental and cultural development of the entire human race. There are many discussions about consciousness, attention, free will and the responsibility of individuals in social contexts (see Walter, 2001). An archaeology of the human mind could represent the outstanding opportunity to tie together scientific and public discussions about one of the oldest and most exciting questions of the human society. Maps and Minds In her highly influential book on 'Linguistic Diversity in Space and Time', Johanna Nichols (1992) has provided a new way of juggling genetic and areal linguistic history. She distinguishes between spread and residual zones; in her terminology the ancient Near East was a spread
Epilogue
257
zone. Clearly, especially the far and the middle east has been affected by a number of Semitic spreads. In historical times we must reckon with three, if not, more such spreads: the early one that gave birth to the languages attested in the oldest documents, Amorite, and the controversial Aramaic spread - the language spoken by Jesus Christ. Today, it seems possible that archaeology can lay out maps of the development of mind (or certain cultural aspects, to be more precise). However, those maps are useless if they do not take into account there mental background which is not always directly reflected in the artifacts of a (deceased) culture. Hugh Brody gives a recent example about the problems to draw maps about cultural aspects in his famous book 'Maps and Dreams' (1981): Brody charted the hunting grounds of native Canadians in charge of the Canadian government; it turned out, however, that hunting was not so much influenced by territorial conditions but by dreams forecasting good hunting. As explained before, Julian Jaynes argues that consciousness arouse during a short period of time between 400 and 600 B.C. - a development marked by the shift of the ancient Greek literature. Jaynes analyses the literature, namely the Illiad, in order to better understand the state of consciousness of the writers and readers of the given time period which leads to some interesting research questions: If there are maps of language in space and time, can we draw maps about states of consciousness in space and time? If we can draw those maps, are they useful to understand certain shifts in cultural developments - or are they just temporarily territorial reflections of 'dreams', that is, mental states which can only be understood in the actual social context? Can we bring together maps of the development of language, artwork and other cultural achievements? Are those maps supported by crosscultural experiments about neurological and psychological processes, such as reading and understanding literature? It has been mentioned earlier that neurological correlates of cultural aptitudes and abilities have been found, e.g. understanding and reading different languages. It has also been pointed out that psychological research has contributed to the understanding of behavior and brain activities accompanying cognition and emotion. Finally, computer and
258
Learning Support Systems
information science can give us the ability to simulate certain aspects of these research findings and to test and to extend their functionality. This proposal suggests that not the natural sciences should guide the way to a better understanding of the human mind because their methods are to demanding in a statistical or mathematical sense. Nor should philosophy or the humanities drive the research activities because there methods are mostly not compatible with those used in natural and computer sciences. It is worthwhile to examine empirical sciences which mostly rely on extensive and exact observation but only on weak models. This holds true for archaeology (or astronomy, for that matter; cf. Kuhn, 1970; Hofmann, 1999). Therefore, an archaeology of mind can break the ground for sound and fast proliferating research activities discovering and unearthen the treasures of the human mind. Invisible Computing and Embedded Learning In the prologue we stated, that technology solves technological problems, only. Learning, however, is not a technological problem. Thus, ICT does not help to overcome didactical faults. We distinguished three levels of digital media in education (Keil-Slawik, 2003): « The primary level provides general infrastructure and software features used for learning, * The secondary level comprises specific software features for learning, and « The tertiary level introduces self learning and adaptive software features. We demanded that knowledge has to be available like the ubiquitous power of electricity with education as its transmitter. New media are the latest transmission technology which can carry education into low developed regions using a lean physical infrastructure. Global trends indicate that new media can help to built the link between people and wealth. As knowledge is not a factor of production easily accessible as financial capital or soil a 'war for talent' has started (cf. Martin & Moldovenau, 2003). Based on these assumption we conclude that there
Epilogue
259
are two major unsolved (and widely ignored) problems that hamper the progress of learning support systems: • We do not have a holistic design model based on human mental processes, nor do we have an appropriate user interface design to step overfromcautomation9 to 'augmentation5. We dot not understand how to apply digital media for sustainable growth and global infrastructures because political, social, and technological hurdles are still dominating local or regional markets and standards.
Figure 54. The efiport system design model (pat. pending): All levels of digital media in education are embedded in the organization centered around skills and competencies which are connected to the actual learning objectives related to the learning options and activities.
260
Learning Support Systems
Nevertheless, we are optimistic that we shall be able - step by step to approximate potential solutions for the two dominating problems. We believe that the following global trends will control the development of e-learning and learning support systems. We shall experience the 'down grading' of ICT solutions and services following the global trends in service industries. For instance, people banks are carrying financial mediation into the low developed regions as 'micro banking' and 'micro finance'; traditional banking institutions, however, cannot cope with micro lending and micro credits and fall behind. Similarly, ICT producers and service providers will miss the opportunities of globalization and sustainable economic growth if they are not able to produce lean infrastructures and 'down graded' ICT solutions for community access and usage. We have seen the taylorism of the industry, and we have mentioned the taylorism of 'knowledge work'. We think that the decomposition of the value chain for economic reasons will drive the development of 'virtual experts team' where some irreplaceable experts will be the 'talents' enjoying high salaries and the rest will be in constant search for new poorly paid jobs. Therefore, the 'thinking networks' of Vilem Flusser will turn out to be 'learning organizations' built around economic values. Learning tasks will be embedded into work tasks, that is, corporate and project management will calculate with constant and relevant learning processes and foster them in order to be more efficient and effective. In the future, we may have jobs which demand only two or three hours of work per day - but the same jobs will demand several hours of daily long learning; the necessity of life-long learning will also grow because all highly developed societies will suffer from 'over-aging' putting more responsibilities on the backs of the elderly people. Until now, the biological evolution used selection to adapt individuals to the natural environment. What does it mean to adapt human beings to artificial environments? Certainly (or probably) not the transformation from hands and fingers into virtual keyboard devices. Moreover, the speed of evolution seems to be extremely slow as compared to the speed of ICT development. The visions of man-machine creatures or 'cyborgs' seem to be part of science fiction novels not reality. But the merger of men and machines has began, already. However, no flesh and metal
Epilogue
261
cyborgs will evolve but embedded neural implants, such as intelligent biosensors with Internet access and artificial ears covered with real flesh from the biogenetic laboratories. Scientific progress allows us to manipulate genomes at the level of single genes. We already have in vitro fertilization and animal cloning; in the near future, we might have human cloning and exploitation of embryonic stem cells. We are developing machines that will surpass the human brain in raw computing power and capabilities to connect to a world of information-processing devices. Eventually, we must come to terms with the fact that genes, machines, and minds are continuous entities both in space and time (cf. Baldi, 2002). Therefore, it has been suggested that individuals may be more keen to safe their minds rather than their genes. Biological evolution is built on the adaptation and extension of information stored in the genome of a certain population. Rapid evolution in small population can develop new races within thousands and hundreds of years (instead of millions of years). We believe that a mind will not exist outside a body, but we are not sure how this 'body' will be and can be produced and maintained in the future. Digital media are changing the way we live and think every day. Generations of children will grow up with a very dense provision of digital media services. This means that digital media are not only changing our minds but also, literally, our brains, because the brain structure develops in two major steps: in early childhood and as a teenager - these are also times of massive media 'consumption'. Digitized human beings and people without extensive access to ICT will not be able to understand each other because their biographies and their brain structures are too different. Mind sets and virtual personalities representing real humans will be stored in computers. Computers will develop into even more 'intelligent machines', but we will ignore it just like we are ignoring it today: Playing chess was considered to be 'intelligent behavior' until computers started to do so; we define all computer-based behavior as 'non intelligent' - and therefore we shall not even notice when computers start to develop their own way of being 'intelligent'. The most important progress, however, will not come from 'big science' and 'intelligent technology' but from 'applied science', 'lean technology' and 'big (economic) value': Sustainability in terms of
262
Learning Support Systems
resources and global reach will be key technologies to win the race for global wealth. The key to develop better e-learning will be the destruction of e-learning with its learning platforms and WBTs: Learning is anywhere and anytime, because it is our nature to learn. Learning platforms will be embedded learning support systems being aware of our current knowledge and learning situation. In the past, we delivered isolated pieces WBT; today, WBT is part of the working and business processes; tomorrow, learning will be part of controlling and adapting work and business processes. The need to do (physical) work will decrease but the necessity to learn will increase even more. What we are waiting for is, that learning systems are self aware when and where they are needed. Form this point of view, we should not discuss 'Learning Support Systems for Organizational Learning' but 'Learning Organization Systems for System Learning'.
Appendices
Abbreviations
ADL: Advanced Distributed Learning Initiative AI: Artificial Intelligence AICC: Aviation Industry CBT Committee CBT : Computer Based Training CI: Computational Intelligence CMC: Computer Mediated Communication DRJVI: Digital Rights Management ERP: Enterprise Resource Planning EVA: Economic Value Added FIPA: Foundation for Intelligent Physical Agents GLM: General Linear Model GPS: Global Positioning System HRD: Human Resource Development HR XML: Human Resource extended Markup Language HR SEP: Human Resource Standard Exchange Protocol HTML: Hypertext Markup Language ICT: Information and Commnucation Technology IEEE LTSC: Institute of Electrical and Electronics Engineers Learning Technology Standards Committee IP: Internet Protocol IPR: Intellectual Property Rights ISO: International Standard Organization ITS: Intelligent Tutoring System LMX: Leadership-Membership-Exchange LOM: Learning Object Meta-Data MIPS: Million Instructions per Second 263
264
Learning Support Systems
MVA: Market Value Added NGO: Non-Govemmental Organization OLTP: Online Transaction Processing SCORM: Shareable Courseware Resource Model SGML: Structured General Markup Language SME: Small and Medium Enterprises / Subject Matter Expert TOC: Total Cost of Ownership TTO: Time To Operation VPN: Virtual Private Network WACC: Weighted Average Cost of Capital WBT: Web Based Training XML: extended Markup Language
Appendices
265
Web Links
Note: Some of the following Web links have been collected and reviewed by Gilly Salmon from Centre for Innovation, Knowledge & Enterprise, Open University Business School, Milton Keynes (UK). General Interest Farrell, G.M. (Ed.) (2001). The Changing Faces of Virtual Education ® www.col.org/virtualed. The Commonwealth of Learning (COL) Principles of knowledge management « www.know-center.at * www.bus.utexas.edu/kman/kmprin.htm Speech technology standards * www.speech.cs.cmu.edu/ ® www.computerworld.com.au/IDG2.NSF/a/0005C942?OpenDocu ment&n=e&c=CT Learning Technology Standards » www.masie.com/standards/S3_Guide.pdf * www.adlnet.org/ Commercial Consultants « www.brandonhall.com/ * www.masie.com/ Learning Support Technology Providers & Institutes ® www.aace.org ® www.efiport.com « www.hyperwave.com » www.iicm.edu ® www.isnm.de Emerging Technologies ® www.newtech.org/addresslO_en.htm * wearables.cs.bris.ac.uk/
266
Learning Support Systems
» wearcam.org/mcluhan-keynote.htm $ www.pjb.co.uk/mobile_comm.htm * www.trainlngzone.co.uk/item/37933 * human-factors.arc.nasa.gov/ihh/psychophysio/ « www.microsoft.com/pocketpc/ * www.datacommresearch.com ® www.xybernaut.com/ ® iswc.gatech.edu/ Collaborative Learning * csalt.lancs.ac.uk/jisc/ * collaborate.shef.ac.uk/spender.htm « www.shef.ac.uk/uni/projects/wrp/sem2.html ® cbl.leeds.ac.uk/~tim/networked_learning/ * www.icbl.hw.ac.uk/jtap-573/cultures.html * www.ucisa.ac.uk/TLIG/conf/tlig00/w26/ Online Tutoring ® oubs.open.ac.uk/e-moderating ® www.e-moderating.com « www.itee.uq.edu.au/~tutoring/Tutor_responsibilities/tutoring_pro blems_and_Performance.htm Online Assessment * www.nwrel.org/learns/resources/measurement/Outcomes_and_Pe rformance_Measurement.pdf « materials.netskills.ac.uk/info/module52.html ® www.derby.ac.uk/ciad/ ® www.scaan.ac.uk/hw_caa.doc « www.ltss.bris.ac.uk/VLEintro_5_3 .htm » www.caaconference.com/ * www.caacentre.ac.uk/ Future Developments * www.tfi.com/ * oubs.open.ac.uk/future * www.wfs.org/index.htm * www.foresight.gov.uk
Appendices
267
About the authors
Joachim P. Hasebrook High school diploma (Abitur) and head an offset printing office in a little city in the north of Germany. Master of science with a major in psychology and bachelor of science with a major in computer science at the University of Marburg. Ph.D. thesis about learning with multimedia and hypermedia systems at the University of Giessen, Germany. Further education at the schools of German booktraders in Frankfurt, Germany, and Executive Education at the Goizueta Business School at Emory University in Atlanta, USA. Expert opinions about effective learning with multimedia and interactive distance learning environments in charge of the German. Parliament and the Scientific Options Assessment (STOA) of the European Parliament. Project manager for electronic media and expert systems for a German publishing and software house in charge of the Federal Institute of Labout Since 1996, member of staff of Bank Academy holding the position of the Head of the Department 'Concept & Programme Development1. Temporarily, Co-CEO of Knowbotic Systems Inc., a developer of software for distributed computational intelligence. Managing director of educational financial portal [efiport] Inc., a subsidiary of Bank Academy. Chair of the Task Force Multimedia of the European Bank Training Network (EBTN) and member of international program committees, e.g. WebNet / Elearn and ICCE/ICCAI, and member of editorial boards of scientific journals, e. g. KES Journal and Journal of Universal Computer Science (J.UCS). Since 2004, Ml professor for e-learning and work design at the International School of New Media (ISNM) of the University of Luebeck, Germany (www.isnm.de).
268
Hermann A. Maurer Sp?---l---•'" ••^4^^S0S^ m&,'.: ''iwK
Learning Support Systems
Study of Mathematics at the Universities of Vienna (Austria) and Calgary (Canada) starting in 1959. System Analyst with the Government of Saskatchewan (Canada) in 1963. Mathematicianprogrammer with IBM Research in Vienna 19641966. Ph.D. in Mathematics from the University of Vienna 1965. Assistant and Associate Professor for Computer Science at the University of Calgary 1966-1971. Full Professor for Applied Computer Science at the University of Karlsruhe, West Germany, 1971-1977, and Visiting Professor at SMU, Dallas, and University of Brasilia (Brazil) for three months, each, and at the University of Waterloo, during the same period. Adjunct Professor at Denver University 1984-1988; Professor for Computer Science at the University of Auckland, New Zealand, in 1993 (on leave from Graz), then Honorary Adjunct Professor and since May 2001 Honorary Research Fellow. Full Professor at the Graz University of Technology since 1978, since October 2000 also Dean of Studies for Telematics. In addition, director of the Research Institute for Applied Information Processing of the Austrian Computer Society 1983-1998; chairman of Institute for Information Processing and Computer Supported New Media since 1988, director of the Institute for Hypermedia Systems of JOANNEUM RESEARCH since 1990, director of the AWAC (Austrian Web Application Center) of the ARCS (Austrian Research Centers) 19972000, member of the board of OCG (Osterreische Computergesellschaft) 1979-2003 and since 2001/01/01 chief scientist of the KNOW Center (K+ Center), the first research center on Knowledge Management in Austria (www.iicm.edu; www.know-center.at).
Bibliography
Aebli, H. (1994). Uber das Verstehen von Witzen - eine kognitionspsychologische Analyse und einige padagogische Schlussfolgerungen. In K. Reusser & M. Reusser-Weyeneth (Eds.), Verstehen. Psychologischer ProzeB und didaktische Aufgabe (89-110). Bern: Hans Huber. Albrecht, J.E., & O'Brien, E. (1993). Updating a mental model: Maintaining both local and global coherence. Journal of Experimental Psychology: Learning, Memory & Cognition, 19, 1061-1070. Allehoff, W. (1985). Berufswahl und berufliche Interessen. Gottingen: Hogrefe. Argyris, C, & Schon, D. (1978). Organisational Learning: A Theory of Action Perspective. Reading: Addison Wesley. Arnold, J., Cooper, C.L., & Robertson, I.T. (1995). Work psychology - understanding human behaviour in the workplace. London: Pitman. Ausbilderhandbuch (1998). Daten und Fakten zur bankbetrieblichen Aus- und Weiterbildung. Stuttgart: Deutscher Sparkassen Verlag. Baddeley, A.D. (1990). Human memory. Theory and practice. Hillsdale, NJ: Lawrence Erlbaum. Baldi, P. (2002). The shattered self. The end of natural evolution. Cambridge: Bradford / MIT. Bandura, A. (1962). Social learning through imitation. In M.R. Jones (Ed.), Nebraska Symposium on Motivation, 1962. Lincoln: University of Nebraska. Bandura, A. (1977). Social learning theory. Englewood Cliffs: Prentice-Hall. Barba, R.H. (1993). The effects of embedding an instructional map in hypermedia courseware. Journal of Research on Computing in Education, 4(25), 405-412. Barba, R.H., & Armstrong, B.E. (1992). The effect of HyperCard and interactive video on earth and space science students' achievements. Journal of Educational Multimedia and Hypermedia, 1, 323-330. Bastien, C. (1992). Ergonomics for hypermedia software. In A. Oliviera (Ed.), Hypermedia courseware: Structures of communication and intelligent help (183187). Heidelberg: Springer. Bateson, G. (1972). Steps to an ecology of mind. New York: Ballantine.
269
270
Learning Support Systems
Bender, C, & Rohling, T. (2001). Ansatze zur Bewertung und Risikomessung von Humankapital. In H. Kossbiel (Hrsg.). Modellgestiltzte Personalentscheidungen (Bd. 5, 27-39). Munchen: R. Hampp. Bennett, R.E. (1993). Toward intelligent assessment: An integration of constructedresponse testing, artificial intelligence, and model-based measurement. In N. Frederiksen, R.J. Mislevy and I.I. Bejar (Eds.), Test theory for a new generation of tests (99-123). Hillsdale, NJ: Erlbaum. Bergmann, C. (1992). Schulisch-berufliche Interessen als Determinanten der Studienbzw. Berufswahl und -bewaltigung: Eine Uberpriifung des Modells von Holland. In A. Krapp & M. Prenzel (Hrsg.), Interesse, Lernen, Leistung. Neuere Ansatze der padagogisch-psychologischen Interessenforschung. Minister: Aschendorff. Berkel, K., & Herzog, R. (1997). Unternehmenskultur und Ethik. Heidelberg: Sauer Bernath, U., & Rubin, E. (1998). A virtual seminar for international professional development in distance education. Informatik Forum, 12(1), 18-23. Blickle, G. (2000). Mentor-Protege-Beziehungen in Organisationen. Zeitschrift fur Arbeits- und Organisationspsychologie, 44(4), 168-178. Bloom, F.E., & Lazerson, A. (1999). Brain, Mind, and Behaviour. Second edition, W.H. Freeman and Company, Worth Publishers. Boettcher, J. (Ed.) (1993). 101 success stories of information technology in higher education: The Joe Wyatt challenge. New York: McGraw-Hill. Boling, N.C., & Robinson, D.H. (1999). Individual study, interactive multimedia, or cooperative learning: Which activity best supplements lecture-based education?. Journal of Educational Psychology, 91(1), 169-174. Brett, J.F., & VandeWalle, D. (1999). Goal orientation and goal content as predictors of performance in a training program. Journal of Applied Psychology, 84(6), 863-873. Bridges, M.P. (1989). Trends and developments: Software for career counseling. In G.R. Walz, J.C. Bleuer & M. Maze (Eds.), Counseling Software Guide (15-19). Alexandria, VA: AACD. Brody, H. (1981). Maps and dreams. Indians and the British Columbia frontier. Vancouver: Douglas & Mclntire. Brooks, R.A: (1999). Cambrian intelligence. The early history of the new AI. Cambridge: Bradford/MIT. Brosius, H.B., & Kayer, S. (1991). Der Einflufi von emotionalen Darstellungen im Fernsehen auf Informationsaufnahme und Urteilsbildung. Medienpsychologie, 3, 236-253. Brosius, H.B., & Mundorf, N. (1990). Eins und eins ist ungleich zwei: Differentielle Aufmerksamkeit, Lebhaftigkeit von Information und Medienwirkung. Publizistik, 35, 398-407. Bubik, R., Quenter, D., & Ruppelt, T. (2000). Informationstechnik - selten geschaftsbezogen gefuhrt. Harvard Business Manager, 22(2), 102-111. Buchtemann, C.F., Schupp, J., & Soloff, D.J. (1993). Ubergange von der Schule in den Beruf - Deutschland und USA im Vergleich. Mitteilungen des Instituts fur Arbeitsmarkt- und Berufsforschung (MitAB), 4/93, 507-520.
Bibliography
271
Calvert, S.L., Huston, A.C., & Wright, J.C. (1987). Effects of television preplay formats on children's attention and story comprehension. Journal of Applied Developmental Psychology, 8, 329-342. Carr, N.G. (2003). IT doesn't matter. In Harvard Business Review, May 2003, 11-22. Does IT matter? In HBR Debate, Web Exclusive, June 2003, 1-17. Challis, B.H., & Velichkovsky, B.M. (1999). Stratification in cognition and consciousness. Amsterdam/Philadelphia: John Benjamins. Chauvin, K., & Hirschey, M. (1994). Goodwill, profitability, and market value of the firm. Journal of Accounting and Public Policy, 13, 159-180. Chomsky, N. (1971). Syntactic structures. The Hague: Mouton. Chruchland, P.S., & Sejnowksi, T.J. (1996). The computational brain. Cambridge, MA: M.I.T./Bradford. Clancey, W.J. (1987). Methodology for building an intelligent tutoring system. In G.P. Kearsley (Ed.), Artificial intelligence and instruction (193-227). Reading, MA: Addison-Wesley. Clark, R.E., & Craig, T.G. (1992). Research and theory on multi-media learning effects. In M. Giardina (Ed.), Interactive multimedia learning environments. Human factors and technical considerations on design issues (19-30). Heidelberg: Springer. Collins, A., & Brown, J.S. (1988). The computer as a tool for learning through reflection. In H. Mandl & A. Lesgold (Eds.), Learning issues for intelligent tutoring systems (1-18). Heidelberg: Springer. Cox, E. (1994). The handbook of fuzzy systems. New York: Academic Press. Cox, R., & Brna, P. (1995). Supporting the use of external representations in problem solving: The need for flexible learning environments. Journal of Artificial Intelligence in Education, 6(2/3), 239-302. Crain, L.C. (1994). Effects of instructional media in immediate and long-term recall. Interpersonal Computing and Technology, 2(2), 19-27. Crick, F. (1994). The Astonishing Hypothesis: The Scientific Search for the Soul. New York: Scribner. Cronbach, L.J., & Gleser, G.C. (1965). Psychological tests and personnel decisions (2nd ed.). Urbana: University of Illinois Press. Cummings, L.E. (1995). Educational technology - a faculty resistance view. Educational Technology Review, 4, 13-18. Cummins, R. (1991). Meaning and mental representation. London: Bradford. Dale, E. (1946). Audiovisual methods in teaching. New York: Holt, Rinehart & Winston. Davis, L. (1991). Handbook of genetic algorithms. New York: Van Nostrand, Reinhold. Deidesheimer Kreis (1993). Feststellung der Studieneignung im Rahmen der Hochschulzulassung. Studienfeldbezogene Verfahren zur Feststellung der Studiereignung bei Hochschulzulassungsentscheidungen im Numerus-clausus- und anderen Studienfachern. Bericht fur das Bundesministerium fur Bildung und Wissenschaft, Bonn. Dennett, D.C. (1986). Consciousness as a virtual machine. Research Group on 'Perception and Action', report no 143/1986. Bielefeld: Center for Interdisciplinary Research (ZiF).
272
Learning Support Systems
Dietinger, T. (1998). GENTLE - General Networked Training and Learning Environment. EdMedia98. Charlottesville: AACE. Dietinger, T., Giitl, C., Knogler, B., Neussl, D., & Schmaranz, K. (1999). Dynamic background libraries - new developments in distance education using HIKS. Journal of Universal Computer Science, 5, 2-10. Doeben-Henisch, G. (1998). The emergence of signs in populations of artificial agents. Paper presented at the Conference on Intelligent Systems and Semiotics (ISAS 98). Gaithersburg, Maryland. Doeben-Henisch, G. (1998). Semiotic machines - an Introduction. In E. W. B. Luettich, J. E. Mueller & A. van Zoest (Eds.), Signs & space, Raum & Zeichen. An international conference on the semiotics of space and culture (313-327). Tubingen: Gunter Narr. Doeben-Henisch, G., & Hasebrook, J. (1999). A Toolbox of Artificial Brain Cells to Simulate Classical and Operand Learning Behavior. 6th Conference on Computational Intelligence. Dortmund: Universitat Dortmund. Dohmen, G. (2001). Das informelle Lernen - Die internationale ErschlieBung einer bisher vernachlassigten Grundform menschlichen Lernens fur das lebenslange Lernen aller. Bonn: Bundesministerium fur Bildung und Forschung (BMBF). Drewniak, U. (1992). Lernen mit Bildern in Texten. Untersuchung zur Optimierung des Lernerfolgs bei Benutzung computerprasentierter Texte und Bilder. Munster: Waxmann. Duchastel, P. (1991). Towards methodologies for building knowledge-based instructional systems. Instructional Science, 20(5/6), 349-358. Eberhart, R., Simpson, P., & Dobbins, R. (1996). Computation Intelligence PC Tools. Academic Press. Edvinsson, L., & Malone, M.S. (1997). Intellectual capital. London: Piatkus. Edvinsson, L., & Brtinig, G. (2000). Aktivposten Wissenskapital - Unsichtbare Werte bilanzierbar machen. Wiesbaden: Gabler. Edvinsson, L., & Kivikas, M. (2003). The new longitude perspective for value creation. Controlling, 3(4), 163-167. Erpenbeck, J., & Heyse, V. (1999). Die Kompetenzbiographie - Strategien der Kompetenzentwicklung durch selbstorganisiertes Lernen und multimediale Kommunikation. Munster: Waxmann. Erpenbeck, J., & Heyse, V. (1999). Kompetenzbiographie - Kompetenzmilieu Kompetenztransfer, (62). Berlin: QUEM. Erpenbeck, J., & von Rosenstiel, L. (Hrsg.) (2003). Handbuch Kompetenzmessung Erkennen, verstehen und bewerten von Kompetenzen in der betrieblichen, padagogischen und psychologischen Praxis. Stuttgart: Schafer-Poeschel. Escalada, L.T., Grabhorn, R., & Zollman, D.A. (1996). Applications of interactive digital video in a physics classroom. Journal of Educational Multimedia and Hypermedia, 5(1), 73-97. Flamholtz, E.G. (1985). Human resource accounting. London, San Francisco: Jossey Bass. Flamholtz, E.G., & Lacey, J. M. (1981). Personnel management, human capital, human resource accounting. Los Angeles: University of California.
Bibliography
273
Flusser, V. (1985). Ins Universum der technischen Bilder. Gottingen: European Photography. Fock, E., & Engelbrecht, G. (1986). MOT, Ergebnisse der Validierungsuntersuchungen und der Vorarbeiten fur die Entwicklung eines studienfachbezogenen Auswertungsverfahrens. Nilrnberg: Bundesanstalt fur Arbeit (Psychologischer Dienst). Fogel, D.B. (2000). Evolutionary Computation, Towards a New Philosophy of Machine Intelligence. New York: IEEE Press. Frese, M., Brodbeck, F.C., Zapf, D., & Prumper, J. (1991). Users' errors and error handling: its relationship with task structure and social support. ACM SIGCHI Bulletin, 23(2), 59-62. Funke, U. (1993). Computerunterstiltzte Eignungsdiagnostik mit komplexen dynamischen Szenarios. Zeitschrift fur Arbeits- und Organisationspsychologie, 37(3), 109-118. Furr, D.L. (2000). Reading clinic. Brain reseach applied to reading. Chicago, 111.: Truman House. Garcia, S. (2001). Kompetenzbilanzierung. Unpublished Master Thesis. Zurich: Eidgenossische Technische Hochschule (ETH) Zurich. Gebauer, K., & Huether, G. (2002). Kinder suchen Orientierung. Anregungen fur eine sinnstiftende Erziehung. Dusseldorf: Patmos / Walter. Gemtinden, H., & Bobin, T. (1999). Projektmanagement in der Softwareentwicklung. Ergebnisbericht. Karlsruhe: Universitat Karlsruhe.
Gershenfeld, N. (1999). When things start to think. New York: Henry Holt. Ghiselli, E.E. (1973). The validity of aptitude test in personnel selection. Personnel Psychology, 26, 461-477. Gierig, R. (2002). Human Capital in Unternehmen: Die Bewertung unterschiedlicher Ansatze zur Messung von Humankapital. Frankfurt am Main: Hochschule fur Bankwirtschaft - Private Fachhochschule der Bankakademie e.V. Glowalla, U. & Hasebrook, J. (1995). An evaluation model based on experimental methods applied to the design of hypermedia user interfaces. In W. Schuler, J. Hannemann & N.A: Streitz (Eds.). Hypermedia user interface design, (99-116). New York: Springer. Goddard, N„ & Hood, N. (1997). Parallel GENESIS for large scale modeling. In J. M. Bower (Ed.), Computational Neuroscience '96. New York: Plenum. Gorz, G. (Ed.)(1993). Einfuhrung in die kunstliche Intelligenz. Bonn: Addison-Wesley. Goodwin, J. & Hoppin, J. (1988). Adult career counseling - new clienteles. ERIC Digest CG021498, access number ED304624. Gruber, T.R. (1993). A translation approach to portable ontologies. Knowledge Acquisition, 5(2), 199-220. Guthrie, J. (2000). The management, measurement and the reporting of intellectual capital. CIMA, Macquarie Graduate School of Management, London. Retrieved July 11, 2003, from the World Wide Web: http://www.cimaglobal.com/downloads/james_guthrie_lecture.pdf
274
Learning Support Systems
Hall, B. (2001). Learning Management Systems 2001: How to Choose the Right System for Your Organization. Sunnyvale (CA). Hanft, A., & Miiskens, W. (2003). Kompetenzbezogene Erfolgskontrollen internetgestiltzten Lernens: Gutachten des Projektes "Lernerfolgskontrolle und Zertifizierung des Lernens im Netz und mit Multimedia unter dem Aspekt der Lernkulturveranderung und Kompetenzentwicklung". Oldenburg, Berlin: Lernen im Netz und mit Multimedia, QUEM. Hanft, A., & Miiskens, W. (ohne Datum). Die Zertifizierung selbstorganisierten Lernens in der beruflichen Weiterbildung - Konsequenzen der Entkopplung von Lernen und Prilfen. Harasim, L.M., Hiltz, S.R., Teles, L., & Turoff, M. (Eds.) (1995). Learning Networks: A Field Guide to Teaching and Learning Online. Boston: MIT Press. Hasebrook, J. (1995). Multimedia-Psychologie. Eine neue Perspektive menschlicher Kommunikation. Heidelberg: Spektrum Akademischer Verlag (Buch + CD-ROM). Hasebrook, J. (1995). Lernen mit Multimedia. German Journal of Eduactional Psychology, 9(2), 95-103. Hasebrook, J. (1996). Short assessment and evaluation of two reports about tele-learning and cooperative learning in Europe. Expert opinion in charge of the Scientific and Technical Options Assessment (STOA) of the European Parliament, Brussels. Hasebrook, J., & Gremm, M. (1996). Struktur beruflicher Interessen bei Jugendlichen (Structure of vocational interests in young adults, in German). German Journal of Educational Psychology, 10(2), 85-103. Hasebrook, J. (1997). Using Multimedia and Software Quizzes to provide vocational guidance. In S. Manning (Ed.), European Symposium on Information Technology in Education (18-42). Dresden University Press. Hasebrook, J., & Nathusius, W. (1997). An expert advisor for vocational guidance. Journal of Artificial Intelligence in Education. 8(1), 21-41. Hasebrook, J., & Steffens, U. (1997). Weiterbildung per Digitalfernsehen. Die Bank 11/97,676-680. Hasebrook, J., & Wagner, J. (1997). Subjektive Bewertung von gedruckten und elektronischen Medien zur Berufsorientierung. Medienpsychologie, 9(2), 89-104. Hasebrook, J. (1998). Multi-Media. In D. Rost (Hrsg.), Handbuch der Padagogischen Psychologie (346-351). Weinheim: Beltz PVU. Hasebrook, J. (1998). Aptitude-Treament-Interaction. In D. Rost (Hrsg.), Handbuch der Padagogischen Psychologie (9-13). Weinheim: Beltz PVU. Hasebrook, J. (1998). Einsicht start Aufsicht: Zuschauen beim Lernen und Lehren. In B. Straufi (Hrsg.), Zuschauer (57-92). Gottingen: Hogrefe. Hasebrook, J., & Gremm, M. (1999). Multimedia for vocational guidance: Effects of testing, videos, and photography on acceptance and recall. Journal of Educational Multimedia and Hypermedia, 8(2), 217-240. Hasebrook, J.P. (1999). Learning in electronic media: Exploring Web-based training and the human mind. Proceedings of the WebNet conference 1999, Charlottesville, VA: AACE, 1025-1032. Hasebrook, J. (1999). Searching the web without losing the mind - traveling the knowledge space. WebNet Journal, 1(2), 24-32.
Bibliography
275
Hasebrook, J. (1999). Web-based training, performance, and controlling. Journal of Network and Computer Applications, 22, 51-64. Hasebrook, J. (2000). Knowledge workers and knowledge robots. Invited paper. Proceedings of International Conference of Computer in Education (ICCE), Taipeh, Taiwan. Hasebrook, J. (2001). Learning for the learning organization. Journal for Universal Computer Science, 7(6), 472-487. Hasebrook, J., Erasmus, L., & Doeben-Henisch, G. (2001). Knowledge robots for knowledge workers: Self-learning agents connecting information and skills. In L.C. Lain & Z. Chen (Eds.), Practical applications of intelligent agents (63-90). New York: Springer. Hasebrook, J. (2002). Cooperative and interactive distance learning: application of teamoriented and selective learning strategies in a European bank. Journal of Universal Computer Science, 8(9), 834-847. Hasebrook, J. & Otte, M. (2002). Die dritte Welle: E-Learning im Zeitalter des Electronic Commerce. Bern: Hans Huber. Hasebrook, J., Rudolph, D.W. & Steffens, U. (2002). E-Learning Business Strategies & Opportunities. Chichester (MI): Datacom Research Report. Hasebrook, J., Herrmann, W., & Rudolph, D. (2003). European perspectives for elearning: Markets, technologies, and strategies. Thessaloniki: CEDEFOP (European Centre for Vocational Training). Hasebrook, J. (2003). Lernen in der New Economy. In L. Beyer, D. Frick, A. Gadatsch, I. Maucher & H. Paul (Hrsg.), Vom E-Business zur E-Society: New Economy im Wandel (152-198). 1. Aufl. Munchen: Hampp. Hasebrook, J. (in press). Vom Informations- zum Kompetenzmanagement am Beispiel der betrieblichen Weiterbildung (From information to competence management applied to vocational training, in German). In U. Hugl. (Hrsg.) Virtuelle Personalentwicklung? Status und Trends IuKT-gestutzten Lernens. Eine kritische Analyse. Wiesbaden: Gabler. Haykin, S. (1994). Neural Networks, A Comprehensive Foundation. New York: Macmillan College Publishing Company. Hegarty, M. (1992). Mental animation: Inferring motion from static displays of mechanical systems. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18(3), 1084-1102. Heinrich, E., Patrick, J., & Kemp, E. (1999). A Flexible Structured Language for the Coding of Human Behaviour - FSCL, Massey University Department of Information Systems Report 1. Heinrich, E., & Maurer, H. (2000). Active Documents: Concept, Implementation and Applications. J.UCS 6, 12, 1197-1202. Heinrich, E., Johnson, R., Luo, D., & Maurer, H. (2001). Learner-Formulated Questions in Technology-Supported Learning Applications. Proceedings of ED-MEDIA 2001, Tampere, AACE, Charlottesville, VA (2001), 720-725. Heyse, V., Erpenbeck, J., & Michel, L. (2002). Kompetenzprofiling Weiterbildungsbedarf und Lernformen in Zukunftsbranchen. Munster: Waxmann.
276
Learning Support Systems
Hiltz, S. R., & Benbunan-Fich, R. (1997). Supporting Collaborative Learning in Asynchronous Learning Networks. Invited Keynote Address for the UNESCO/ Open University Symposium on Virtual Learning Environments and the role of the Teacher. Milton Keynes (UK). Hoelscher, G.R. (1994). Kind und Computer. Spielen und Lernen am PC. Heidelberg: Springer. Hofmann, G. (Hrsg.) (2001). Auf dem Weg zu Basel II. Konzepte, Meinungen, Modelle (1. Aufl.). Frankfurt/Main: Bankakademie Verlag. Hofmann, S.G. (1999). Introducing the grandmother test into psychological science. Journal of Theoretical and Philosophical Psychology, 19, 167-176. Holland, J.L. (1985). Making vocational choices. A theory of vocational personalities and work environments. Englewood-Cliffs, NJ: Prentice-Hall. Holzinger, A., Pichler, A., Aimer, W., & Maurer, H. (2001). TRIANGLE: A multi-Media test-bed for examining incidental learning, motivation and the Tamagotchi-Effect within a Game-Show like Computer Based Learning Module. Proceedings of EDMEDIA 2001, Tampere, AACE, Charlottesville, VA (2001), 766-771. Howell, J.M., & Hall-Meranda, K.E. (1999). The ties that bind: The impact of LeaderMember Exchange, transformational and transactional leadership, and distance on predicted follower performance. Journal of Applied Psychology, 84(5), 680-694. Institut fur Test- und Begabungsforschung (Hrsg.) (1988a). Test fur medizinische Studiengange (2. Auflage). Gottingen: Hogrefe. Institut fur Test- und Begabungsforschung (Hrsg.) (1988b). Der neue TMS. Originalversion des Tests fur medizinische Studiengange im besonderen Auswahlverfahren (2. Auflage). Gottingen: Hogrefe. Irle, M., & Allehoff, W. (1984). Berufs-Interessen-Test II (B-I-TII). Gottingen: Hogrefe. Ives, W., Torrey, B., & Gordon, C. (1998). Knowledge Management: An emerging discipline with a long history. Journal of Knowledge Management 1, 4, 269 - 274. Jain, L.C., Chen, Z., & Ichalkaranje, N. (2002). Intelligent agents and their applications. Heidelberg: Physica. Jaynes, J. (1993). The origin of consciousness in the breakdown of the bicameral mind. London: Penguin. Joreskog, K.G., & Sorbom, D. (1989). LISREL 7: User's reference guide. Chicago, IL: Scientific Software International. Jonassen, D.H. (1993). Effects of semantically structured hypertext knowledge bases on users knowledge structures. In C. McKnight, A. Dillon & J. Richardson (Eds.), Hypertext: A psychological perspective. London: Ellis Norwood. Jonassen, D.H., Beissner, K., & Yacci, M. (1993). Structural knowledge. Techniques for representing, conveying, and acquiring structural knowledge. Hillsdale, NJ: Lawrence Erlbaum. Johnson-Laird, P.N. (1983). Mental models. Cambridge: Cambridge University Press. Johnston, P., & Moretti, C. (1998). Worldwide and U.S. IT training and education markets and trends. New York: International Data Corporation (IDC). Kadishi, B. (1998). Anerkennung und Validierung (aufier-)beruflich erworbener Kompetenzen am Beispiel der Familien- und Hausarbeit. Grundlagen der Weiterbildung, 9(5), 217-220.
Bibliography
277
Kapplinger, B. (2002). Anerkennung von Kompetenzen: Definitional, Kontexte und Praxiserfahrungen in Europa. Bonn: Deutsches Institut fur Erwachsenenbildung. Kahlert, H. (2002). Bilanzierung und Bewertung bei Unternehmenszusammenschliissen nach US-GAAP. Hochschule fur Bankwirtschaft - Private Fachhochschule der Bankakademie e.V., Frankfurt a. M. Kahneman, D. & Tversky, A. (1984). Choices, values, and frames. American Psychologist, 39, 341-350. Kahneman, D. & Tversky, A. (1979). Prospect theory: An analysis of decisions under risk. Econometrica, 47, 263-291. Kaiser, A. (1998). Carte de competence: Wie lassen sich Kompetenzen feststellen? Grundlagen der Weiterbildung, 9(5), 199-201. Kaplan, R., & Norton, D. (1997). Balanced Scorecards. Strategien erfolgreich umsetzen. Stuttgart. Katz, M.R. (1993). Computer-assisted career decision making. Hillsdale, NJ: Erlbaum. Keil-Slawik, R., & Selke, H. (1998). Der Aufbau von lernforderlichen Infrastrukturen. Bibliothek 22(1), 51-59. Kelley, J. (1988). Limits of intelligent tutoring systems. International Journal of Educational Research, 12(8), 863-873. Kerka, S. (1987). Adult career counseling: An interactive model. Overview. ERIC Digest No. 65 (access number ED289996). Kiesler, S. (1992). Talking, teaching, and learning in network groups: Lessons from research. In Kaye, A.R. (Ed.), Collaborative learning through computer conferencing. Heidelberg: Springer, 145-165. Kim, D.H. (1993). Creating Learning Organizations: Understanding the Link Between Individual and Organizational Learning. MIT Sloan School of Management, 1-33. Klein, R.G. & Blake, E. (2003). The dawn of human culture. Hoboken, NJ: Wiley. Klemme, M., & Hasebrook, J. (1999). Web-based training in banking and finance: Requirements and strategies. Submitted to WebNet 99. Charlottesville: AACE. Kluge, A., & Schilling, J. (2000). Organisationales Lernen und Lernende Organisation ein Uberblick zum Stand von Theorie und Empirie. Zeitschrift fur Arbeits- und Organisationspsychologie, 44(4), 179-191. Kopp, O. (1999). Wissensmanagement in Banken: Bestandsaufftahme und Konzeption eines Wissensmanagementmodells am Beispiel des Firmenkundengeschafts. Diplomarbeit, Hochschule fur Bankwirtschaft, Frankfurt. Kosko, B. (1992). Neural Networks for Signal Processing. Englewood-Cliffs, NJ: Prentice-Hall. Kosslyn, S.M., & Koenig, O. (1995). Wet mind. The new cognitive neuroscience. NN: Simon & Schuster. Kramer, S. (1988). Symbolische Maschinen. Die Idee der Formalisierung in geschichtlichem Abriss. Darmstadt: Wissenschaftliche Buchgesellschaft. Kroeber-Riel, W. (1996). Bildkommunikation. Imagerystrategien fur die Werbung. Munchen: Vahlen. Kredelbach, T. (1998). Weiterbildungscontrolling in der Kreditwirtschaft. Die Bank 7/98. Kuhn, T.S. (1970) The structure of scientific revolutions. (2nd ed.). Chicago: University of Chicago Press.
278
Learning Support Systems
Ktthn, I.-W. (2002). Human Kapital in Unternehmen: MOglichkeiten der Bewertung des Human Kapitals bei Fusionen. Hochschule fur Bankwirtschaft - Private Fachhochschule der Bankakademie e.V., Frankfurt a. M. Kulik, C.-L., & Kulik, J.A. (1991). Effectiveness of computer-based instruction: An updated analysis. Computers in Human Behavior, 7, 75-94. Kulik, J.A., Bangert-Downs, R.L., & Williams, G.W. (1983). Effects of computer-based teaching on secondary school students. Journal of Educational Psychology, 75, 1926. Kulik, C.-L., Kulik, J.A., & Cohen, P. (1980). Effectiveness of computer-based college teaching: A meta-analysis of findings. Review of Educational Research, 50, 252544. Kunz, G.C., Drewniak, U., & Schott, F. (1994). On-line and off-line assessment of selfregulation in learning from instructional text. Learning and Instruction, 2, 287-302. Kurzweil, R. (1990). The age of intelligent machines. Cambridge: MIT Press. Kurzweil, R. (1999). The age of spiritual machines. New York: Penguin Putnam. Lajoie, S., & Deny, S.J. (1993). Computers as cognitive tools. Hillsdale, NJ: Lawrence Erlbaum. Landy, F J. (1985). Psychology of Work Behavior. Homewood: Lane, R.D., Nadel, L., & Ahem, G. (2000). Cognitive neuroscience of emotion. Oxford: Oxford Univerity Press. Larkin, J.H., & Chabay, R.W. (Eds.) (1992). Computer assisted instruction and intelligent tutoring systems: Shared goals and complementary approaches. Hillsdale, NJ: Erlbaum. Lee, T.B. (1999). Weaving the web. San Francisco: Harper Collins. Lee, T.B. (2001). The semantic web. Scientific American, May 2001. Lennon, J. (1997). Hypermedia Systems and Applications; Springer Pub. Co., Germany. www.gentle-wbt.com Lennon, J., & Maurer, H. (1995). Digital Libraries as Learning and Teaching Support. J.UCS 1/11, 719-727; Proc. ICCE'95, Singapore (Dec. 1995), AACE, Charlottesville, USA, 28-33. Lennon, J., & Maurer, H. (2001). MUSLI: A hypermedia interface for dynamic, interactive and symbolic communication. JNCA 24, 4, 273-292. Levie, H.W., & Lentz, R. (1982). Effects of text illustration: A review of research. Educational Communication and Technology Journal, 30, 195-232. Levin, J.R., Anglin, G.J., & Carney, R.N. (1987). On empirically validating functions of pictures in prose. In D.M. Willows & H.A. Houghton (Eds.), The psychology of illustration: Vol. 1 basic research (51-85). New York: Springer. Livingstone, S. (1990). Making sense of television. The psychology of audience interpretation. Oxford: Pergamon Press. Lochel, H. (2000). Die okonomischen Dimensionen der "New Economy". Arbeitsberichte der Hochschule fur Bankwirtschaft, 25, Frankfurt/Main. Lukesch, H. (1989). Video im Alltag der Jugend: Quantitative und qualitative Aspekte des Videokonsums, des Videospielens und der Nutzung anderer Medien bei Kindern, Jugendlichen und Erwachsenen. Regensburg: Roderer.
Bibliography
279
Maass, W. (1997). Networks of Spiking Neurons: The Third Generation of Neural Network Models. Graz: Institute for Theoretical Computer Science, TU Graz. Marchionini, G., & Maurer, H. (1995). The roles of digital libraries in teaching and learning. C.ACM 38, 4, 67-75. Martin, R. L., & Moldoveanu, M. C. (2003). Capital versus talent. The battle that's reshaping business. Harvard Business Review, July 2003, 36-41. Maurer, H. (1996). HyperWave: The Next Generation Web Solution. London: AddisonWesley. Maurer, H. (1998). Using the WWW System Hyperwave as the Basis of a General Networked Teaching and Learning Environment, CIT, 6(1), 63-72. Maurer, H. (2001). Beyond classical digital libraries Global Digital Library Development in the New Millenium. Proceedings NIT Conference, Beijing, Tsinghua University Press, 165-173. Maurer, H. (2003). Necessary Aspects of Quality eLearning Systems; Proc. Conference on Quality Education @ a Distance, Deakin University, Geelong, Australia, to appear. Maurer, H., & Oliver, R.: A scenario for the computer 2010 and its implications for society; forthcoming paper, consult www.iicm.edu/maurer for details. Maurer, H., & Pivec, M. (2001). Situation Learning: A New Approach to Knowledge Mediation. Proceedings of ED-MEDIA 2001, Tampere, AACE, Charlottesville, VA, 1254-1259. Maurer, H., Rozsenich, N., & Sapper, M. (1999) How to Make Discussion Forums Work on the WWW, Proc. WebNet 99, AACE, Charlottesville, USA, 717-722 Maurer, H., Stubenrauch, R., & Camhy, D. (2003). Foundations of MIRACLE Multimedia Information Repository: A Computer-based Language Effort; J.UCS 9, 4, to appear. Maurer, H., & Tochtermann, K. (2002). On a New Powerful Model for Knowledge Management and its Applications. JUCS 8, 1, 85-96. Mayer, R.E. (1989). Models for understanding. Review of Educational Research, 59(1), 43-64. Mayer, R.E. (1992). Cognition and Instruction: Their historic meeting within educational psychology. Journal of Educational Psychology, 84(4), 405-412. Mayer, R.E., & Anderson, R.B (1991). Animations need narrations: An experimental test of a dual-coding hypothesis. Journal of Educational Psychology, 83(4), 484-490. Mayer, R.E., & Anderson, R.B. (1992). The instructive animation: Helping students build connections between words and pictures in multimedia learning. Journal of Educational Psychology, 84(4), 444-452. Mayer, R.E., & Gallini, J.K. (1990). When is an illustration worth ten thousand words? Journal of Educational Psychology, 82(4), 715-726. Mayer, R.E., & Sims, V.K. (1994). For whom is a picture worth a thousand words? Extensions of a dual-coding theory of multimedia. Journal of Educational Psychology, 86(3), 389-401. McDermott, D., & Poggio, T. (1989). Artificial Intelligence. In J.E. Hopcroft & K.W. Kennedy (Eds.), Computer Science Achievements and Opportunities, Report of the
280
Learning Support Systems
NSF Advisory Committee for Computer Research (41-50). Philadelphia: Society for Industrial and Applied Mathematics. McGraw, K.L. (1994). Performance support systems: Integrating AI, hypermedia, and CBT to enhance user performance. Journal of Artificial Intelligence in Education, 5(1), 3-26. Meersman, R., Tari, Z., & Stevens, S. (Eds.) (1999). Database Semantics. Kluwer Academic Publishers, USA. Mehler, J., & Franck, S. (1995). Cognition on cognition. Cambridge, MA: M.I.T. Press. Meier, H. (1995). Strategische Planung und Personalentwicklung in Banken und Sparkassen. Wiesbaden: Gabler. Menzel, R. (1996). Neuronale Plastizitat, Lernen und Gedachtnis. In J. Dudel, R. Menzel & R.F. Schmidt (Eds.), Neurowissenschaft. Vom Molekiil zur Kognition (485518). New York: Springer. Michalski, R. (1986). Understanding the nature of learning. In R. Michalski, J. Carbonell, & T. Mitchell (Eds.), Machine learning - an artificial intelligence approach (Vol. II, 471-498). Los Altos, CA: Morgan Kaufmann. Miller, G.A., Galanter, I., & Pribram, K.H. (1960). Plans and the structure of behavior. New York: Holt. Minkowski, M. (1963). On Aphasia in Polyglots. In L. Halpern (Ed.), Problems of Dynamic Neurology. Hadassah University, Jerusalem, 119-161. Moravec, H. (1998). When will computer hardware match the human brain? Journal of Transhumanism, 1, 9-21. Moravec, H.P. (2000). Robot: Mere Machine to Transcendent Mind. Oxford; Oxford Univerity Press. Morik, K. (1993). Maschinelles Lernen. In G. Gorz (Ed.), Einfuhrung in die kilnstliche Intelligenz (247-301). Bonn: Addion-Wesley. Moormann, J. (1999). Umbruch in der Bankinformatik - Status Quo und Perspektiven fur eine Neugestaltung. In J. Moormann, & T. Fischer (Eds.), Handbuch Informationstechnologie in Banken. Wiesbaden: Gabler, 3-20. Myers, B.A. (1998). A brief history of human-computer interaction technology. Interactions, 5(2), 44-54. Nagel, S. (Ed.) (1993). Computer-aided decision analysis: Theory and application. Westport, CT: Greenwood. Neumann, J. von, & Morgenstern, O. (1967). Theory of games and economic behavior (3rd ed.) New York: Wiley. Nichols, J. (1992). Linguistic Diversity in Space and Time. Chicago: Chicago University Press. Nonaka, 1.(1994). A dynamic theory of organizational knowledge creation. Organization Science, 5(1), 14-37. North, K., Probst, G., & Romhardt, K. (1998). Wissen messen - Ansatze, Erfahrungen und kritische Fragen. Zeitschrift fur Ftlhrung + Organisation^), 158-166. Och, F.J. (2003). Statistical machine translation: from single-word models to alignment templates. Technical Report 2003-06. Aachen: RWTH Aachen. OECD (2001). The new economy beyond the hype. The OECD growth project. Paris: Organization for Economic C-Operation and Development (OECD).
Bibliography
281
OECD (2003). ICT and economic growth. Evidence from OECD countries, industries and firms. Paris: Organization for Economic C-Operation and Development (OECD). Ogata, H., & Yano, Y. (1998). Supporting awareness for augmenting participation in collaborative learning. Proceedings of the WebNet98. Charlottesville: AACE., 1040-1045. Ohsuga, S. (Ed.) (1992). Information modeling and knowledge bases III: Foundations, theory and applications. Tokyo: IOS Press. Oliver, R. (2001). Seeking best practice in online learning: Flexible Learning Toolboxes in the Australian VET sector. Australian Journal of Educational Technology, 17, (2), 204-222. Oliver, R., Wirski, R., Omari, A., Hingston, P., & Brownfield, G.(2003). Exploring the reusability of Web-based learning resources. Proc. of ED-MEDIA 2003, to appear. Ortony, A. (Ed.) (1993). Metaphor and thought. Cambridge, MA: Cambridge University Press. Paivio, A. (1986). Mental representations: A dual-coding approach. New York, Oxford: Oxford University Press. Paradis, M. (1977). Bilingualism and Aphasia. In H. Whitaker & H. Whitaker (Eds.), Studies in Neurolinguistics. (3), New York: Academic Press, 65-121. Paradis, M., Goldblum, M.C., Abidi, R. (1982). Alternate Antagonism with Paradoxical Translation Behaviour in Two Bilingual Aphasic Patients. In Brain and Language 15,55-69. Park, O., & Tennyson, R.D. (1983). Computer-based instructional systems for adaptive education: A review. Contemporary Education, 2, 121-135. Paul, S., & Siewert, K.-J. (1996). Bank-Controlling 1: Ertragsmanagement in Kreditinstituten. Frankfurt: Bankakademie Verlag. Pfurtscheller, G., Neuper, C, Guger, C , Harkam, W., Ramoser, H., Schlogl, A., Obermaier, B., & Pregenzer, M. (2000). Current Trends in Graz Brain-Computer Interface Resarch; IEEE Transactions Rehab. Engineering 8, (2), 216-219. Piaget, J. (1974). Weisheit und Illusion in der Philosophic Frankfurt/Main: Suhrkamp. Piaget, J. (1976). The Grasp of Consciousness: Action and Concept in the Young Child. Harvard, MA: Harvard University Press. Pichler, M. (1996). Der Ertrag der Weiterbildung laBt sich rechnen. Dusseldorf: Handelsblatt v. 23.08.1996. Picking, R. (1994). A comparative study of computer-based document manipulation techniques. Hypermedia, 6, 87-100. Picot, A., Reichwald, R., & Wigand, R. (1998): Die Grenzenlose Unternehmung: Information, Organisation und Management (3. Aufl.). Wiesbaden: Gabler. Plowman, L. (1994). The 'primitive mode of representation' and the evolution of interactive multimedia. Journal of Educational Multimedia and Hypermedia, 3(4), 275-293. Prahalad, C.K., & Hammond, A. (2002). Serving the world's poor, profitably. Harvard Business Reviews, September 2002, 48-57.
282
Learning Support Systems
Press, W.H., Flannery, B.P., Teukolky, S.A., & Vetterlin, W.T. (1989). Numerical recipes in Pascal. The art of scientific computing. Cambridge: Cambridge University Press. Probst, G., & Raub, S.P. (1998). Kompetenzorientiertes Wissensmanagement. Zeitschrift fur Fuhnmg + Organisation^), 132-138. Probst, G., Raub, S., & Romhardt, K. (1998). Wissen managen. Wie Unternehmen ihre wertvollste Ressource optimal nutzen. Wiesbaden. Pulic, A. (1998). Measuring the Performance of Intellectual Potential in Knowledge Economy. Austrian IC Research Center. Retrieved July 03, 2003, from the World Wide Web: http://www.measuring-ip.at/OPapers/PulicA^aictxt/vaictxt.html Ramachandran, V.S., & Blakeslee, S. (1998). Phantoms in the brain. Probing the misteries of the human mind. New York: William Morrow. Rauterberg, M. (1999). Different effects of auditory feedback in man-machine interfaces. In: N.A. Stanton & J. Edworthy (Eds.), Human factors in auditory warnings (pp. 225-242). Aldershot (UK): Ashgate Publishing. Retterer, OJ. (1991). Learning from a hypertext: The effect of reading interactive text containing non-sequential, associative linkages on comprehension. UMI 9200771, Dissertation: University of Toledo. Reynolds, S.B., & Dansereau, D.F. (1990). The knowledge hypermap: An alternative to hypertext. Computers & Education, 14,409-416. Rice, M.L., Huston, A.C., & Wirgt, J.C. (1986). Replay as repetitions: Young children's interpretation of television forms. Journal of Applied Developmental Psychology, 7, 61-76. Richman, W.L., Kiesler, S., Weisband, S., & Drasgow, F. (1999). A meta-analytical study of social desirability distortion in computer-administrated questionnaires, traditional questionnaires, and interviews. Journal of Applied Psychology, 84 (5), 754-775. Rieber, L.P. (1990). Using computer animated graphics in science instruction with children. Journal of Educational Psychology, 82(1), 135-140. Rieber, L.P. (1991). Animation, incidental learning, and continuing motivation. Journal of Educational Psychology, 83(3), 318-328. Rinck, M., & Glowalla, U. (1995). Was niltzt Multimedia? Eine empirische Effektivitatsprtlfung bei der Darstellung quantitativer Daten. In E. Schoop, R. Witt & U. Glowalla (Eds.), Hypermedia in der Aus- und Weiterbildung. Dresdner Symposium zum computerunterstiitzten Lernen (107-117). Konstanz: Universitatsverlag Konstanz. Rollandelli, D. (1989). Children and television: the visual superiority effect reconsidered. Journal of Broadcasting and Electronic Media, 33, 69-81. Romer, P. (1993). Beyond the knowledge worker (ideas and things). The Economist, 11.09.1993. Rosenberg, R. (1990). A critical analysis of research on intelligent tutoring systems. In The Education Technology Anthology Series: Expert systems and intelligent computer-assisted instruction (184-190). Englewood Cliffs, NJ: Educational Technology Publications.
Bibliography
283
Rosenzweig, J.A. (1998). Winning the global game. A strategy for linking people to profits. New York: Free Press. Salmon, G. (2002). Future learning encounters. Keynote presentation EUROCALL 2002. Centre for Innovation, Knowledge & Enterprise. Milton Keynes (UK): Open University Business School. Salomon, G. (1984). Television is 'easy' and print is 'tough': The differential investment of mental effort in learning as a function of perceptions and attribution. Journal of Educational Psychology, 76, 647-658. Schein, E.H.(1988). Organizational Psychology. Englewood. Schooling, H. (1998). Neue partizipative Momente und Strukturen bei Methoden und Verfahren der Kompetenzbestimmung. Grundlagen der Weiterbildung, 9(5), 205207. Schulmeister, R. (1996). Grundlagen hypermedialer Lernsysteme. Theorie, Didaktik, Design. Bonn: Addison-Wesley. Schulz von Thun, F. (1998). Miteinander reden 1. Reinbeck. Scott, P. (1983). Learning: The construction of posteriori knowledge structures. Proceedings of the AAAI-83, Washington. Seifert, K.H. (1994). Improving prediction of career adjustment with measures of career development. The Career Development Quarterly, 42, 353-366. Self, J. (1992). Unrevealing the learner model mystery. In F.L. Engel & D.G. Bouwhuis (Eds.), Cognitive modeling and interactive environments in language learning (1725). Heidelberg: Springer. Shannon, C.E., & Weaver, W. (1949). The mathematical theory of communication. Urbana, IL: University of Illinois Press. Sheperd, G.M. (Ed.) (1994). Neurobiology. Oxford, New York: Oxford University Press. Sheperd, G.M. (Ed.) (1998). The synaptic organization of the brain. Oxford, New York: Oxford University Press. Siegert, T. (1999). Humankapital - Erfolgsmessung und Partizipation. In W. Btihler & T. Siegert (Eds.), Unternehmenssteuerung und Anreizsysteme Kongrefidokumentation / 52. Deutscher Betriebswirtschafter-Tag 1998 (17-46). Stuttgart: Schaffer-Poeschel. Skyrme, David J. (1999). From Measurement Myopia to Knowledge Leadership. London: David Skyrme Associates. Software Publishers Association (1995). Report on the effectiveness of technology in schools, 1990-1994. Washington, DC: SPA. Sonnenberg, H.-G. (1993). Computerunterstiitzte psychologische Diagnoseverfahren bei der Auswahl von Fuhrungskraften. Zeitschrift fur Arbeits- und Organisationspsychologie, 37(3), 146-149. Sproull, L., & Kiesler, S. (1991). Connections: New ways of working in the networked organization. Cambridge, MA: MIT Press. Stevens, J. (1992). Applied Multivariate Statistics for the Social Sciences. Hillsdale, NJ: Erlbaum. Stone, P. (2000). Layered learning in multiagent systems. A winning approach to robotic soccer. Cambridge: Bradford / MIT.
284
Learning Support Systems
Sullivan, P.H. (2000). A brief history of the ICM movement. Retrieved July 11, 2003, from the World Wide Web: http://www.sveiby.com/articles/icmmovement.html Sveiby, K.-E. (2001). Intellectual Capital and Knowledge Management. www.sveiby.com/articles/intellectualcapital.html Sveiby, K.-E. (2001). The "Invisible" Balance Sheet. www.sveiby.com/articles/InvisibleBalance.html Sweetland, R.C., & Keyer, D.J. (1984). Test - a city. Kansas City: Test Cooperation of America. Tesluk, P.E., & Mathieu, J.E. (1999). Overcoming roadblocks to effectiveness: Incorporating management of performance barriers into models of workgroup effectiveness. Journal of Applied Psychology, 84(2), 200-217. Thuraisigham, B. (2003). XML databases and the semantic web. Boca Raton, FL: CRC Press. Todt, E. (1967). Differentieller Interessentest (DIT). Bern: Huber. Tsang, E. W. K. (1997). Organizational learning and the learning organization: A dichotomy between descriptive and prescriptive research. In Human Relations, 50, (1), 73-89. Tuomi, I. (1999). Corporate knowledge: Theory and practice of intelligent organizations. Oxford: Oxford University Press. Tuomi, I. (2003): Networks of innovation: Change and meaning in the age of the Internet. Oxford: Oxford University Press. Turing, A.M. (1950). Computing Machinery and Intelligence. Mind, 59(1), 433-460. Ueckert, H. (1995). Expertensysteme. In W. Sarges (Ed.), Management-Diagnostik (2nd ed.), (789-796). Gottingen: Hogrefe. Van den Berg, S., & Watt, J.H. (1991). Effects of educational setting on student responses to structured hypertext. Journal of Computer-Based Instruction, 4(18), 118-124. Walter, H. (2001). Neurophilosophy of Free Will: From Libertarian Illusions to a Concept of Natural Autonomy. Cambridge, MA: M.I.T. Press. Walz, B.R., Bleuer, J.C., & Maze, M. (1989). Counseling Software Guide. Alexandria, VI: AACD. Wardrip-Fruin, N. & Montfort, N. (2003). The new media reader. Cambridge: MIT Press. Weber, G. (1995). Individual selection of examples in an intelligent learning environment. Journal of Artificial Intelligence in Education, 7(1), 3-31. Weick, K. E., & Westley, F. (1996). Organizational learning: Affirming an oximoron. In S. R. Clegg, C. Hardy, & W. R. Nord (eds.), Handbook of Organization studies., (440-458). London: Sage. Weidenmann, B. (1988). Psychische Prozesse beim Verstehen von Bildern. Bern: Hans Huber. Weidenmann, B. (1994). Wissenserwerb mit Bildern. Bern: Hans Huber. Weisband, S., & Atwater, L. (1999). Evaluating self and others in electronic and face-toface groups. Journal of Applied Psychology, 84(4), 632-639. Weiser, M. (1991). The computer for the twenty-first century, Scientific American, Sept. 1991, 94-100.
Bibliography
285
Weiss, D.J., & Vale, CD. (1987). Adaptive testing. Applied Psychology - an International Review, 36, 249-262. Weiss, S.W., & Kulikowski, C.A. (1991). Computer systems that learn: Classification and prediction methods from statistics, neural nets, machine learning, and expert systems. San Mateo, CA: Morgan Kaufmann. Wells, S. (2003). The journey of man: A genetic odyssey. Princeton, NJ: Princeton University Press. Wiener, N. (1961). Cybernetics (2nd ed.). Cambridge, MA: The M.I.T. Press. Williams, S.M. (2001). Is a company's intellectual capital performance and intellectual capital disclosure practices related?: Evidence from publicly listed companies from the FTSE 100. Paper presented at the 4th World Congress on Intellectual Capital, Hamilton. Wolf, K., Hasebrook, J., & Rinck, M. (1999). Wand oder keine Wand? Die Reprasentation raumlicher Veranderungen in Situationsmodellen. Zeitschrift fur Experimentelle Psychologie, 46(3), 152-163. Woolridge, M. (2000). Reasoning about rational agents. Cambridge: Bradford / MIT. Zimmermann, B. (1990). Self-regulated learning and academic achievement: An overview. Educational Psychologist, 25, 3-17.
This page is intentionally left blank
Index
Active Documents 85, 105 ADL 205 Africa 24 Agent Intelligent-Agent 220 Agent Multi-Agent-Platform 202, 204, 237 Agent Rational 202 AI 155,201 AICC 205 Aphasia 229 Ariadne 205 Asia 22 Augmented Reality 216 BDI Model 202 Bible 226, 237 Blended Learning 125 Brain Drain 18 Career Counseling 182 Career Decision 161 Carreer Counseling 181 CBT 6 CI 155 Civilization 253 Clinical Psychology 32 CMC 140,226 COBOL 201 Cog 201 Competence Management 34 Computational Intelligence 227 Conditioning Circuit 231 Conflict Management 52 Cultural Differences 27 Customer Capital 57 Cybernetics 235 Developing Countries 18 Development Costs 67 Double Loop Learning 45 DRM 13 Dual Encoding 180 Ecology 254
Educational Controlling 34 E-Government 5 E-Learning Benefits 8 E-Learning Markets 14 E-Learning Perspectives 14 ERP 122 EVAMVA 60 Evolution 260 FIPA 205, 237 FORTRAN 201 G7 Countries 9 Gesture Recognition 238, 246 Globalization 19 GNP 19 GPS 241, 248 HR XML Consortium 48 HRD 68 HR-SEP 205 Human Capital 57 Human Culture 255 IAS 58 ICT 7 ICT Investments 9, 37 ICT vs. Didactics 28 IEEELTSC 205 Illiad 257 IMS 205 Information Theory 235 Intangible Assets 57 Intellectual Capital 57 Intelligent Robotics 201 Invisible Assets 57 ITS 28, 93, 155, 159, 180 Knowledge Management 46 Knowlegde Based Economy 46 Language 235 Language Natural 107 Learning Culture 119 Learning Organization 31, 42 Linguistics 256
288 LMS 122 LMX 55, 131 LOM 49 LSS 122, 126 Machine Learning 201 MANOVA 186 Mental Models 228 Meta Cognition 96 Meta-Analysis 211 MIPS 225 MIT 219 M-Learning 12 Moore's Law 5 Multimedia 41 Neural Networks 220,231 Neurobiology 230 New Economy 39 Nobel Prize 31 NSA 237 OECD 7 OLTP 209 Ontology 107,219 PDA 248 Personnel Budget 66 Poverty Premium 19 PSS 68, 93 Rating Standards 59 RDF 210 Reinforcement Learning 202 Robotics 236
Learning Support Systems School Connectivity 10 SCORM 205 Semantic Web 204 SGML 208 SHOE 218 Skandia Navigator 57 Skills Management 47, 48 SME 131, 140 Social Theory 53 Speech Recognition 236 Speech Synthesis 236 Spiking Neurons 231 Theories X and Y 53 TOC 38 Training Investment 77 TTO 67 Ubiquitous Computing 216 UMUC 121 Validity 76 Value Creation 59 Value Extraction 59 Virtual Keyboard 245 Virtual Reality 215 VPN 12 WACC 60 War For Talent 62 World Bank 17 XML 208 XMLQL 219 XSL 219
Learning Support Systems for Organizational Learning The major trends in e-learning are determined by the global demand of academic, elderly and nontraditional target groups for training and education. The advent of the learning organization reflects these major shifts of the educational markets within companies. Automation of learning processes does not enhance a company's productivity; augmentation of individual and collaborative learning processes is needed. This book reflects seven years of applied research (1997-2003) in the fields of adaptive multimedia systems, knowledge-based and collaborative learning environments, and intelligent software agents.
ISBN 981-238-831-1
World Scientific www.worldscientific.com 5529 he
9 "789812"388315"