Excluding Violent Youths from Juvenile Court: The Effectiveness of Legislative Waiver (Criminal Justice: Recent Scholarship)

Criminal Justice Recent Scholarship

Edited by Marilyn McShane and Frank P. Williams III

A Series from LFB Scholarly

This page intentionally left blank

Excluding Violent Youths from Juvenile Court

The Effectiveness of Legislative Waiver

David L. Myers

LFB Scholarly Publishing LLC New York 2001

Copyright (c) 2001 by LFB Scholarly Publishing LLC All rights reserved. Library of Congress Cataloging-in-Publication Data Myers, David L., 1968Excluding violent youths from juvenile court : the effectiveness of legislative waiver / David L. Myers. p. cm. -- (Criminal justice) Includes bibliographical references and index. ISBN 1-931202-02-8 1. Juvenile justice, Administration of--United States. 2. Prosecution--United States--Decision making. 3. Juvenile recidivists--United States. 4. Violent offenders--United States. I. Title. II. Criminal justice (LFB Scholarly Publishing LLC) KF9812 .M97 2001 345.73'08--dc21 2001000674

ISBN 1-931202-02-8 Printed on acid-free 250-year-life paper. Manufactured in the United States of America.

Dedication

To my wife, Nicole, who has supported me in more ways than I can count, and our daughter, Abigail, my new source of inspiration.

This page intentionally left blank

Table of Contents

List of Tables

ix

Acknowledgments Introduction

xi

1

Deficiencies in Transfer Research The Current Research 10

9

Chapter 1: Cracking Down on Youth Violence Origins of Juvenile Justice 12 Criticisms and Reforms 15 The Surge in Youth Violence 18 Methods of Transfer 21 Transfer Use and Rationale 26 Juvenile Transfer in Pennsylvania 29

11

Chapter 2: The Deterrent Effects of Formal Sanctions 33 Origins of Deterrence Theory 34 Deterrence Research 36 The Specific Deterrent Effect of Formal Sanctions 40 Summary 55 Chapter 3: Punishment and Deterrence Through Treating Juvenile Offenders as Adults 57 Punishment Certainty 58 Punishment Severity 62 Punishment Celerity 66 Deterrent Effects 68 Punishment and Deterrence: Is Transfer Effective? 72 The Present Study 74

vii

viii

Table of Contents

Chapter 4: Research Agenda Hypotheses 77 Data and Methods 83

77

Chapter 5: Quantitative Findings Bivariate Results 109 Multivariate Results 119 Summary 161

109

Chapter 6: Qualitative Findings 167 Overall Effectiveness of Legislative Waiver 167 Legislative Waiver Versus Judicial Waiver: Strengths and Weaknesses 171 Reverse Waiver to Juvenile Court 174 Evidence of Deterrence 176 Alternatives to Treating Violent Youths as Adults 179 Summary 180 Chapter 7: Discussion and Conclusions 183 Evidence of a "Custody Gap" 183 No Evidence of a "Leniency Gap" 184 Little Evidence of Deterrence 186 Policy Implications 190 Recommendations for Future Research 198 References Index

223

201

List of Tables

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

Descriptive Statistics for All Variables 90 Descriptive Statistics by Transfer Status 91 Bivariate Correlations Among Variables 110 Logistic Regression Estimates for the Determinants of Release 121 Logistic Regression Estimates for the Determinants of Conviction 124 Logistic Regression Estimates for the Determinants of Target Conviction 126 Logistic Regression Estimates for the Determinants of Incarceration 129 OLS Regression Estimates for the Determinants of the Natural Log of Incarceration Length 131 OLS Regression Estimates for the Determinants of the Natural Log of Case Processing Time 134 Logistic Regression Estimates for the Determinants of Pre-Dispositional Arrest 137 Logistic Regression Estimates for the Determinants of Pre-Dispositional Violent Felony Arrest 141 Logistic Regression Estimates for the Determinants of Post-Dispositional Arrest 144 Logistic Regression Estimates for the Determinants of Post-Dispositional Violent Felony Arrest 150 Cox Regression Estimates for the Determinants of Post-Dispositional Survival Time 156

ix

This page intentionally left blank

Acknowledgments

There are a number of people I would like to thank for the help they provided during the completion of my dissertation and this subsequent book. To begin, my dissertation committee at the University of Maryland supplied outstanding guidance while I was completing my research and graduate education. As chair, Raymond Paternoster read numerous drafts, provided useful comments and suggestions, offered words of encouragement, and was a valuable source of humor and friendship. The other members of my committee, Alan Neustadtl, Lawrence Sherman, Sally Simpson, and Charles Wellford, all willingly contributed their time, knowledge, and expertise, and their remarks and critiques greatly improved the quality of my work. Of course, any weaknesses that remain should be attributed to me. I also would like to recognize The Center for Juvenile Justice Training and Research, located at Shippensburg University, for granting me access to their data. More specifically, my thanks go out to Clay Yeager (Director of The Center), Linda Bender, and John Lemmon for the time and assistance they provided. Without their cooperation, completing this study would not have been possible. Finally, I would like to acknowledge several other individuals who have contributed in various ways to the completion of this book. Kraig Kiehl, my graduate assistant, spent countless hours working on the formatting, the index, and proofreading. Dennis Giever, Department of Criminology Chairperson, provided much technological advice and support, and he has worked extremely hard to create an environment that is conducive to scholarly activity. My parents, Leon and Gloria, continue to set the example that I try to follow. Most importantly, I am deeply grateful to my wife, Nicole, whose love and support are my foundation, and our daughter, Abigail, who makes loving life easy. Indiana, Pennsylvania

October 2000

xi

This page intentionally left blank

Introduction

A certain amount of delinquency has been a traditional mark of adolescence. However, during the past decade, youth violence emerged as a major public concern (Biskup & Cozic, 1992; Howell, 1997; Howell, Krisberg, Hawkins, & Wilson, 1995; Moore & Tonry, 1998; Snyder & Sickmund, 1995). Driving this concern was both the rising level and the lethality of violent acts committed by and against young people. The juvenile violent crime arrest rate surged from the mid1980s until the mid-1990s, increasing 70% from 1987 to 1994 (Sickmund, Snyder, & Poe-Yamagata, 1997, p. 18). In particular, the murder arrest rate of juveniles exhibited dramatic growth by doubling between 1987 and 1993 (Sickmund et al., 1997, p. 20). These increases corresponded with a sizable upswing in violent crime victimization among youths (Snyder & Sickmund, 1995, pp. 23-26), and firearm use appeared to be a key factor in these trends (Blumstein, 1995b; Cook & Laub, 1998; Fagan & Wilkinson, 1998; Greenbaum, 1997). Despite recent national decreases in juvenile violent crime rates, current levels remain higher than those of the mid-1980s (Sickmund et al., 1997; Snyder, 1997), and youth violence continues to receive a considerable amount of public attention (see, e.g., Masters & Shear, 1998; “More Teens Report,” 1998). News accounts routinely depict seemingly senseless acts of lethal violence, involving gangs, guns, drugs, or all three (see, e.g., Fields & Overberg, 1998; McCoy, Lowe, & Gibbons, 1998). A reader or viewer is often left with the perception that young people are out of control, the juvenile system is ineffective, and something must be done to deal with these violent youthful offenders. In response to this problem, many changes have been carried out in juvenile justice systems across the country. In contrast to the 1

2

Excluding Violent Youths From Juvenile Court

conventional juvenile court’s emphasis on “child-saving” and serving the “best interests of children,” current reforms reflect a perceived need to “get tough” with violent adolescents. The general perception seems to be that youth “get off easy” in juvenile court, which contributes greatly to high levels of juvenile crime (Greenwood, 1995). Accord-ingly, the outlook on handling delinquents has shifted from a rehabil-itative philosophy to one that is more harmonious with retribution, incapacitation, and deterrence. To this end, recent changes include im-proved and more accessible juvenile records, fingerprinting of young offenders, and the use of public hearings and mandatory sentencing in juvenile courts (Greenwood, 1995; Torbet et al., 1996). Probably the most popular reform effort has been in the area of 1 transferring serious and violent youthful offenders to adult criminal court. Although state legal codes exhibit wide variation in the age 2 limitations placed on juvenile court jurisdiction, all states do have provisions that allow juveniles to be tried in adult court (DeFrances & Strom, 1997; Sickmund et al., 1997). In recent years, many states have amended their juvenile codes to facilitate this process (Feld, 1987, 1993; 1

A number of terms have been used to represent this process: transfer, waiver, certification, and remand, to name a few (Champion & Mays, 1991). In general, there is little or no difference in the meaning of these expressions, and throughout this study they are used interchangeably. 2 Minimum age restrictions for juvenile court jurisdiction are often not stipulated in state codes, meaning individual courts can either develop their own policy or rely on the common law notion that no person can be held criminally responsible for their actions until the age of 7 (Krisberg and Austin, 1993, p. 67). Maximum ages for original juvenile court jurisdiction currently range from 15 to 17 years of age, with 37 states and the District of Columbia using 17 as the upper age limit (DeFrances & Strom, 1997; Sickmund et al., 1997). States also vary in the maximum age at which the juvenile court can retain control beyond the original oldest age of jurisdiction. Some states may retain control until the age of 21, or even higher (e.g., youths committed to the California Youth Authority for certain offenses may be retained up to age 25; see Krisberg & Austin, 1993, p. 70). Finally, there is state variation in the youngest age at which juveniles may be transferred to adult criminal court. Many states have no minimum age limit (i.e., there is no exclusive juvenile court jurisdiction), while the lower age limit for transfer in other states ranges from 10 to 16 (Krisberg & Austin, 1993, p. 69).

Introduction

3

Fritsch & Hemmens, 1995; Sickmund et al., 1997; Torbet et al., 1996). This legislative action has resulted in increasing numbers of juveniles being sent to adult court, particularly for violent offenses (DeFrances & Strom, 1997; Howell, 1997; Sickmund et al., 1997). In general, it is believed that these youthful offenders will receive harsher treatment in adult court, which in turn will have a beneficial impact on juvenile crime, through both greater deterrence and longer incapacitation (Bishop, Frazier, Lanza-Kaduce, & Winner, 1996). In a 1993 USA Today/CNN/Gallup Poll (Meddis, 1993), 73% of the adults surveyed believed violent juvenile offenders should be treated as adults. Recent juvenile justice system reforms reflect the public point of view. Unfortunately, changes in juvenile waiver policies have not been guided by systematic research. Howell (1997) asserts, “Transfer is a sociolegal policy based on very little information” (pp. 108-109). There have been few scientific studies comparing the effectiveness of juvenile and criminal court processing. This lack of evidence limits the conclusions that can be made regarding the relative merits of treating juveniles as adults: It is not at all certain that we gain increased deterrence, retribution, or incapacitation in this way. Youth committed by the adult court to adult prisons might become hardened and more, rather than less, likely to offend again upon release. What is needed is much more careful research following comparable samples of offenders through these different experiences to provide a better understanding and confident policies about the division of jurisdiction between the two courts, the relative effectiveness of the dispositional options they provide, and the efficacy of the criteria used to select offenders for differential processing and disposition (Farrington, Ohlin, & Wilson, 1986, p. 125). Essentially, then, there are two competing arguments surrounding this issue. First, a policy of this type may reduce juvenile offending, through the imposition of stronger punishment that provides lengthier incapacitation and greater deterrence. On the other hand, treating juveniles as adults may make things worse, by producing offenders who are more criminally disposed and motivated.

4

Excluding Violent Youths From Juvenile Court

As increasing efforts continue toward sending violent youths to adult court, the need for further research becomes greater. Certification to the criminal justice system is the most severe disposition that can be imposed by a juvenile court, and it has even been characterized as “the capital punishment of juvenile justice” (Zimring, 1981, p. 193). Consequently, it is important to know the effectiveness of this policy option. Efforts to treat young violent offenders as adults are supported by claims of greater accountability, stronger punishment, and enhanced community safety. In other words, transferring cases to criminal court “carries with it the expectation that punishment will be swifter, more certain, more severe, and more effective as a crime control strategy” (Fagan, 1995, p. 252). Some evidence exists regarding these issues, but it is often of uneven quality, and studies focusing on juvenile transfer tend to emphasize certain aspects of the process more than others. Several authors have provided excellent presentations of the historical development of the transfer process (see, e.g., Feld, 1987, 1993; Forst & Blomquist, 1991; Thomas & Bilchik, 1985). Nearly 100 years ago, with the establishment of the first juvenile courts, delinquents were deemed to require different treatment under the law 3 than adult criminals. However, juvenile court judges were also granted the discretion to transfer serious youthful offenders from juvenile court to adult court. Over the next century, as states further developed and refined their juvenile systems, various mechanisms were established to waive juvenile court jurisdiction in certain instances. Almost all contemporary juvenile court judges retain the power to transfer cases to adult court. In addition, this authority has also been granted to some prosecutors, and legislatures are increasingly excluding certain types of offenses, offenders, or both, from juvenile court jurisdiction. Therefore, transferring juveniles to criminal court is not a new phenomenon, but until recently transfer procedures were used more sparingly. A substantial amount of research has examined the offense characteristics of transferred youth (Barnes & Franz, 1989; Bishop & 3

Historical accounts reveal that the creation of the juvenile court was strongly influenced by a "progressive movement," which stressed positivistic beliefs that behavior is amenable to rehabilitation and that the causes of juvenile delinquency and adult crime are different (Krisberg & Austin, 1993; Rothman, 1980).

Introduction

5

Frazier, 1991; Bishop, Frazier, & Henretta, 1989; Bortner, 1986; Champion, 1989; Clarke, 1996; Gillespie & Norman, 1984; Hamparian et al., 1982; Houghtalin & Mays, 1991; Keiter, 1973; Kinder, Veneziano, Fichter, & Azuma, 1995; Lee, 1994; Lemmon, Sontheimer, & Saylor, 1991; Osbun & Rode, 1984; Podkopacz & Feld, 1996; Poulos & Orchowsky, 1994; Sagatun, McCollum, & Edwards, 1985; Thomas & Bilchik, 1985). Studies using data collected in the 1970s and 1980s indicated that the largest percentage of waived juveniles had been transferred on property offenses (Bishop & Frazier, 1991; Bishop et al., 1989; Bortner, 1986; Champion, 1989; Hamparian et al., 1982; Lemmon et al., 1991; Thomas & Bilchik, 1985). More recent research has suggested a change, as juveniles charged with personal or violent offenses now make up the largest percentage of transfers (Barnes & Franz, 1989; DeFrances & Strom, 1997; Houghtalin & Mays, 1991; Kinder et al., 1995; Podkopacz & Feld, 1996; Sickmund et al., 1997). This shift corresponds well with the increasing public concern over youth violence, and seems, in part, to reflect recent changes in waiver laws that target violent offenses for exclusion from juvenile court. Nearly all studies that measure the race of transferred juveniles find that Blacks are over-represented, generally making up between 50% and 90% of all waived youths (see, e.g., Clarke, 1996; DeFrances & Strom, 1997; Houghtalin & Mays, 1991; Kinder et al., 1995; Thomas & Bilchik, 1985). This hints at racial bias, as roughly 15% of all juveniles are Black, and they account for approximately 32% of all juvenile court referrals and 49% of all juvenile arrests for violent index crimes (Snyder & Sickmund, 1995, p. 91). Still, relatively few studies have analyzed a possible race effect on the transfer decision. Some evidence exists that race does have a significant influence (Eigen, 1981a, 1981b; Keiter, 1973). However, more recent research, with stronger controls for offense seriousness and prior record, has failed to find a direct effect (Fagan & Deschenes, 1990; Fagan, Forst, & Vivona, 1987; Podkopacz & Feld, 1995, 1996; Poulos & Orchowsky, 1994; Singer, 1993). It is possible that a race effect could be more indirect than direct, as the 4 decision to transfer comes late in the juvenile system process. Other 4

For example, prior record is affected by police response to juvenile crime, and police response may be influenced strongly by neighborhood social context and status of minority group members (Smith, 1986).

6

Excluding Violent Youths From Juvenile Court

research focusing on juvenile arrest, intake, detention, adjudication, and disposition has found evidence of racial bias (see, e.g., Leonard, Pope, & Feyerherm, 1995; Pope & Feyerherm, 1990). Case outcomes in adult criminal court have been another major concentration in waiver research. Here, studies have examined primarily two matters: conviction and incarceration. Most studies find high conviction rates for transferred juveniles, generally ranging between 75% and 95% (Bishop & Frazier, 1991; Bishop et al., 1989; Bortner, 1986; Champion, 1989; Clarke, 1996; Eigen, 1981a, 1981b; Fagan, 1990; Gillespie & Norman, 1984; Hamparian et al., 1982; Houghtalin & Mays, 1991; Lemmon et al., 1991; Podkopacz & Feld, 1996; Rudman, Hartstone, Fagan, & Moore, 1986; Thomas & Bilchik, 1985; for exceptions, see Kinder et al., 1995; Sagatun et al., 1985; Singer, 1996). However, it is unclear if these conviction rates in criminal court are significantly different from those of comparable offenders in juvenile court. The limited research that has addressed this question is mixed, with a few studies finding evidence of higher conviction rates in criminal court (Eigen, 1981a, 1981b; Fagan, 1995; Podkopacz & Feld, 1996), while others have found higher rates in juvenile court (Kinder et al., 1995) or little difference between the two systems (Fagan, 1990; Rudman et al., 1986). Concerning incarceration, researchers have examined both incarceration rates of transferred juveniles and length of incarceration. Studies show that youthful property offenders tend to be treated leniently in criminal court, often receiving sentences of probation (Barnes & Franz, 1989; Bishop & Frazier, 1991; Bishop et al., 1989; Bortner, 1986; Champion, 1989; Hamparian et al., 1982; Podkopacz & Feld, 1996). On the other hand, violent juvenile offenders appear to be treated harshly in adult court, where a jail or prison term is often imposed (Barnes & Franz, 1989; Clarke, 1996; Eigen, 1981a, 1981b; Fagan, 1990, 1995; Houghtalin & Mays, 1991; Podkopacz & Feld, 1996; Rudman et al., 1986). Nevertheless, as with conviction rates, it is still somewhat unclear if these incarceration rates in criminal court are very different from those of comparable offenders in juvenile court. Some early research on juveniles prosecuted in criminal court found evidence of a “leniency gap” for these young offenders, who seemed to appear relatively less criminal than other older and more experienced adult defendants. Studies suggested that based on their younger age and lack of an extensive criminal record, juveniles certified

Introduction

7

to adult court typically were not imprisoned and received more lenient sentencing than they would have in juvenile court (Bortner, 1986; Champion, 1989; Emerson, 1981; Gillespie & Norman, 1984; Hamparian et al., 1982; Royscher & Edelman, 1981; Sagatun et al., 1985). However, other research assessing this situation has found higher incarceration rates in criminal court than in juvenile court, particularly for violent offenders (Barnes & Franz, 1989; Bishop et al., 1996; Eigen, 1981a, 1981b; Fagan, 1990, 1995; Podkopacz & Feld, 1996; Rudman et al., 1986). Research examining incarceration length reveals that for those transferred youths who are incarcerated, lengthy sentences are common (Bishop & Frazier, 1991; Bishop et al., 1989; Lemmon et al., 1991; Singer, 1996; Thomas & Bilchik, 1985). Average prison sentences range from 1 to 4 years, with the longest imposed on violent offenders. However, it is still questionable if longer periods of incarceration are prescribed in adult criminal court than for comparable cases in juvenile court. The evidence appears stronger that lengthier sentences are imposed in adult court (Bishop et al., 1996; Eigen, 1981a, 1981b; Fagan, 1990; Fritsch, Caeti, & Hemmens, 1996; Rudman et al., 1986; Podkopacz & Feld, 1996), but at least one recent study that compared the length of juvenile and criminal court sentences found them to be very similar (Fagan, 1995). Furthermore, when actual time served is taken into consideration, transferred youths may often serve shorter sentences than those available in juvenile court (Fritsch et al., 1996). In addition to these major areas of research, a few other aspects of treating juvenile offenders as adults have been examined, but to a much lesser extent. First, the juvenile court has been criticized recently for not operating fast enough (Butts, 1996, 1997; Butts & Halemba, 1994). th The 6 Amendment to the Constitution guarantees a “speedy and public trial” for any citizen charged with a crime, but the definition of “speedy” is unclear. Furthermore, the United States Supreme Court has never confronted the issue of speedy trial rights for juveniles. Nevertheless, juveniles waived to adult court may be subjected to longer case processing times than they would have experienced in juvenile court. A small amount of research indicates that transferred juveniles have their cases processed more slowly than similar cases in juvenile court (Fagan, 1995; Kinder et al., 1995; Rudman et al., 1986). Case processing time is an emerging issue in juvenile justice, with increased emphasis being placed on “immediate interventions” (Howell,

8

Excluding Violent Youths From Juvenile Court

1997; Krisberg, Currie, Onek, & Wiebush, 1995). Therefore, it seems important to know how rapidly comparable cases are processed in the juvenile and adult systems. In national and local discussions regarding youth violence, deterrence is a primary topic that often emerges. Although the principles of retribution and incapacitation are also addressed, deterrence is a fundamental justification for waiving youths to criminal court (Bishop et al., 1996; Fagan, 1995; Singer, 1996). This issue of deterrence can be framed in two questions. First, does the transfer of juveniles to adult court affect the rate of juvenile crime (i.e., is there a general deterrent effect)? Second, do transferred juveniles refrain from offending as a consequence of being prosecuted in criminal court (i.e., is there a specific deterrent effect)? With the recent popularity of transferring juveniles, and the inherent intent of providing a mechanism that results in deterrence, it is surprising that few researchers have designed studies to answer these questions. Only two studies have looked at the impact of expanded juvenile transfer laws on aggregate adolescent crime rates (Jensen & Metsger, 1994; Singer & McDowall, 1988; see also Singer, 1996). Neither study found evidence of a general deterrent effect. Recently, three studies examined recidivism among comparable youths in the juvenile and adult systems (Bishop et al., 1996; Fagan, 1995; Podkopacz & Feld, 1996; see also Winner, Lanza-Kaduce, Bishop, & Frazier, 1997). Their combined results do not support a specific deterrent effect, as the waived youths exhibited higher rates of rearrest and reincarceration, shorter time until failure, and involvement in more serious crime. Efforts to relocate youthful offenders to criminal court have been supported by the expectation of increased accountability (i.e., more certain, severe, and swift punishment) and enhanced public safety. This corresponds well with the central arguments of deterrence theory. Through treating juvenile offenders as adults, policy-makers seek to deter young people from committing crime by threatening and actually punishing them with formal criminal sanctions. However, based on the research just discussed, it is highly questionable if the goals of greater accountability and stronger deterrence are met through this policy option.

Introduction

9

Deficiencies in Transfer Research Although the topic of transferring juvenile offenders to adult court has been addressed in the literature, there is a need to improve and extend the research in this area. Most past studies are descriptive in nature. Early research in the 1970s and 1980s often focused solely on offense and offender characteristics of waived youths and did not provide adequate controls to examine the objectivity of juvenile justice system decision-making (see, e.g., Bishop & Frazier, 1991; Bishop et al., 1989; Bortner, 1986; Champion, 1989; Eigen, 1981a, 1981b; Gillespie & Norman, 1984; Hamparian et al., 1982; Keiter, 1973; Osbun & Rode, 1984; Thomas & Bilchik, 1985). Recently, researchers have begun to use more sophisticated multivariate statistical techniques and comparison groups of similar offenders retained in the juvenile system. However, data on critical variables (e.g., prior record, weapon use, victim injury) is still often lacking (see, e.g., Clarke, 1996; Kinder et al., 1995; Lee, 1994). Most of the available studies focus on judicial waiver, where a juvenile court judge makes a decision to remand a case to adult criminal court. This emphasis on judicial waiver is intriguing, because “these cases represent less than 10% of all juveniles that are moved into the criminal justice system” (Howell, 1997, p. 108). Much less is known about prosecutorial waiver, where a prosecutor directly files a case in adult court, and legislative waiver, where certain youthful offenders are statutorily excluded from juvenile court jurisdiction. The lack of information available regarding legislative waiver is particularly disturbing, as many legislatures today are moving to limit the discretionary decision-making of juvenile court judges and prosecutors and subsequently relocate more youths to adult court (Fritsch & Hemmens, 1995; Sickmund et al., 1997; Torbet et al., 1996). Finally, there have been very few studies of high quality that have addressed the effectiveness of treating juvenile offenders as adults, in terms of the certainty, severity, and swiftness of punishments imposed and subsequent community safety. Most of the research that has addressed case outcomes of transferred youths presents only descriptive statistics on conviction rates, incarceration rates, and length of incarceration. Studies that actually compare the effectiveness of juvenile and criminal court processing and add a consideration of

10

Excluding Violent Youths From Juvenile Court

recidivism are rare, with the notable exceptions of Bishop et al. (1996), Fagan (1995), and Podkopacz and Feld (1996). Thus, the shift toward treating increased numbers of adolescent offenders as adults is not based on a strong foundation of research. The problem is summarized well in a recent report from the Office of Juvenile Justice and Delinquency Prevention: Studies on the impact of criminal court transfer have not yielded definitive conclusions. The debate over the efficacy of criminal court transfer has been underway for at least 50 years. Yet, there are still no definitive answers to basic questions about the effects of the practice. In many ways, policymakers are operating in the dark on this issue. (Snyder & Sickmund, 1995, pp. 155-156)

The Current Research The purpose of this study is to investigate the overall effectiveness of a policy that targets violent juvenile offenders for prosecution in adult criminal court. We will examine the nature of punishment imposed on specified violent youths in both the juvenile and adult justice systems, in terms of the certainty, severity, and swiftness of prescribed sanctions. We also will assess the specific deterrent effect of juvenile transfer by conducting analyses of recidivism, which will determine whether youths waived to adult court are more likely to refrain from future criminal behavior than similar adolescents retained in juvenile court. Data will be used from the Commonwealth of Pennsylvania’s Center for Juvenile Justice Training and Research, and findings from interviews with various juvenile justice system officials will also be discussed.

CHAPTER 1

Cracking Down On Youth Violence

During the past thirty years, there has been vigorous debate over the juvenile justice system’s philosophy, structure, and procedures. Critical attacks have come from a variety of angles, focusing on such issues as insufficient enforcement of due process rights, inadequate treatment and rehabilitation services, abuse of the juvenile court’s power, lenient treatment of offenders, and a general lack of direction in dealing with juvenile crime (Feld, 1993; Greenwood, 1995; Krisberg & Austin, 1993; Moore & Wakeling, 1997; Schwartz, 1989). These criticisms, combined with recent increases in youth violence, have led to an erosion of the traditional juvenile court’s philosophy and authority. A “get tough” ideology, which originated in the adult criminal system in the 1970s, now extends into the juvenile system as well. A central issue is the transfer of juveniles to adult court, which is often described as a move toward “criminalizing” delinquent behavior (Feld, 1993; Fagan, 1995; Singer, 1996). To provide a context for understanding the modern response to youthful offending, this chapter will first offer a brief review of the original juvenile court philosophy. The evolution of juvenile courts will then be discussed, as well as the recent concern over violent juvenile crime. Changes in juvenile court jurisdiction will also be addressed, through an examination of the policy of treating young offenders as adults. The chapter concludes by discussing recent modifications in Pennsylvania’s juvenile system.

11

12

Excluding Violent Youths From Juvenile Court

Origins of Juvenile Justice To a certain extent, juvenile offenders have been treated differently th from adult criminals for approximately 600 years. By the late 14 century, English common law began to recognize the defense of infancy or immaturity (Thomas & Bilchik, 1985). Children under the age of seven could not be found guilty of a crime. From ages 7 through 14, children were presumed to lack any criminal capacity, but this presumption could be rebutted. Anyone over the age of 14 could not raise infancy or immaturity as a defense and was treated as an adult. During the Colonial Period, American legal doctrine similarly 5 recognized the age of law violators. However, the creation of a separate system of juvenile justice is most often associated with the establishment of houses of refuge in the 1820’s (Bernard, 1992). The first house of refuge opened in New York City in 1825, and the idea of a separate institution exclusively for juveniles quickly spread to Boston and Philadelphia. However, most children committed to these institutions were not there for serious criminal behavior (Bernard, 1992; Thomas & Bilchik, 1985). The institutions were more like poorhouses, with their purpose being to prevent children from growing up to be “paupers” (i.e., poor people undeserving of charity, due to their lazy and wicked ways). Many youths were placed in houses of refuge without committing any criminal offense, and the courts that did so also failed to provide due process rights. Thus, constitutional challenges th arose. By the late 19 century, a new legal mechanism was needed to reestablish a separate system of justice for youthful offenders. At the turn of the century, the Progressive reform movement sought to change the way children were treated by the legal system (Forst & Blomquist, 1991; Krisberg & Austin, 1993). A separate court system was envisioned that would take into account that juveniles lacked the reasoning ability and moral development of adults. When the first juvenile court opened in Chicago in 1899, it was based on the doctrine of “parens patriae,” meaning it was the state’s responsibility to care for children who were unable to care for themselves and whose parents 5

There were exceptions. In some American colonies, only youths over the age of 16 could receive capital punishment or harsh physical punishments for noncapital offenses, and courts and juries were often lenient toward youths (Platt, 1969).

13

Cracking Down On Youth Violence 6

either could not or would not care for them. Early juvenile courts emphasized an informal and protective environment, resting on a philosophy of rehabilitation and individualized justice (Feld, 1987, 1993). Consequently, indeterminate sentencing became the norm, and discretionary decision-making was pervasive throughout the system. The philosophy of parens patriae represented a distinct break from the traditional punitive responses of the criminal court. Juvenile court actions were designed to determine if a child was in need of care and what should be done to prevent a future life of crime. This orientation is evident in the following description of early juvenile court procedures in Boston: The sheltered location of the room, the absence of decoration, the dispensing with attendants and the exclusion of outsiders give the simplicity that is necessary to gain the undivided attention of the child, and give the quiet which is indispensable for hearing clearly what the child says and speaking to him in the calmest tone. . . . The child does not stand in front of the desk, because that would prevent the judge from seeing the whole of him, and the way the child stands and even the condition of his shoes are often useful aids to a proper diagnosis of the case. The child stands at the end of the platform where the judge can see him from top to toe, and the judge sits near the end, so he is close to the child and can reassure him if necessary by a friendly hand on the shoulder. (Baker, 1979, p. 148) In this setting, punishment was not viewed as a central goal. Delinquency was seen as an illness, brought on by urbanization, industrialization, poverty, parental neglect, and other social diseases (Bernard, 1992; Feld, 1987, 1993; Forst & Blomquist, 1991; Krisberg & Austin, 1993). This perspective corresponded well with changing ideological assumptions in the field of criminology. In the latter half of th the 19 century, positivism challenged the classical school belief that crime was a product of free will. The positive school emphasized 6

For a review of the history of the doctrine of parens patriae, see Rendleman (1979).

14

Excluding Violent Youths From Juvenile Court

determinism, or the need to identify causal variables that produce crime and delinquency. Thus, positive criminology combined with the progressive movement to produce a “rehabilitative ideal,” whereby the juvenile court was to be given the tools and strategies necessary to effect change in the behavior of young people (Feld, 1987, 1993). Similar to the houses of refuge that were established 75 years earlier, the new juvenile court generated great optimism that it could solve the problem of youthful offending. By 1925, all but two states had adopted juvenile courts, and by 1945 the last state had done so (Bernard, 1992; Thomas & Bilchik, 1985). However, there was no uniform court structure in place. Juvenile courts varied in their age limits for jurisdiction and in the types of young people they supervised (i.e., dependent, neglected, and delinquent youth). Furthermore, as the juvenile system developed, a relaxed approach was taken regarding due process and rules of procedure. This was justified on the grounds that juvenile courts were serving the “best interests” of children, which required an individualized diagnosis of problems and a flexible response. Theoretically, the juvenile court’s proceedings were designed to identify the underlying causes of youthful misbehavior and provide the treatment necessary to prevent more serious criminal behavior from happening in the future. A distinct language was developed to portray a sense of concern and further distinguish the boundaries between juvenile and criminal jurisdiction (Bernard, 1992). In spite of this orientation, it should be noted that juvenile court judges were also given the discretion to transfer serious young offenders to adult criminal court: Legislation in many states permitted juvenile court judges to transfer any given case to an adult court, an allowance that they occasionally took advantage of when the charge was especially serious. The Cook County juvenile court, for example, asked grand juries to weigh the merits of a regular indictment in about fifteen cases a year—a figure that represented no more than one percent of its cases but did include the most notorious. Typically these boys were older (sixteen, not twelve) and were arrested for “deeds of violence, daring holdups, carrying guns, thefts of considerable amounts,

Cracking Down On Youth Violence

15

and rape.” These transfers probably muted criticisms of the courts for coddling the criminal. (Rothman, 1980, p. 285) Therefore, the notion of treating juvenile offenders as adults is not something new. Many of the considerations pertaining to this practice also seem to be consistent across time. It is interesting that from the outset, the juvenile court appeared to “wash its hands” of the youths who were possibly most in need of help. Nevertheless, until recently, transfer procedures were used sparingly as the philosophy of parens patriae ensured that most young offenders would remain within the juvenile system.

Criticisms and Reforms Although there were occasional constitutional debates and sporadic criticisms of the juvenile court following its creation, the philosophy and procedures just described went largely unchallenged until after the th mid-20 century (Bernard, 1992; Empey & Stafford, 1991). In the 1950s and 1960s, critics became increasingly vocal about the deficiencies of the juvenile system. The first round of criticisms focused on the lack of procedural safeguards granted to youthful offenders in juvenile court (Forst & Blomquist, 1991). The informality of court hearings and the broad discretionary power granted to juvenile justice officials were specifically attacked. The concern for due process rights grew as it became apparent that the juvenile system often exhibited the same concern for punishment and incapacitation as the adult criminal 7 system. During the “due process revolution” of the late 1960s and early 1970s, several United States Supreme Court decisions established that

7

"By the 1950s, optimism about the juvenile court had broken down and a more realistic view began to emerge. This view was based on an assessment of the actual performance of the juvenile court rather than the good intentions of its founders. In practice, the juvenile court often did not 'treat' juveniles or act in their 'best interests,' but only punished them for their offenses. This new realism laid the basis for a reintroduction of due process protections: if juveniles are being punished, then they need at least some of the protections provided to adults in criminal court" (Bernard, 1992, p. 108).

16

Excluding Violent Youths From Juvenile Court

juveniles could not be totally cut off from fundamental constitutional 8 rights under the guise of “individualized justice.” In the landmark case of Kent v. United States (1966), the Court ruled that prior to being transferred to criminal court by a juvenile court judge, youths have the right to a formal hearing regarding the waiver criteria and a right to counsel at this hearing. The Court also noted that the benefits and protections supposedly extended to youths in juvenile court do not negate the need for certain due process rights: While there can be no doubt of the original laudable purpose of juvenile courts, studies and critiques in recent years raise serious questions as to whether actual performance measures well enough against theoretical purpose to make tolerable the immunity of the process from the reach of constitutional guaranties applicable to adults. . . . There is evidence, in fact, that there may be grounds for concern that the child receives the worst of both worlds: that he gets neither the protections accorded to adults nor the solicitous care and regenerative treatment postulated for children. (see Bernard, 1992, p. 113) The following year, the case of Gerald Gault (In re Gault, 1967) established that juveniles who face institutional confinement require the procedural safeguards of advanced notice of charges, legal representation, the opportunity to confront and cross-examine witnesses, and the privilege against self-incrimination. The case involved a 15-year-old boy who had been arrested for making an 9 “obscene” telephone call to a female neighbor and was subsequently st placed in a state industrial school for up to his 21 birthday. No record was kept of the adjudication hearing, and the victim did not appear. In extending several more due process rights to juveniles, the Supreme Court asserted that the benefits of the juvenile court and the juvenile court’s unique procedures for processing youths separately from adults 8

For a lengthy presentation and discussion of these cases, see Faust and Brantingham (1979). For a good summarization, see Bernard (1992). 9 The offense allegedly involved the following questions: "Do you give any?""Are your cherries ripe today?" and "Do you have big bombers?" (see Bernard, 1992, p. 114).

Cracking Down On Youth Violence

17

should be available without any reduction in the law’s protection. The Court also directed strong words at juvenile courts that fail to recognize basic due process rights: “It would be extraordinary if our Constitution did not require the procedural regularity and the exercise of care implied in the phrase ‘due process.’ Under our Constitution, the condition of being a boy does not justify a kangaroo court” (see Bernard, 1992, p. 117). In the 1970s, the Supreme Court continued to address the issue of due process protections for juveniles. With the decision in In re Winship (1970), the Court asserted that in delinquency proceedings, the standard of proof required is “beyond a reasonable doubt” rather than “by preponderance of the evidence.” In Breed v. Jones (1975), the th Court found that the protections of the double jeopardy clause of the 5 Amendment require juvenile courts to make a decision to waive jurisdiction to criminal court prior to any determination of guilt. Although the Supreme Court stopped short of extending to juveniles all 10 the procedural safeguards given to adults, these cases continued to move the juvenile court away from its original Progressive orientation and toward a more procedurally formal system that emphasized substantive justice. While initial criticisms of the juvenile court focused on constitutional rights and procedural fairness, “the second wave of criticisms and reforms that emerged in the mid-1970s into the 1980s was specifically directed at changing the goals and structure of the system” (Forst & Blomquist, 1991, p. 331). Here, critics focused primarily on two areas: the perceived ineffectiveness of rehabilitation programs and rapidly rising crime rates. In a short period of time, concerns over these two factors would lead to rather dramatic changes in juvenile justice philosophy and practice. In the 1970s, several influential reviews of rehabilitation programs seemed to suggest, “nothing works” (Lipton, Martinson, & Wilks, 1975; Martinson, 1974; Sechrest, White, & Brown, 1979). Reservation about the potential for rehabilitation conflicted strongly with the juvenile court’s traditional emphasis on treatment and individualized care. Furthermore, court dispositions that removed youths from the 10

For example, the right to a jury trial (see McKeiver v. Pennsylvania, 1971) and the right to bail (see Schall v. Martin, 1984) were not granted.

18

Excluding Violent Youths From Juvenile Court

community and placed them in a secure facility were increasingly seen as inherently punitive rather than rehabilitative. In addition to the apparent inability of the juvenile system to provide effective rehabilitative services, crimes committed by young people increased rapidly from the mid-1960s through the mid-1970s. By the end of this period, the popular view emerged that the juvenile court was too lenient, particularly with serious and violent offenders (Empey & Stafford, 1991; Forst & Blomquist, 1991). Assertions were made that the juvenile system was unable to ensure public safety (Regnery, 1985, 1986; Wilson, 1983; Wolfgang, 1982). Combined with earlier constitutional challenges, concerns over the efficacy of rehabilitation and rising youth crime encouraged a reexamination of the juvenile court’s structure and philosophy. The response was a rapid shift toward a more punitive system, emphasizing 11 accountability, deterrence, and incapacitation. Commitment to the rehabilitative ideal gave way to a concentration on “just deserts,” and a “get tough on crime” approach was supported by politicians at all levels of government (Thomas & Bilchik, 1985). Jurisdiction after jurisdiction made modifications to their juvenile codes. Popular strategies included easing the transfer of youths to adult court, incorporating the goals of accountability and public safety into juvenile codes, and using determinate and mandatory sentencing in juvenile courts (Forst & Blomquist, 1991). It was hoped that these changes would be an effective response to the increasing concern over juvenile crime.

The Surge in Youth Violence Over the past 15 years, juveniles have accounted for less than 35% of all serious property crime arrests and less than 20% of all violent crime arrests (Cook & Laub, 1998; Empey & Stafford, 1991; Howell, 1997; Snyder & Sickmund, 1995). Nearly 70% of all juvenile arrests are for nonindex crimes (i.e., crimes other than murder, rape, robbery, aggravated assault, burglary, larceny, motor vehicle theft, and arson). 11

A similar change took place in the adult criminal justice system, beginning slightly before this shift in juvenile justice philosophy. In the 1970s, the adult system began to move away from an emphasis on rehabilitation and indeterminate sentencing, toward a more retributive philosophy characterized by sentencing guidelines and mandatory sentencing (Tonry, 1996).

Cracking Down On Youth Violence

19

Furthermore, each year a small percentage of all juveniles (somewhere in the neighborhood of 6%) are arrested for any offense. Despite these statistics, in the mid-1980s a disturbing trend emerged that served to heighten fears about youthful offending. Beginning in 1985 and continuing through 1994, the juvenile violent crime arrest rate increased by 75% (Sickmund et al., 1997, p. 18). This rise followed a decade of fairly stable rates, during which time 12 the juvenile justice system had shifted to a more punitive orientation. While juvenile arrest rates for robbery and aggravated assault exhibited rapid growth, the surge in youth homicide rates was particularly disturbing. Not only did the juvenile murder arrest rate rise at a much faster pace than that of adults, but also for older youths the increase was over 100% (Blumstein, 1995b). This increase in homicide rates corresponded with the accumulation of guns among youths (Blumstein, 1995b; Cook & Laub, 1998; Fagan & Wilkinson, 1998; Greenbaum, 1997). Juvenile arrest rates for weapons law violations also climbed steadily, with the greatest upturn occurring after 1987 (Snyder, 1997). Not surprisingly, juvenile gun homicides rose dramatically, while homicides involving other weapons remained stable (Greenbaum, 1997). Blumstein (1995b) attributed this development to the recruitment of young people into illicit drug markets and the corresponding need to arm themselves. Furthermore, he argued that gun homicides by young people are often committed against strangers, which is instrumental in generating public fear and feelings of vulnerability. Undoubtedly, policy-makers were influenced by this public fear, as efforts to “get tough” and “crack down” on violent juvenile crime escalated during the 1990s. While the growing concern over youth violence was enough to accelerate punitive juvenile justice policy trends that originated in the late 1970s, demographic influences have added even more to the perceived need to “do something” about violent juvenile crime. To begin, crime rates are age sensitive. Involvement in most types of crime 12

In contrast to the juvenile violent crime arrest rate, the juvenile arrest rate for property crime did not increase substantially from the mid-1980s until the mid1990s. Although a modest increase occurred between 1989 and 1991, in 1993 the juvenile property crime arrest rate was at nearly the same level as it was from 1986 to 1989 (Snyder, 1997, p. 6).

20

Excluding Violent Youths From Juvenile Court

peaks during mid-to-late adolescence and then steadily declines (Blumstein, 1988; 1995a; 1995b; Blumstein et al., 1986; Farrington, 1986, 1998). This fact has been used to at least partially explain the dramatic increase in official crime from the mid-1960s through the mid1970s. During this time period, the youth population also grew, as the post-World War II “baby boomers” reached their teenage years (Empey & Stafford, 1991; Thomas & Bilchik, 1985). Combined with an atmosphere of civil rights protests, urban rioting, campus rebellion, and Vietnam War opposition, increases in adolescent offending suggested that a large generation of youth had gone out of control. Today, a fear of another large generation of young Americans is prevalent. Beginning in the 1980s, children of the baby boom generation (i.e., the “echo boom”) started to reach adolescence. Those age 18 and under now rival baby boomers for the greatest percentage of the United States population, and this age group is expected to remain large over the next decade. Based on the projected size of the at-risk population of 14 to 17 year olds, total juvenile arrests for violent offenses are also expected to remain high and even increase (Howell, Krisberg, & Jones, 1995). The number of total arrests will depend greatly on the youth violence rate over the same time period. For example, in 1994 the National Council on Crime and Delinquency suggested that if youth violence rates continued to increase as they did from the mid-1980s through the mid-1990s, total juvenile arrests could increase more than 100% by the year 2002 (see Howell et al., 1995, p. 10). Predictions like these have further incited fears about youth violence. In spite of the recent downturn in the juvenile violent crime arrest rate (Sickmund et al., 1997; Snyder, 1997), the projected size of the teenage population continues to receive attention. Politicians and criminologists alike regularly mention its significance, with warnings that we need to “get ready” (Wilson, 1995, p. 507). Of greatest concern is that the possible juvenile crime wave will be particularly vicious, due to the ready availability of guns and drugs in our society. Although it is possible that the forecast of severe youthful offending could be incorrect (Cook & Laub, 1998; Howell, 1997), descriptions of future juvenile “super-predators” flooding the nations streets (DiIulio, 1995, 1996) have been very influential on public policy. From 1992 through 1995, laws were passed in 47 states and the District of Columbia that sought to generally toughen juvenile systems (Sickmund et al., 1997; Torbet et al.,

Cracking Down On Youth Violence

21

1996). Changes most often pertained to the areas of confidentiality, sentencing, and jurisdiction. Regarding issues of privacy, many states moved to improve juvenile court records and make them more accessible. Also, most states now allow police to fingerprint and photograph specified juveniles, and open hearings are being used for certain cases. Concerning sentencing, many state legislatures acted to remove discretion from judicial decisions by requiring mandatory minimum periods of confinement for certain violent or serious offenders. Other modifications included raising the maximum age of the juvenile court’s continuing jurisdiction over youthful offenders and imposing “blended sentences” that combine juvenile and adult correctional alternatives. Finally, probably the biggest area of change occurred in jurisdictional authority, as new laws sought to remove an increasing number of serious and violent offenders from juvenile court and prosecute them in adult criminal court (Sickmund et al., 1997; Torbet et al., 1996).

Methods of Transfer Although there is variation among states in their minimum and maximum ages for juvenile court jurisdiction, today all states do have provisions that allow juveniles to be tried in adult criminal court (DeFrances & Strom, 1997; Sickmund et al., 1997). There are three primary ways to remove a youth from juvenile court jurisdiction: 13 judicial waiver, prosecutorial waiver, and legislative waiver. Each represents a different way to identify which young offenders to prosecute as adults, and most states have some combination of these three procedures in place (Sickmund et al., 1997; Torbet et al., 1996). Judicial Waiver Historically, transferring juveniles to adult criminal court has been accomplished most often through judicial waiver, whereby a juvenile court judge makes the key decision in the transfer process. Currently, 46 states and the District of Columbia have judicial waiver provisions 13

This section will briefly describe each of these procedures. For further information, see Champion and Mays (1991), Feld (1987), Forst and Blomquist (1991), Fritsch and Hemmens (1995), and Krisberg and Austin (1993).

22

Excluding Violent Youths From Juvenile Court

(DeFrances & Strom, 1997). A judge’s decision to remand a case to adult court requires a determination that the youth is no longer amenable to the treatment offered in the juvenile system, based on such factors as age, offense seriousness, prior record, and public safety (Feld, 1987, 1993). This assessment reflects the individualized nature of dispositions that is characteristic of traditional juvenile courts. The United States Supreme Court addressed the practice of judicial waiver in Kent v. United States (1966) and Breed v. Jones (1975). Although these two cases established a procedural framework for judges to use in making the waiver decision, they by no means eliminated judicial discretion from the process. In Kent, the majority opinion set forth several factors a judge might consider in making a decision to transfer, but vague phrases such as “amenability to treatment,” “best interests of the public,” and “dangerousness” remain a central feature of judicial waiver laws (Forst & Blomquist, 1991). This leads to the argument that judicial waiver provides judges with broad, standardless discretion, which allows for inequities and disparities to occur (Feld, 1987, 1993). In contrast, supporters of this method contend that individualized transfer procedures provide the flexibility necessary for sound decision-making to take place (Fagan, 1990; Sanborn, 1994; Zimring, 1991). Research addressing this debate has revealed that an older age at the time of offense will often predict judicial certification to criminal court (Eigen, 1981a, 1981b; Fagan & Deschenes, 1990; Feld, 1989; Podkopacz & Feld, 1996; Poulos & Orchowsky, 1994). This practice of waiving older youths is undoubtedly influenced by the length of time remaining within the juvenile court’s jurisdiction (Feld, 1989). The juvenile court’s sanctioning powers are limited when a youth is approaching the maximum age for dispositional jurisdiction, and this restriction appears to increase the likelihood of transfer when a longer sentence is called for than can be imposed by a juvenile court judge. In addition, states vary dramatically in the age and offense restrictions they prescribe with judicial waiver laws (DeFrances & Strom, 1997; Sickmund et al., 1997; Torbet et al., 1996;). Even within a single state or county, it is unlikely that judges interpret and apply waiver laws consistently. This appears to create a situation of “justice by geography,” in which location has a significant influence on the transfer decision (Fagan & Deschenes, 1990; Feld, 1987, 1989; Hamparian et al., 1982; Podkopacz & Feld, 1996; Poulos &

Cracking Down On Youth Violence

23

Orchowsky, 1994). For example, it appears that rural youths are more likely to be judicially waived than comparable urban offenders (Feld, 1989; Hamparian et al., 1982; Poulos & Orchowsky, 1994). This is perhaps due to a higher threshold in urban juvenile courts for defining serious offending, or it may be that rural juvenile courts have lesser access to treatment programs and secure correctional facilities. Finally, a juvenile’s race may affect a judge’s waiver decision (Feld, 1987, 1993; Hamparian et al., 1982; Keiter, 1973). A study by Bortner (1986) revealed that in an unnamed western metropolitan county, Blacks comprised only 9% of all juveniles referred for serious offenses, but 24% of those juveniles judicially waived. Eigen (1981a, 1981b) found evidence of a possible “race-of-victim effect,” whereby Black juvenile homicide defendants in Philadelphia were more likely to be transferred if the victim was White. However, more recent researchers have employed multivariate statistical techniques to better control for critical legal variables (i.e., offense seriousness and prior record) and have subsequently failed to find direct racial bias toward minority youths (Fagan & Deschenes, 1990; Fagan et al., 1987; Podkopacz & Feld, 1995, 1996; Poulos & Orchowsky, 1994). Still, based on probable differential treatment earlier in the juvenile justice process (e.g., at the time of arrest; see Smith, 1986), Fagan and his colleagues (1987) concluded, “it appears that the effects of race are indirect, but visible nonetheless” (p. 276). Prosecutorial Waiver Sometimes referred to as concurrent jurisdiction, this method allows a prosecutor to file charges in either juvenile or adult court. Prosecutorial waiver is often limited to some combination of specific offenses (e.g., violent crimes), juveniles of a certain age, and youths with prior adjudications in juvenile court. Furthermore, states allowing for this practice commonly permit the criminal court judge to send a case back to juvenile court if deemed appropriate, a process known as “reverse waiver” (DeFrances & Strom, 1997). Prosecutorial waiver is utilized in a relatively small number of states, and it is probably the most controversial method of transfer (Champion & Mays, 1991). Currently, 10 states and the District of Columbia have concurrent jurisdiction statutes (DeFrances & Strom, 1997). Policy-makers may seek to increase the number of juveniles

24

Excluding Violent Youths From Juvenile Court

transferred to adult court by enacting this practice into law, but only a handful of states have added or expanded prosecutorial waiver statutes in recent years (Sickmund et al., 1997; Torbet et al., 1996). Research on prosecutorial waiver has been confined, for the most part, to the state of Florida (Bishop & Frazier, 1991, Bishop et al., 1989; Bishop et al., 1996; Howell, 1997; Thomas & Bilchik, 1985; Winner et al., 1997). Prosecutors appear to enjoy the authority they are granted with this procedure and feel it expedites the transfer process. However, critics contend that prosecutors can be too easily influenced by a perceived public demand for punishment, especially since there is no opportunity for appellate review (Bishop & Frazier, 1991; Bishop et al., 1989). Furthermore, the shift in discretion that takes place, from juvenile court judges to prosecutors (Sanborn, 1994), may actually introduce additional variability among jurisdictions in the use of transfer (Bishop & Frazier, 1991; Bishop et al., 1989). Legislative Waiver Over the past 15 years, literature has been building in opposition to both judicial and prosecutorial waiver, with calls for more objective criteria to be applied to the transfer process (Bishop & Frazier, 1991; Bishop et al., 1989; Fagan & Deschenes, 1990; Feld, 1987; Osbun & Rode, 1984). Perhaps in response to these efforts, there has been increasing political support for a third method of transfer that seeks to eliminate discretionary decision-making by judges and prosecutors. 14 Legislative waiver, or statutory exclusion, places eligible youths into the adult system at the time of arrest, thereby removing the initial discretionary power of juvenile officials (Singer, 1993, 1996). Legislative waiver laws, which emphasize the use of offense criteria in determining the appropriate court of jurisdiction, are an 14

Sanborn (1994) argued that judicial and prosecutorial waiver are the only two methods of transfer, as legislative waiver simply extends the power of court officials to transfer youths. According to his reasoning, legislative waiver only occurs when a prosecutor charges the necessary offense, and this amounts to a limited form of prosecutorial waiver. Also, he stated that lowering the maximum age of juvenile court jurisdiction should not be regarded as legislative waiver (see Forst & Blomquist, 1991), as youths over the maximum age can no longer be defendants in the juvenile system.

Cracking Down On Youth Violence

25

indicator of the shift in juvenile justice from an individualized treatment philosophy to a more retributive orientation (Feld, 1987, 1993). Supporters of this policy contend that, as compared to the highly discretionary practices of judicial and prosecutorial waiver, there is improved uniformity in determining the correct court of jurisdiction. However, others point out that the process can be rigid and overinclusive, as some excluded offenders may benefit from the treatment services offered in the juvenile system (Sanborn, 1994; Singer, 1996). Furthermore, legislative waiver may simply switch discretionary decision-making from the juvenile court to the adult criminal court (Zimring, 1991). This argument is based on the fact that most states with legislative waiver provisions also allow cases to be waived back to juvenile court for adjudication, disposition, or both (DeFrances & Strom, 1997). Also, prosecutors may reduce or dismiss charges, thereby eliminating criminal court jurisdiction. Legislatures in 36 states and the District of Columbia have excluded certain offenses, offenders, or both from juvenile court jurisdiction (DeFrances & Strom, 1997). The most commonly identified crimes are murder and other serious crimes against persons, while youth who are charged with a felony and have a prior adjudication of delinquency may also be excluded. Over the past few years, states that made changes to the boundaries of their juvenile court jurisdictions most often did so through the use of statutory exclusion (Feld, 1987; Fritsch & Hemmens, 1995; Sickmund et al., 1997; Torbet et al., 1996). It is important to note that the most rapidly expanding method of transfer, legislative waiver, is also the one that has received the least amount of attention from the research community (Howell, 1997). The increase in the use of legislative waiver laws, which tend to focus on serious and violent offenders, corresponds to a shift in the offense characteristics of juveniles transferred to adult court. Ironically, a great deal of past research revealed that transferred youths were not usually charged with violent offenses or crimes against persons, but they instead appeared to be chronic property offenders (Bishop & Frazier, 1991; Bishop et al., 1989; Bortner, 1986; Champion, 1989; Hamparian et al., 1982; Lemmon et al., 1991; Thomas & Bilchik, 1985). Related to these findings, in a study of judicial waiver decisions for violent youths in four urban juvenile courts, Fagan and Deschenes (1990) concluded, “Violent youths account for less than one-third of all youths transferred, yet they are a central focus of juvenile justice

26

Excluding Violent Youths From Juvenile Court

attention. Accordingly, initial expectations that virtually all violent youths would be transferred were not met” (p. 345). Although this research seemed to indicate that violent youths did not make up a majority of those transferred, more recent studies show that this situation has changed. Evidence from a variety of sources demonstrates that juveniles charged with personal or violent crimes now make up the largest percentage of waived cases (Barnes & Franz, 1989; DeFrances & Strom, 1997; Houghtalin & Mays, 1991; Kinder et al., 1995; Podkopacz & Feld, 1996; Sickmund et al., 1997). In discussing judicially waived cases across the nation, DeFrances and Strom (1997) reported, “Between 1985 and 1994, the profile of cases waived has changed considerably. Until 1991, property offenses comprised the largest number of cases judicially waived. Since 1991, however, offenses against persons have outnumbered property offenses as the most serious charge” (p. 5). Similarly, in discussing a 41% increase in judicial waivers between 1989 and 1993, Howell (1997) observed, “The proportion of juvenile referrals judicially waived to criminal courts that were violent offense cases increased from 28% in 1989 to 42% in 1993. Property offense cases decreased from 49% of the cases waived in 1989 to 38% in 1993” (p. 105). It is evident that the dominant type of crimes committed by transferred juveniles has changed from property offenses to personal or violent offenses. Furthermore, this switch corresponds well with the growing concern over youth violence during the past 10 years. With the increasing adoption and expansion of legislative waiver laws, it is likely that violent offenses will continue to be targeted for transfer, as they are the charges most commonly excluded from juvenile court jurisdiction (DeFrances & Strom, 1997).

Transfer Use and Rationale From 1992 through 1995, 41 states passed laws seeking to ease the transfer of juveniles to adult court (Sickmund et al., 1997; Torbet et al., 1996). As stated above, these modifications most often involved the use of statutory exclusion, as over 30 states either established or expanded their legislative waiver laws during this time period (Sickmund et al., 1997, p. 30). Early evidence indicates that these changes in state statutes have had the desired effect of sending a greater number of youthful offenders to criminal court.

Cracking Down On Youth Violence

27

National estimates of the number of juvenile offenders who are waived to adult court are available only for those youths who are transferred by judicial waiver (DeFrances & Strom, 1997; Sickmund et al., 1997). From 1988 to 1994, the total number of cases waived to adult court by a juvenile court judge rose from 7,000 to 12,300, an increase of 75% (DeFrances & Strom, 1997). However, as noted by Howell (1997, p. 108), judicially waived cases only represent roughly 10% of all juveniles who are transferred to the adult system. The other 90% are tried in adult court under legislative and prosecutorial waiver 15 provisions. Even though complete national statistics are not available on youths who are statutorily excluded from juvenile court jurisdiction, police dispositions of juvenile offenders that result in referral to adult court would appear to reflect the use of legislative waiver, as these youths are treated as adults at the time of arrest. Therefore, with the current expanded use of legislative waiver, one would expect that police would refer an increasing number of juvenile offenders to adult court, and an increasing percentage of police dispositions of juvenile offenders would result in an adult court referral. Figures obtained from the Uniform Crime Reports support these expectations. More than 170,000 juvenile offenders were referred to adult court by police in 1996 and 1997, representing over 6% of the total police dispositions for juveniles during those years (Federal Bureau of Investigation, 1997, 1998). In 1995, police referred less than 45,000 juvenile offenders to adult court, representing only slightly more than 3% of the total police dispositions for juveniles during that year (Federal Bureau of Investigation, 1996). Current national statistics are also not available on young offenders who end up in adult court as a result of prosecutorial waiver, but in states that employ this method, juveniles waived by a prosecutor are likely to outnumber judicially waived youths by a large margin. For example, as discussed by Snyder and Sickmund (1995, p. 156), recent 15

It should also be noted that 13 states have set the upper age limit for juvenile court jurisdiction at 15 or 16, rather than 17. Although not generally defined as transfers, approximately 180,000 cases involving youths either 16 or 17 years of age were tried in adult criminal court in 1994 because they were legally defined as adults in their respective states (Sickmund et al., 1997, p. 30).

28

Excluding Violent Youths From Juvenile Court

estimates in Florida indicated that prosecutorial transfers accounted for more than 80% of the youths under the age of 18 who were handled in adult court in that state in 1993. Furthermore, as another sign of the increasing use of transfer, waivers in Florida increased 216% between 1981 and 1993. Finally, it is noteworthy that an estimated 2,000 prosecutorial transfers occurred nationwide in 1982, while in 1993, Florida prosecutors alone filed charges in adult court for approximately 7,000 cases involving offenders under the age of 18. In sum, the figures discussed above illustrate the growing trend toward treating more and more juvenile offenders as adults. The basic rationale for this practice is that the juvenile court is unable to serve the needs of certain young people, and therefore, the criminal court should take over their cases. The perceived inability of the juvenile court to handle certain cases may be based on a lack of faith in juvenile correctional facilities, a belief that harsher punishment is needed than can be provided in the juvenile system, or the position that a youth is too dangerous to remain outside the criminal system. In spite of the recent shift in juvenile justice philosophy, from the rehabilitative ideal to a more punitive orientation, the decision to transfer a case still denotes that a youthful offender is beyond whatever treatment capacity remains in the juvenile justice system. Ironically, support for treating juvenile offenders as adults comes from both sides of the political spectrum. “Due process liberals” (Thomas & Bilchik, 1985) believe that the informal and discretionary atmosphere that remains in many juvenile courts can breed discrimination and violation of constitutional rights. Consequently, they argue that rules of procedure in juvenile courts should be essentially the same as those in criminal court, and more emphasis should be placed on the offense rather than the offender. Suggestions have even been made to do away with the juvenile court altogether (Feld, 1993, 1998). On the other hand, “crime control conservatives” (Thomas & Bilchik, 1985) believe that many of today’s youthful offenders are no different than adult criminals, and that the juvenile system provides neither effective rehabilitation nor adequate punishment required to ensure public safety. Therefore, they assert that these youths need to be prosecuted in the criminal system, where the goals of retribution, deterrence, and incapacitation can best be met. Both groups attack the idea of a separate system of justice for juvenile offenders, but they use entirely different reasoning to support

Cracking Down On Youth Violence

29

their position. A third group can also be added to the mix: supporters of the traditional juvenile court who would like to maintain a distinct system. These individuals see current reforms as an effort to “recriminalize delinquency,” or return to the response to youthful offending that existed prior to the creation of juvenile courts (Singer, 1996). Also, some do not accept criticisms of the rehabilitative ideal, arguing instead that past evaluations of treatment programs were weak, and that many quality programs exist and produce positive results (Fagan, 1990; Howell, 1997; Krisberg & Austin, 1993). However, the nature of modern juvenile justice reforms shows that this third party has not had a substantial influence on public policy.

Juvenile Transfer in Pennsylvania Recent legislative actions in Pennsylvania typify the way many states have enacted laws pertaining to a variety of juvenile justice issues, including the criminal prosecution of youthful offenders. Amid much political fanfare, legislators in Pennsylvania modified their juvenile code, and in doing so moved away from judicial waiver and toward statutory exclusion. This was done with the expectation that the number of juveniles processed in adult criminal court would increase substantially. Prior to March 1996, Pennsylvania’s Juvenile Act (Pennsylvania Juvenile Court Judges’ Commission, 1992), in Section 6355, provided that the transfer of juveniles to criminal court could occur through either judicial waiver or statutory exclusion. However, the only crimes that fell under the legislative waiver component were those of murder and crimes committed by youths who were previously found guilty in a criminal proceeding for a misdemeanor or felony. Hence, for those juveniles who were never previously found guilty in adult court, the act of murder constituted the only crime for which they could be statutorily excluded from juvenile court jurisdiction. Furthermore, the law also allowed for statutorily excluded youths to be returned to juvenile court for adjudication, disposition, or both (i.e., reverse waived), basically at the criminal court’s discretion. The more common method of transfer was judicial waiver, according to the standards set forth in Kent v. United States (1966) and Breed v. Jones (1975). Prior to transfer or any determination of guilt, a formal waiver hearing had to be held and a number of criteria had to be

30

Excluding Violent Youths From Juvenile Court

met, as specified in Section 6355 of the Juvenile Act (Pennsylvania Juvenile Court Judges’ Commission, 1992). Objective criteria included a minimum age of 14 at the time of the alleged incident, the establishment of a prima facie case against the juvenile, and the classification of the alleged offense as a felony. If those criteria were satisfied, the juvenile court judge then had to determine that the youth was no longer amenable to treatment in the juvenile system. To do so, the judge was to consider the juvenile’s age, mental capacity, maturity, degree of criminal sophistication, prior delinquent involvement, prior treatment experience, chances for rehabilitation, nature of the current offense, and any other relevant factors. If the judge found that the youth was no longer amenable to treatment in the juvenile system, the case could be waived to adult court. In addition, the judge could transfer the proceedings at the request of the juvenile, as this request allowed the court to order the listed criteria as not applicable. In short, Pennsylvania’s juvenile code granted judges the type of broad discretionary power to transfer cases that has been criticized in the past (Feld, 1987, 1993). Under this system, from 1985 through 1995, approximately 1% of the total number of juvenile court dispositions in Pennsylvania resulted in a judicial waiver (Lemmon et al., 1991; Myers, 1997). In terms of total judicial waivers during this time period, there was a gradual increase from 227 in 1985 (0.78% of the total dispositions), to 375 in 1990 (1.06% of the total dispositions), and finally to 533 in 1995 (1.44% of the total dispositions) (Myers, 1997). In 1995, significant legislative attention was given to the juvenile justice system in Pennsylvania. Subsequently, the Commonwealth’s Juvenile Act (Pennsylvania Juvenile Court Judges’ Commission, 1992) was modified. New legislation was enacted that pertained to such matters as the use of juvenile records in determining bail, fingerprinting of juveniles, expungement of juvenile records, public access to delinquency hearings, restitution by youthful offenders, creation of a DNA database, background checks for gun purchases, parental participation in treatment programming, parental responsibility for truancy, and reporting of delinquency dispositions. However, probably the most important change concerned the prosecution of youthful offenders in adult criminal court. Overall, the new Juvenile Act (Pennsylvania Juvenile Court Judges’ Commission, 1996) was structured to meet several objectives. This is

Cracking Down On Youth Violence

31

evident in an opening statement of the code, which states that a purpose of the act is: Consistent with the protection of the public interest, to provide for children committing delinquent acts programs of supervision, care and rehabilitation which provide balanced attention to the protection of the community, the imposition of accountability for offenses committed and the development of competencies to enable children to become responsible and productive members of the community. (Pennsylvania Juvenile Court Judges’ Commission, 1996, p. 1) As part of this “balanced attention,” the new juvenile code excluded from juvenile court jurisdiction any youth charged with murder or any of the following violent offenses, when the child was between 15 and 18 years of age at the time of the alleged offense and a 16 deadly weapon was used during the offense: rape, involuntary deviate sexual intercourse, aggravated assault, robbery, robbery of a motor vehicle, aggravated indecent assault, kidnapping, voluntary manslaughter, or an attempt, conspiracy, or solicitation to commit murder or any of these other listed offenses (Pennsylvania Juvenile Court Judges’ Commission, 1996, pp. 2-3). The new act also excluded any youth between 15 and 18 years of age at the time of the alleged offense who commits any of the above listed crimes, except for aggravated assault, after previously being adjudicated delinquent on the 17 basis of any of these offenses. Also, it is noteworthy that the new juvenile code allows for all excluded cases to be transferred back to

16

Defined by 18 PA C.S. Sec. 2301 as any firearm, whether loaded or unloaded, or any device designed as a weapon and capable of producing death or serious bodily injury, or any other device or instrumentality which, in the manner in which it is used or intended to be used, is calculated or likely to produce death or serious bodily injury. 17 Aggravated assault was intentionally omitted from this repeat violent offender clause, apparently based on the belief of policy-makers and practitioners that aggravated assault without a deadly weapon could encompass too many violent but less-serious acts (e.g., a fist-fight at school).

32

Excluding Violent Youths From Juvenile Court

juvenile court, if the offender establishes by a preponderance of the evidence that a reverse waiver will serve the public interest. In addition to excluding the above-designated violent felonies from juvenile court jurisdiction, the new act still allows for discretionary judicial waiver of cases (Pennsylvania Juvenile Court Judges’ Commission, 1996, pp. 32-34). Here, the burden of establishing by a preponderance of the evidence that the public interest is served by transferring the case to criminal court rests with the prosecution, unless a deadly weapon was used and the youth was 14 at the time of the offense, or the youth was 15 or older, was previously adjudicated delinquent on the basis of a felony, and is alleged to have committed any of a series of designated felonies set forth in the juvenile act. If either of these last two criteria is met, the burden of establishing by a preponderance of the evidence that retaining the case in juvenile court would serve the public interest rests with the juvenile. As with the former juvenile code, the new act provides for two transfer mechanisms: judicial waiver and statutory exclusion. However, the focus is now clearly on the exclusion of violent offenders from juvenile court jurisdiction. Youths between 15 and 18 years of age who commit a violent offense with a deadly weapon, as well as some repeat violent offenders, essentially receive “automatic adulthood” (Feld, 1993, p. 239). In addition, a presumptive waiver component was added to the traditional judicial waiver process, again with a focus on offenses committed with a deadly weapon and repeat offenders. Overall, the new juvenile code in Pennsylvania corresponds well with the nationwide shift toward a more retributive model of juvenile justice, as it emphasizes holding youths accountable for violent offending. Whether this type of approach is an effective response to youth violence remains to be seen.

CHAPTER 2

The Deterrent Effects Of Formal Sanctions

In recent times, politicians have reacted to public concerns and fear of crime by supporting various “get tough” policies. Popular strategies include “three strikes and you’re out” laws, “truth in sentencing” provisions, expanded use of the death penalty, boot camps, and stricter law enforcement (DiIulio, Smith, & Saiger, 1995; Wilson, 1995). While these measures may be backed by arguments of increased retribution and incapacitation, they are also based on the belief that punishment deters criminal behavior. This idea is very appealing to the general public. Not only is it assumed that formal sanctions have a deterrent effect, but also that harsher punishments are needed: Despite their good instincts for the right answers, the people, frustrated by the restraints (many wise, some foolish) on swiftness and certainty, vote for proposals to increase severity: if the penalty is ten years, let us make it twenty or thirty; if the penalty is life imprisonment, let us make it death; if the penalty is jail, let us make it caning. (Wilson, 1995, p. 494) Transferring juveniles to adult criminal court fits in well with this view. Supporters of this practice contend that adult court is the appropriate place for youthful offenders who exhibit serious criminal behavior. In adult court, it is argued, a message is sent that the lenient treatment of the juvenile system is no longer an option (Bishop et al., 1996). Instead, harsh criminal court sanctions will be imposed, which will increase public safety and reduce motivations to commit future

33

34

Excluding Violent Youths From Juvenile Court

crimes. In sum, adult criminal court is believed to provide stronger punishment and greater deterrence. These underlying beliefs suggest that the effectiveness of treating juvenile offenders as adults is based on the nature of the punishment it produces (i.e., the certainty, severity, and swiftness of the punishment) and subsequent criminal behavior (Fagan, 1995). To the extent that this is true, deterrence theory would appear to provide an excellent framework for studying a policy of this type. This chapter will briefly review the theoretical origins of deterrence theory and discuss how scholars have addressed its central propositions. In addition, the specific deterrent effects of formal sanctions will be closely examined.

Origins of Deterrence Theory Deterrence theory can be traced to the development of the classical th school of criminology in the latter half of the 18 century. According to classical criminology, humans are rational beings who are guided by their own free will. Therefore, both criminal and law-abiding behavior results from conscious choice. Based on this underlying belief, Cesare Beccaria (1764/1986) proposed a more rational system of justice in his treatise On Crimes and Punishments. In reaction to the often arbitrary and cruel systems of justice that were in place during the 1700’s, Beccaria presented a series of criminal justice reforms. His proposals covered such topics as making laws public and simple to understand, eliminating the torture of suspects, a presumption of innocence until proven guilty, equality under the law, and abolishing the death penalty. Furthermore, he stressed that the purpose of punishment should be deterrence, and to achieve maximum deterrence, punishment should be based on the principles of certainty, severity, and celerity. Because subsequent deterrence theorists and researchers have devoted their attention to the three principles of punishment first identified by Beccaria, it is important to examine them in more detail. Although the celerity, or swiftness, of punishment has been generally ignored in modern deterrence literature, Beccaria placed great emphasis on it: The more prompt the punishment is and the sooner it follows the crime, the more just and useful it will be. I say more just, because it spares the criminal the useless and cruel torments of

The Deterrent Effects of Formal Sanctions

35

uncertainty, . . . I have said that promptness of punishment is more useful, for the less time that passes between the misdeed and its chastisement, the stronger and more permanent is the human mind’s association of the two ideas of crime and punishment, so that imperceptibly the one will come to be considered as the cause and the other as the necessary and inevitable result. (Beccaria, 1764/1986, p. 36) In addition to the importance of swiftness, Beccaria also stressed that punishments should be a certain response to criminal behavior: One of the greatest checks on crime is not the cruelty of punishments but their inevitability. The certainty of a chastisement, even if it be moderate, will always make a greater impression than the fear of a more terrible punishment that is united with the hope of impunity; for, when they are certain, even the least of evils always terrifies men’s minds, while hope, that heavenly gift that often fills us completely, always removes from us the idea of worse punishments, especially if that hope is reinforced by the examples of impunity which weakness and greed frequently accord. (Beccaria, 1764/1986, p. 46) Finally, in addressing the importance of certain punishments, Beccaria also cautioned against punishments that were too severe: The very savagery of a punishment makes the criminal all the bolder in taking risks to avoid it precisely because the evil with which he is threatened is so great, so much so that he commits several crimes in order to escape the punishment for a single one of them. In order for a penalty to achieve its objective, all that is required is that the harm of the punishment should exceed the benefit resulting from the crime. Everything more than this is thus superfluous and therefore tyrannical. (Beccaria, 1764/1986, p. 46) These three principles of punishment continue to receive attention today, from criminologists, politicians, and the general public. However, contemporary scholars have emphasized the certainty and—

36

Excluding Violent Youths From Juvenile Court

to a lesser extent—the severity of punishment, while the swiftness of punishment has received little consideration (Andenaes, 1974; Blumstein, Cohen, & Nagin, 1978; Gibbs, 1975; Paternoster, 1987; 18 Zimring & Hawkins, 1973). Citizens and politicians tend to be even more narrowly focused, placing their support squarely behind those policies that promise the harshest punishment (Wilson, 1995). This emphasis on severe punishment directly contradicts the writings of Beccaria. Furthermore, considering what is known about the deterrent effects of formal sanctions, policies that focus mainly on increasing the severity of punishment are highly questionable.

Deterrence Research Although deterrence theory is over 200 years old, its propositions have been empirically tested only during the past few decades (Gibbs, 1975; Paternoster, 1987; Zimring & Hawkins, 1973). Over this time, researchers have distinguished between two types of deterrence: 19 specific deterrence and general deterrence. Specific deterrence pertains to the effect of punishment on the behavior of the individual that is sanctioned. As stated by Andenaes (1968), when someone is “deterred by the actual experience of punishment, we speak of special [specific] deterrence” (p. 78). In contrast, general deterrence refers to the effect of punishment on potential offenders in the community. In other words, general deterrence is the “imposition of sanctions on one person [in order to] demonstrate to the rest of the public the expected costs of a criminal act, and thereby discourage criminal behavior in the general population” (Nagin, 1978, p. 96). Overall, specific deterrence 18

The lack of attention given to celerity may be due to ambiguity surrounding the potential effect of swift punishment. Although immediate sanctions may correspond well with increased punishment certainty, delayed sanctions may increase severity, due to the time spent thinking about the impending punishment. Therefore, two contrasting predictions exist: (a) swift punishments provide greater deterrence, and (b) delayed punishments provide greater deterrence. 19 This distinction can also be traced to Beccaria (1764/1986), who wrote, "The purpose of punishment, then, is nothing other than to dissuade the criminal from doing fresh harm to his compatriots and to keep other people from doing the same" (p. 23).

The Deterrent Effects of Formal Sanctions

37

has been thought to impact on offenders who have been caught and punished, while general deterrence has been applied to those in the 20 general public who have not yet offended or experienced punishment. 21 Studies on deterrence theory are usually classified as pertaining to one type of deterrence or the other, with most involving general rather than specific deterrence (Stafford & Warr, 1993). Early research focused on deterrence at the objective level, whereby the relationship between aggregate levels of risk and officially recorded crime rates was explored (Nagin, 1978). Researchers correlated state level variations in punishment certainty (measured as the aggregate risk of arrest or imprisonment for a specific crime) and severity (measured as average time served for the offense) with variations in officially recorded crime rates (Paternoster, 1987). These studies generally found inconsistent support for the deterrence doctrine (i.e., the proposed inverse relationship between the certainty and severity of punishment and crime rates), but it did appear that the certainty of punishment had a greater deterrent effect than punishment severity (Chiricos & Waldo, 1970; Gibbs, 1968, 1975; Logan, 1975; Tittle, 1969; Tittle & Rowe, 1974; Zimring & Hawkins, 1973). However, critics raised the issues of correct temporal ordering (Logan, 1975) and possible measurement errors in crime data that could produce spurious negative correlations between arrest or incarceration rates and crime rates (Nagin, 1978). Perhaps in response to these problems, researchers soon discussed the possibility that deterrence could depend more on individual 20

Stafford and Warr (1993) reconceptualized deterrence as a process based on the effects of four kinds of experience: direct experience with punishment and punishment avoidance (i.e., specific deterrence) and indirect experience with punishment and punishment avoidance (i.e., general deterrence) (see also Paternoster & Piquero, 1995). However, most deterrence scholars continue to differentiate between specific and general deterrence (see, e.g., DeJong, 1997; Grasmick, Bursik, & Arneklev, 1993; Miller & Iovanni, 1994; Ward & Tittle, 1993; Yu, 1994). 21 Modern deterrence scholars have defined the "deterrence doctrine" as a series of three propositions concerning the inverse relationship between punishment threats and criminal behavior: "The greater the celerity, certainty, and severity of punishment for a type of crime, the more individuals are deterred from that type of crime" (Gibbs, 1975, p. 15; see also Paternoster, 1989b, p. 7).

38

Excluding Violent Youths From Juvenile Court

perceptions of the certainty and severity of punishment rather than on the objective levels of punishment certainty and severity (Geerken & Gove, 1975; Waldo & Chiricos, 1972). This recognition of deterrence as a perceptual theory has guided most of the modern deterrence research. By the early 1980s, numerous cross-sectional studies appeared to indicate a moderately strong inverse relationship between perceptions of certain punishment and various criminal behaviors (for a thorough review, see Paternoster, 1987). These same studies generally failed to find a similar inverse association between the perceived severity of punishment and illegal acts. Initial perceptual studies were also soon questioned on the basis of temporal ordering. Researchers demonstrated with panel data that active participants in crime often do not get caught, which leads to lower estimates of risk of punishment (Minor & Harry, 1982; Paternoster, Saltzman, Chiricos, & Waldo, 1982; Paternoster, Saltzman, Waldo, & Chiricos, 1983; Saltzman, Paternoster, Waldo, & Chiricos, 1982). The earlier perceptual studies were actually measuring an “experiential effect” of prior behavior on current perceptions of risk, rather than a “true” deterrent effect of perceived risk on behavior. When the experiential and true deterrent effects were separated using longitudinal data, a much weaker deterrent effect from perceived certainty of punishment was found than earlier research had indicated. Furthermore, panel studies that also controlled for greater numbers of other explanatory variables found the least evidence of a deterrent effect from the perceived risk of formal sanctions (Paternoster, 1989b; Paternoster & Iovanni, 1986; Paternoster et al., 1983; Piliavin, Gartner, Thorton, & Matsueda, 1986). By the mid-1980s, perceptual deterrence research indicated little reason to be optimistic about the deterrent effects of formal sanctions. The null findings generally obtained from the most methodologically rigorous studies led to the conclusion that, “no matter how sophisticated the study or how valiant the effort, very little relationship may exist between people’s estimates of the certainty and severity of punishment and their behavior” (Paternoster, 1987, p. 214). However, panel studies had also revealed that risk perceptions were unstable, as shown by generally low correlations between the perceptual variables at different time points (Paternoster, 1987). This finding led to a major improvement in the most recent deterrence research, resulting in conclusions that are more supportive of the deterrence doctrine.

The Deterrent Effects of Formal Sanctions

39

Over the past decade, researchers have used scenarios to elicit perceptions of the risk of punishment and behavioral intentions at the same time. This method permits respondents to consider detailed information regarding proposed criminal behavior. Presumably, this allows for a more informed response to questions of sanction risk and projected behavior (Klepper & Nagin, 1989). The use of scenarios also overcomes the problems of experiential effects and instability of risk perceptions that plagued earlier cross-sectional and panel studies. Furthermore, recent research has addressed the “science of sophomores” (Paternoster, 1987, p. 214) criticism which was applied to past studies that only examined student samples and minor forms of offending (for an exception, see Piliavin et al., 1986). Scenario studies have surveyed young adults and older individuals who would be adversely affected by criminal prosecution, and they have used a rational choice perspective to consider both the costs and incentives for criminal behavior (Cornish & Clarke, 1986, 1987). With these improvements, modern deterrence research has demonstrated that perceived threats of punishment do operate as a deterrent, at least for some people. While controlling for such factors as potential rewards, moral beliefs, informal sanctions, peer delinquency, possible victim reaction, and low self-control, perceived sanction risk has been found to have a significant deterrent effect on a variety of criminal behaviors (Bachman, Paternoster, & Ward, 1992; Grasmick & Bursik, 1990; Klepper & Nagin, 1989; Nagin & Paternoster, 1993; Piquero & Tibbetts, 1996). These studies show that earlier research, which suffered from methodological problems and often failed to find support for the deterrence doctrine, may have been premature in dismissing the deterrent effect of formal sanctions, especially with regard to the perceived certainty of punishment. Overall, contemporary research on deterrence theory has been concerned primarily with the inhibiting effect of perceptions of legal punishments on criminal behavior. A related—but much less studied— issue concerns the factors that influence the formation of these perceptions (Miller & Iovanni, 1994). If, in fact, perceptions of punishment risk are inversely related to involvement in proscribed behavior, then it would seem important to know how individual experiences with crime and subsequent formal sanctions affect perceptions of punishment risk (Apospori, Alpert, & Paternoster, 1992; Horney & Marshall, 1992). As argued by Geerken and Gove (1975),

40

Excluding Violent Youths From Juvenile Court

“the success of any deterrence process will be determined by the degree to which [a] message is successfully transmitted to the population of offenders” (p. 499). Earlier panel studies revealed a lowering of risk estimates among offenders actively engaging in crime, which was interpreted as an experiential effect (Minor & Harry, 1982; Paternoster et al., 1982; Paternoster et al., 1983; Saltzman et al., 1982). However, if offenders are caught and punished on a regular basis, this experiential effect might be modified (Horney & Marshall, 1992). These offenders may raise their perceptions of punishment risk and, consequently, curtail their offending. This same process may occur when an offender gains knowledge of punishment of criminal behavior from other associates (Stafford & Warr, 1993; Paternoster & Piquero, 1995). Unfortunately, research on the effect of formal sanctions on perceptions of risk has produced equivocal results. A few studies have found evidence that experiencing some type of formal sanction leads to greater estimates of sanction risk (Horney & Marshall, 1992; Paternoster, Saltzman, Waldo, & Chiricos, 1985; Piliavin et al., 1986). Other research has produced less supportive findings, showing that formal sanctions were related to lower estimates of sanction risk or that no significant relationship existed (Apospori et al., 1992; Bridges & Stone, 1986; Cohen, 1978; Lanza-Kaduce, 1988; Paternoster & Piquero, 1995; Richards & Tittle, 1981, 1982; Schneider & Ervin, 1990; Thomas & Bishop, 1984). In sum, the effect of legal punishments on perceptions of risk is open to debate, but the weight of the evidence is not supportive of a positive relationship. Furthermore, the findings from this research serve as a prelude to another important area of study, which is the specific deterrent effect of legal punishments on future criminal behavior.

The Specific Deterrent Effect of Formal Sanctions While research has addressed the issues of the effect of personal perceptions of punishment on criminal behavior and the effect of experiences with punishment on personal perceptions, probably the most important question for the current study is how experiences with punishment affect future criminal behavior. As mentioned earlier, a major justification for transferring juveniles to adult court is that the

The Deterrent Effects of Formal Sanctions

41

criminal court can provide stronger punishment and subsequently reduce inclinations toward future criminal behavior. Specific deterrence, in its simplest form, occurs when individuals who have experienced a punishment refrain from committing crime in the future, because of the fear of further punishment (Andenaes, 1968). This process can be expanded to include not only a cessation in offending, but also a reduction in the seriousness of offenses committed or in the rate at which they occur (Paternoster & Piquero, 1995, p. 251; see also DeJong, 1997; Gibbs, 1975; Paternoster, 1989a). According to the deterrence doctrine, individuals who experience a more certain, swift, and severe punishment will be more likely to reduce or refrain from future offending than those who are punished with less certainty, severity, and swiftness (Gibbs, 1975; Paternoster, 1989b). Not coincidentally, recent public opinion surveys indicate that citizens believe adult criminal courts can provide punishment that is more certain, swift, and particularly more 22 severe than that of juvenile courts (Schwartz, Guo, & Kerbs, 1993). In turn, this forceful punishment is expected to have a beneficial impact on future criminal offending. Unfortunately, the question of how legal punishments affect future criminal behavior is not an easy one to answer. An enduring debate in the field of criminology concerns whether formal sanctions reduce or amplify future criminal offending (Sherman, 1993; Smith & Gartin, 1989). Deterrence theory has been opposed by labeling theory, which argues that legal punishments can increase future offending among those sanctioned (Becker, 1963; Lemert, 1951, 1972;). This increase in criminal behavior may be due to diminished opportunities for success in legitimate activities, or it may be caused by a process of selfidentification and value development in which the individual that is “labeled” adopts the norms and behavior patterns that are characteristic of the label (Paternoster & Iovanni, 1989; Thomas and Bishop, 1984). In any case, “a fundamental empirical prediction of labeling theory is that being sanctioned or negatively labeled will increase one’s 22

Although the public seems to strongly favor the practice of sending serious and violent juvenile offenders to adult court, where the administration of harsher punishment is expected, it should also be noted that these same surveys show citizen opposition to confining juveniles in adult prisons and allowing the use of capital punishment for juvenile offenders (Schwartz et al., 1993).

42

Excluding Violent Youths From Juvenile Court

involvement in future deviant conduct—a deviance amplification effect” (Smith & Paternoster, 1990, pp. 1109-1110). While studying the effects of legal sanctions requires recognition of two contradicting behavioral predictions, the task is made even more difficult by the potential obstacle of selection bias (Smith & Paternoster, 1990). A simple test of whether formal punishments have any effect on subsequent criminal behavior could involve comparisons of future offending across groups who naturally received different court dispositions. However, this design is problematic: Assignment to treatment groups (diversion vs. referred to juvenile court, for example) is the result of a nonrandom process in which high-risk youth are more likely to receive severe dispositions. Thus, those individuals assigned more severe sanctions would be more likely to commit new offenses whether or not any relationship existed between juvenile court dispositions and future offending. (Smith & Paternoster, 1990, pp. 1111-1112) The difficulty of selection bias can be approached in various fashions. In conjunction with large samples, the ideal way to assure the comparability of treatment and control groups is to use random assignment of subjects into these groups (Farrington, 1983). 23 Unfortunately, this is not always possible in justice system research. Therefore, other methods have been employed, although the danger of selection bias remains. One strategy is to use a matching design, whereby the subjects in treatment and control groups are matched on the basis of specified variables. The major problem with this method is obtaining large groups matched on more than a few relevant variables (Smith & Paternoster, 1990; but see Bishop et al., 1996, for an exception). An alternative, and more often used, strategy is to obtain statistical control over extraneous variables by using multivariate statistical models. The remaining difficulty here is the possible 23

This seems particularly true for research on juveniles transferred to criminal court. Experiments comparing youth randomly assigned to juvenile and adult court would most likely be met with strong political opposition, and, therefore, alternative strategies are required.

The Deterrent Effects of Formal Sanctions

43

omission of any variables that influence both the probability of formal sanctions being imposed and the likelihood of future offending, which could create a selection artifact (Smith & Paternoster, 1990). In sum, research on the specific deterrent effects of formal sanctions needs to be 24 interpreted with caution. Gold and Williams (1969) conducted an early study in this area. Using 35 matched pairs of juveniles who self-reported their delinquent activity, the authors tested the proposition that youths who are apprehended by police for an offense will have more subsequent offenses than comparable youths who are not apprehended. In 57% of the pairs, the apprehended youth reported more subsequent delinquency than the unapprehended youth (i.e., a deviance amplification effect). In 29% of the pairs, the apprehended youth committed fewer offenses (i.e., a specific deterrent effect). In the remaining 14% of the pairs, the two juveniles committed an equal number of offenses. Gold and Williams (1969) argued that these results support the hypothesis derived from labeling theory. However, their study presents various methodological weaknesses. Most obvious is the small number of matched pairs, which points to the difficulty in matching groups on the basis of several criteria. A total of 74 apprehended youths had been originally selected for the study, but it was possible to match only 35 with unapprehended juveniles. Furthermore, there was a lack of any control for the seriousness of reported prior offenses, which raises doubt about the similarity of the “matched” offenders. Overall, this study provides weak support for the argument that formal sanctions increase future delinquent behavior. Subsequent research by Klemke (1978) produced similar findings. Self-report data on shoplifting were collected from 1,189 high school students in the Pacific Northwest. The ensuing analysis revealed that 24

Because the current study focuses on the formal punishment of youthful offenders, the following discussion will emphasize research of a similar nature. For general reviews of the literature regarding specific deterrence and labeling, see Farrington (1983); Gibbs (1975); Mahoney (1974); Paternoster and Iovanni (1989); Sherman (1993); Wilson (1983); and Zimring and Hawkins (1973). Examples of recent studies examining the specific deterrent effect of formal punishments on adults include Baumer (1997); DeJong (1997); Joo, Ekland-Olson, and Kelly (1995); Makkai and Braithwaite (1994); and Yu (1994).

44

Excluding Violent Youths From Juvenile Court

apprehension by store personnel and by parents was directly related to future shoplifting activity and the identification of oneself as a “troublemaker,” as predicted by labeling theory. In addition, subsequent shoplifting and a deviant self-concept were much more strongly related to parental apprehension than to apprehension by store personnel. This finding is supported by more recent research that “highlights the importance of informal labels and their subjective perception” (Smith & Brame, 1994, p. 624; see also Matsueda, 1992; Ward & Tittle, 1993; Wellford & Triplett, 1993). Finally, youth who experienced police intervention exhibited greater subsequent shoplifting than those who did not go through police involvement. However, this study also employed a weak control for prior delinquency (an ordinal measure merely of recent shoplifting), which allows for the possibility that apprehended youths and those exposed to the police were simply more serious delinquents. A later study by Smith and Gartin (1989) improved upon the efforts of Gold and Williams (1969) and Klemke (1978). Using data on police contacts for all males born in Racine, Wisconsin, in 1949 and followed until age 25, the authors examined the influence of arrest on four different dimensions of subsequent criminal behavior: termination from future criminal behavior, future offending rate, duration of the future offending period, and time until next offense. The final sample consisted of 325 males who had at least one recorded police contact for a misdemeanor or felony. Contrary to earlier research, Smith and Gartin (1989) uncovered evidence that supported a specific deterrence hypothesis. With controls for the seriousness of the current offense and the extent of prior offending, being arrested was found to affect specific aspects of a criminal career differently. The combined effect was a reduction in future criminal behavior among those offenders who were arrested relative to those who were contacted by police but not arrested. Being arrested was associated with an increased probability of desistence among novice offenders, but as offenders acquired more police contacts, this effect of arrest on future police contacts diminished. Still, arrest was consistently related to reduced rates of future police contacts, and this finding became more pronounced as criminal careers persisted. Furthermore, for those offenders who persisted in their criminal behavior, arrest appeared to extend the duration of their criminal career, but also increase the time until next police contact. Finally, arrest was

The Deterrent Effects of Formal Sanctions

45

found to have a consistently negative effect on the total number of future police contacts. Smith and Gartin’s (1989) research provided strong evidence of a specific deterrent effect from formal sanctions. Their study was methodologically stronger than earlier works, with better controls for offense seriousness and prior offending and a comprehensive examination of future criminal activity. However, the measurement of criminal behavior was based solely on official records of police contacts, which underestimate true rates of offending. Also, official records were used to select a rather small sample of known offenders. A larger sample would have permitted the authors “to disaggregate police contacts by offense type and estimate models separately for different types of crimes or for felonies relative to misdemeanors” (Smith & Gartin, 1989, p. 103). Finally, the degree of actual punishment associated with arrest could have varied greatly among offenders, as prosecutors could have subsequently dismissed many cases. Thus, while support for a specific deterrent effect from legal punishment was presented, several avenues remained open for further investigation. A recent study by Paternoster and Piquero (1995) illustrates this point. Based on self-reported data obtained from over 1,400 high school students from nine schools in and around a southeastern city, the authors tested a reconceptualization of the deterrence doctrine presented by Stafford and Warr (1993). In doing so, they examined the effect of both personal experience with punishment and punishment avoidance and vicarious experience with punishment and punishment avoidance. Central to the current discussion, the authors found that those youths who reported past experience with being apprehended by police, taken to a police station, arrested, or taken to juvenile court were more likely to drink alcohol and use marijuana in the future than those who were not legally punished. The significant positive effect from formal sanctions remained even when the number of previous delinquent acts was controlled. In general, this finding contradicts the evidence presented by Smith and Gartin (1989) and supports the earlier results of Gold and Williams (1969) and Klemke (1978). However, the research by Paternoster and Piquero (1995) considered relatively minor forms of substance use and employed a sample of juveniles who had limited experience with the legal system. In addition, as with the previous studies, possible

46

Excluding Violent Youths From Juvenile Court

differences in the level of punishment actually experienced were not well measured. While the research discussed to this point focused on the future offending of juveniles who were apprehended relative to those who were not, other studies have looked at the effects of varying degrees of punishment on known offenders. In the much debated and somewhat controversial “Provo Experiment,” delinquent boys in Utah were randomly assigned to experimental and control groups (Empey & Erickson, 1972). The experimental intervention was community-based and provided intensive treatment services. Comparisons were made with a randomly selected group that had been placed on regular probation, as well as a matched group that was institutionalized in a training school. The effectiveness of the treatment program was evaluated based on arrests both during program participation and after the treatment had ended. Empey and Erickson (1972) reported that boys in the communitybased program were significantly less delinquent while under supervision than those placed on regular probation, and they were no more delinquent than youths in the training school. Juveniles in the latter group committed as much delinquency while home on short furloughs or following escape as did the experimental group that was free in the community. An analysis of post-program recidivism revealed little difference between the experimental and control groups. However, when pre-program arrests were compared to post-program arrests, the results supported both the community program and institutionalization as an effective deterrent to offending. Arrests were reduced by 25% for those juveniles who were institutionalized and by 70% for those who remained free in the community. Although these findings seem to support a deterrent, rehabilitative, or combined effect from correctional programs, and in doing so contradict the conclusion that “nothing works” (Lipton et. al., 1975; Martinson, 1974), they also need to be treated with caution. The reductions in arrests following correctional intervention could have been due to natural maturation on the part of the juveniles or because their delinquent behavior peaked just before the sanction was imposed and then declined to a normal level (Empey & Stafford, 1991, p. 480). Other potential problems with the Provo Experiment include selection bias, the reliance on official arrest data to measure recidivism, and a

The Deterrent Effects of Formal Sanctions

47

relatively short follow-up period that could create an illusory deterrent effect (Maltz, 1984; Wooldredge, 1988). Unfortunately, these same criticisms can be applied to a later study that produced somewhat similar findings. Murray and Cox (1979) examined chronic delinquents in Chicago who were either incarcerated in a state reformatory or were diverted to one of several less-custodial community-based programs. Although all of the youths had lengthy records of delinquency, those who were chosen for the community programs were presumed to be somewhat less dangerous and perhaps more amenable to treatment. However, the study produced several striking outcomes, which would appear to support a deterrent effect from increasingly severe formal sanctions. A comparison of pre-program and post-program delinquent behavior revealed a larger reduction in offending among youths sent to reformatories relative to those who remained in the community. Additionally, for youths who remained in the community, the greatest reduction in arrests occurred among those who experienced the most restrictive forms of supervision. Murray and Cox (1979) suggested that these findings indicate evidence of a “suppression effect” from increasingly restrictive sanctions, and that court mandated intervention is the most effective deterrent to chronic juvenile offending. However, as with the Provo Experiment, their conclusions can be challenged on the grounds of subject maturation and the issue of whether the decline in official arrests represents a decrease in criminal behavior or an increase in ability to avoid apprehension. Subsequent researchers studying the effects of varying levels of formal punishment sought to methodologically improve upon earlier works. Klein (1986) randomly assigned 306 juvenile arrestees, all initially considered referable for further processing, to four different conditions: release, referral to a social service program, referral to a social service program with purchase of treatment, and formal petition for prosecution. By the end of a 6 month follow-up period, a trend had emerged that as the level of case processing increased, so did the chances of rearrest. This trend continued through a 27-month follow-up period, with released offenders rearrested less commonly than all others, and petitioned offenders rearrested most commonly. In the middle, there was a widening gap between the two social service conditions. The purchase of treatment condition, designed to ensure a

48

Excluding Violent Youths From Juvenile Court

greater level of service, actually yielded an increasingly greater number of arrests. In general, Klein’s (1986) findings on future rearrests indicated that diversion to a community agency was less harmful than petitioning to juvenile court, but more harmful than outright release. However, no similar differences emerged in terms of self-reported delinquent behavior. This could be due to the fact that only 185 of the original 306 subjects referred for random assignment were subsequently interviewed regarding their future offending. A higher interview rate may have confirmed the findings from the official data. On the other hand, the results concerning rearrests could be due to police selection bias. Klein (1986, p. 78) reported that there were attempts made by police officers to circumvent the randomization process. This allows for the possibility that youths perceived as less serious offenders received a lower level of formal processing. In addition, juveniles perceived as more serious offenders may have received a higher level of formal processing, and they may also have been targeted for rearrest. In a more detailed study of the same issue, Wooldredge (1988) examined the relative effectiveness of 12 different juvenile court dispositions on eliminating recidivism among 2038 youthful offenders in Illinois. Court actions ranged from case dismissals to incarceration, with a variety of combinations employed. Recidivism was measured based on future juvenile and adult arrests during a follow-up period of 3 to 7 years, and both logit and survival analyses were conducted. While controlling for a wide range of individual and environmental characteristics (but not eliminating the possibility of selection bias), a combination of probation supervision and community treatment was found to be associated with the least recidivism. Contrary to the findings of Klein (1986), 8 of 11 dispositional options yielded lower recidivism rates than the case dismissal option. Shorter terms of supervision and longer terms of community treatment were found to be effective in eliminating and prolonging recidivism, while longer terms of incarceration were found to be counterproductive. Based on these results, Wooldredge (1988) concluded: While it appears that “doing something” is better than “doing nothing” for eliminating recidivism, this study suggests that differences in “something” may also yield differences in recidivism rates. Specifically, two years of court supervision

The Deterrent Effects of Formal Sanctions

49

with community treatment is superior to any other sentence examined in this study for eliminating and prolonging recidivism. On the other hand, sentences involving detention should be carefully considered in relation to the types of delinquents they may be effective on. Failure to make such considerations could serve to inflate aggregate recidivism rates. While such sentencing may be inconsequential for specific deterrence in some cases, it can actually be damaging in others. (p. 293) Building on the work of Klein (1986) and Wooldredge (1988), Schneider and Ervin (1990) used in-person interviews and official records of 876 adjudicated delinquents from six cities to examine intentions of engaging in delinquent activity (measured at time 1) and subsequent offenses during a 2 to 3 year follow-up period (time 2). The sample over-represented serious and chronic juvenile offenders, or those “for whom general deterrence has already proven to be ineffective” and at whom “crime control policies are directed” (Schneider & Ervin, 1990, p. 591). Through pooling the data from the six juvenile courts, the authors found that the dispositional options of incarceration and regular probation produced significantly greater recidivism (i.e., more future contacts with the juvenile or adult court) than the option of restitution. Furthermore, youths who had been punished more severely, as indicated by the number of days incarcerated, reported being less certain of getting caught and committed more (rather than fewer) subsequent crimes. Although the results varied somewhat by city, and the lingering question of selection bias remained, Schneider and Ervin (1990) concluded that: The results from this study clearly suggest that punishmentoriented policies may set in motion unintended effects on selfimage or other values that culminated in more crimes being committed. Yet, many public officials and many social scientists are confident that increases in the severity or certainty of punishment will reduce crime. (p. 599) A study by Gottfredson and Barton (1993) would appear, at least at first glance, to contradict the findings of Klein (1986), Wooldredge

50

Excluding Violent Youths From Juvenile Court

(1988), and Schneider and Ervin (1990). The researchers investigated the effects of the 1988 closing of the Montrose Training School in Maryland. A group of 673 youths who had been incarcerated in the institution was compared to a group of 254 youths who had a high statistical probability of being institutionalized, but juveniles in the latter group did not go to Montrose because they were referred to the Department of Juvenile Services after admissions had ceased. These youths spent little or no time in any other institution and were instead placed in community-based programs. Recidivism was measured through both official records and face-to-face interviews, although only 750 interviews were attempted and the overall response rate was 62.5%. Gottfredson and Barton (1993) found that during a 2.5-year followup period, the incarcerated juveniles had significantly fewer arrests than the noninstitutionalized youths, but the difference was less substantial when only serious offenses were considered. Furthermore, the results from the self-reports of criminal behavior did not show the same degree of differential recidivism that was indicated by official records. This was possibly due to a number of factors: a longer follow-up period for interviews relative to the examination of official records; youths who were interviewed being more delinquent than those who were not, which could diminish group differences; or the inclusion of relatively minor crimes in the self-reports (Gottfredson & Barton, 1993, pp. 603604). Nevertheless, the results indicated that recidivism was greater among youths in the community-based programs than among those who were incarcerated, which would imply a greater deterrent, rehabilitative, or combined effect from the more restrictive sanction. However, this conclusion should be treated with caution. The study was conducted immediately after the closing of the institution, when the community programs were at the earliest stages of implementation. As noted by Gottfredson and Barton (1993), their findings suggest, “that the alternatives available when Montrose was closed were less effective in reducing crime than institutionalization would have been” (p. 604). It may be that if the less restrictive programs had been operating for a longer period of time, the quality of services provided would have been greater, and different results would have occurred. This argument is consistent with evaluations of the Massachusetts deinstitutionalization experience. An early evaluation found greater recidivism among youths committed to the Department of Youth Services following the closing of the state training schools, relative to

The Deterrent Effects of Formal Sanctions

51

juveniles released from institutions prior to the reform movement (Coates, Miller, & Ohlin, 1978). However, more positive results were found in regions with stronger community-based program implementation. In addition, a later study of the Massachusetts approach found recidivism rates that were as good or better than those of most other jurisdictions and were also lower than those of the predeinstitutionalization time period (Krisberg, Austin, & Steele, 1989). To continue with the theme of the effects of varying degrees of punishment, several studies have examined the juvenile court’s response to offenders referred for the first time. In a series of articles, Brown and his colleagues (Brown, Miller, & Jenkins, 1987; Brown, Miller, Jenkins, & Rhodes, 1989; Brown, Miller, Jenkins, & Rhodes, 1991) reported on a random sample of 500 juveniles who were adjudicated delinquent in Pennsylvania between 1960 and 1975. Their analyses indicated that while the type of disposition (i.e., probation vs. placement facility) was not related to future criminal behavior, adjudication at first court referral was associated with less recidivism than was postponed adjudication. Additionally, juveniles who were not adjudicated on their first referral had a rate of ending up in prison as an adult that was double that of youths who were adjudicated on their first referral. In a similar study, Jacobs (1990) reported slightly different findings. Among first-time juvenile court referrals, youths who were formally processed displayed less recidivism than those who were diverted. However, for those juveniles placed under formal court supervision, youths who received out-of-home placements exhibited greater recidivism than did those who received in-home supervision. Finally, in a recent study from the same mold, Minor, Hartmann, and Terry (1997) found that for first-time juvenile court referrals, there was a minimal relationship between type of court action and recidivism. The one exception was that for persons entering adulthood during the follow-up period, those who had been formally petitioned for further processing during the first court action were significantly more likely to be charged in adult court than those who were diverted. Based on these results, the authors concluded: Rather than encouraging high levels of recidivism, extensive and repeated use of diversion in this court was associated with reasonably constrained levels. These findings . . . imply that a

52

Excluding Violent Youths From Juvenile Court policy of “getting tough” with juvenile offenders by petitioning increasingly high proportions of cases, transferring more cases to adult court, and so forth may be misguided. (Minor et al., 1997, pp. 341-342)

The findings of Brown et al. (1987, 1989, 1991), Jacobs (1990), and Minor et al. (1997), while varying somewhat, appear consistent with those of Klein (1986), Wooldredge (1988), and Schneider and Ervin (1990). None of these studies found support for the harshest of formal sanctions. Although some of the findings indicate better results from outright release relative to more intermediate community-based programs (Klein, 1986; Minor et. al., 1997), the weight of the evidence appears to support formal supervision in a structured community setting. In any event, the results of these studies contrast with those of Murray and Cox (1979) and—to a lesser extent—Gottfredson and Barton (1993), who found support for institutionalization over other less restrictive programs. It is important to point out that the research by Murray and Cox (1979) and Gottfredson and Barton (1993) focused on more serious and frequent offenders. It is possible that the specific deterrent efficacy of legal punishments varies with the extent of past involvement in delinquent or criminal activity among those who are sanctioned (Smith & Gartin, 1989; see also DeJong, 1997). In other words, formal sanctions may have a different effect on novice offenders than on experienced offenders. Smith and Gartin (1989) found arrest to be associated with an increased probability of desistance among less experienced offenders, but not among those in later stages of involvement in criminal activity. Among more experienced offenders, however, arrest did reduce future rates of offending. Along these same lines, the literature appears to indicate that youths referred to juvenile court for the first time and offenders referred for relatively minor delinquent behavior exhibit less recidivism when they are not forced to undergo incarceration (Jacobs, 1990; Klein, 1986; Minor et. al., 1997; Wooldredge, 1988). On the other hand, for the most serious and frequent offenders, institutional confinement may be an effective response that reduces the likelihood of recidivism (Gottfredson & Barton, 1993; Murray & Cox, 1979). Unfortunately, the final answer is probably not as clear-cut as the previous paragraph would indicate. The findings of Murray and Cox (1979) and Gottfredson and Barton (1993) are directly contradicted by

The Deterrent Effects of Formal Sanctions

53

those of Schneider and Ervin (1990), who also employed a sample of serious and chronic youthful offenders. Furthermore, several longitudinal studies, conducted in very diverse locations (London, Philadelphia, and Racine, Wisconsin), suggest that more severe legal punishments, such as incarceration, lead to increases in future juvenile offending (Farrington, 1977; Farrington, Osborn, & West, 1978; Shannon, 1980; Wolfgang, Figlio, & Sellin, 1972). With these findings in mind, Smith and Gartin (1989) concluded: The important distinction is that [our] study compares the future criminal activity of those who are offending and punished (arrested) with that of those who are offending but not punished (released by police without arrest). In contrast, the typical study of specific deterrence compares the future criminal activity of two or more groups that differ, not in terms of whether they are punished or not, but in terms of how severely they are punished. It is possible that punishment will reduce future offending relative to no punishment at all, but that among those who are punished, more severe punishment may lead to increased future criminal activity. (p. 103) Overall, based on the research discussed above, it would appear that legal punishments could have a specific deterrent effect, but that harsher sanctions, such as incarceration, may have little additional impact and could actually lead to greater recidivism. This conclusion would seem to support the specific deterrent effects of certain punishment over more severe punishment, which contrasts with the basic philosophy of the “get tough” movement. However, the final answer to the question of whether or not formal sanctions produce a specific deterrent effect may be closer to the argument of Sherman (1993), who proposed that legal punishments can either reduce, increase, or have no effect on future criminal activity, depending on a variety of factors related to the offender, offense, and social setting. In presenting a theory of “defiance,” Sherman (1993) suggested that the future criminal behavior of formally sanctioned individuals will be influenced by such things as the degree to which offenders perceive the sanctioning as legitimate, the strength of social bonds that offenders have with the sanctioning agent and community, and the extent to which offenders accept the sanctioning without becoming angry and feeling rejected.

54

Excluding Violent Youths From Juvenile Court

Recent research has shown that people tend to obey the law more when they believe a criminal sanction has been administered fairly (Lanza-Kaduce & Radosevich, 1987; Makkai & Braithwaite, 1994; Paternoster, Brame, Bachman, & Sherman, 1997; Tyler, 1990). Furthermore, randomized experiments in arrest for domestic violence have found arrest to reduce repeat domestic violence among employed men, but increase it among those who are unemployed (Sherman, 1992; Sherman & Smith, 1992). More important for the current study, there is at least limited evidence that older people, who generally have more of a stake in conformity, are more effectively deterred by formal sanctions than are younger people (Sherman, 1993). If juveniles are less likely than adults to be deterred by formal sanctions, it may be due greatly to the legitimacy they perceive in legal punishments. It is well known that virtually all youths commit some type of delinquent act (Empey & Stafford, 1991), yet many or most are not caught or formally sanctioned. In discussing their findings of a positive effect of punishment on future delinquent behavior, Paternoster and Piquero (1995) asserted: In understanding the point that imposed sanctions may backfire, and lead to additional offending, it should be kept in mind that most of the offenders in this sample offended without consequence. It is entirely possible that those youths who committed delinquent acts and did get caught and sanctioned saw their punishment as unfair. When so many others “get away with it,” their own apprehension for offending may have been a bitter enough experience for the infrequently punished to become defiant. (pp. 270-271) In sum, research on the specific deterrent effects of formal sanctions has not established that punishing offenders (particularly youthful ones) will consistently reduce their future criminal behavior. It appears that a more certain justice system response provides more benefit than a more severe response, but even this conclusion is based on mixed evidence. The way that sanctioned offenders perceive the fairness of their punishment is an emerging question of great significance, and “the issue of disentangling deterrence from defiance also promises to be an important area of future research” (Paternoster & Piquero, 1995, p. 282).

The Deterrent Effects of Formal Sanctions

55

Summary During the past 20 years, there has been increasing public and political support for various “get tough on crime” policies. The most popular approaches to dealing with crime have been those that promise the harshest punishment. While often supported on the grounds of increased retribution and incapacitation, a central argument of the “get tough” movement is that stronger punishments are needed to deter criminal behavior. Treating juvenile offenders as adults fits in well with this view, as the criminal court is expected to provide formal sanctions that are greater and more effective than those imposed in juvenile court. In turn, this response to youthful offending is expected to have a beneficial impact on juvenile crime. Deterrence theory is over 200 years old, and its key argument of punishment based on certainty, severity, and swiftness continues to receive a great deal of attention today. Criminologists who have empirically tested the deterrence doctrine have focused almost exclusively on the certainty and severity of punishment, or the expected inverse relationship between the certainty and severity of punishment and crime. As this research has progressed through various stages, the findings have suggested that the perceived certainty of punishment has a greater deterrent effect than the perceived severity of punishment. With this in mind, it would seem logical that the experience of formal punishments would raise perceptions of punishment certainty, and thereby reduce future inclinations toward offending. Unfortunately, research has not consistently supported these expectations. Research on the effect of formal sanctions on perceptions of risk has produced mixed results, with the weight of the evidence suggesting that legal punishments are related to lower estimates of risk or that no significant relationship exists. Furthermore, the effect of formal sanctions on future criminal behavior is also open to debate. The literature does seem to indicate that legal punishments that are certain but moderate (e.g., arrest, adjudication, supervision with treatment) can be expected to produce greater specific deterrent effects than more severe punishment (e.g., incarceration). Also, harsh punishment in the form of longer incarceration appears counterproductive, possibly increasing the likelihood and frequency of future offending. Nevertheless, it may be that for the most serious and chronic offenders, specific deterrence will only take place with harsher punishment. To

56

Excluding Violent Youths From Juvenile Court

add to the complexity of this matter, the way that offenders are affected by punishment may be influenced by a number of factors, such as whether or not the sanctioning is perceived as being fair. For the current study, the next issues to address pertain to the nature of punishment received by juveniles in adult criminal court, and the effect these sanctions have on subsequent offending. Before going on to a consideration of the research that has addressed these matters, it should be noted that supporters of a policy of transferring youths to criminal court often criticize juvenile courts for being too lenient. Lenient treatment of juvenile offenders has been blamed for the dramatic increase in youth violence over the past decade, and many argue that criminal courts can correct this situation by providing greater accountability and stronger punishment. While this may sound good to the general public, research on the deterrence doctrine and the specific deterrent effects of formal sanctions is not so supportive. Nonetheless, if the harshest of punishment is needed to deter the most serious offenders, and if, in fact, the criminal court provides harsh punishment to those youths who come before it, transferring serious juvenile offenders to adult court may be a sound policy. The next chapter will address these issues.

CHAPTER 3

Punishment and Deterrence Through Treating Juvenile Offenders as Adults

As we have seen, transferring juveniles to adult criminal court is a policy that has evolved over time and constitutes a major part of the recent effort to crack down on violent youthful offending. Critics of the juvenile justice system contend that juvenile court sanctions are too lenient, and that adult criminal courts can provide increased accountability and greater deterrence. However, a review of deterrence theory and the associated research does not provide extensive evidence to support a policy of harsh punishment. If formal sanctions do act as a deterrent, it would appear to be due to their certainty, rather than their severity. On the other hand, there is some evidence that for the most serious offenders, harsh punishments may provide the only means for specific deterrence to take place. In Chapter 2, it was revealed that during the past 5 to 10 years, a shift occurred in the types of offenses juveniles are most likely to be charged with when they are waived to criminal court. Until recently, most youths were transferred on property crimes. Today, juveniles charged with personal or violent crimes make up the largest percentage of those transferred. This trend is likely to continue, as the most popular current method of transfer—legislative waiver--generally targets violent offenses and repeat violent offenders for exclusion from juvenile court. Based on the increasing public concern over youth violence, most people would agree that these are the youths who should be certified (Meddis, 1993; Schwartz et al., 1993).

57

58

Excluding Violent Youths From Juvenile Court

It does appear, then, that successful efforts are being made to place serious youthful offenders in adult criminal court. To a certain extent, deterrence research could be used to support this practice (assuming that these offenders are punished more severely than they would have been in juvenile court, and that this harsh treatment has the expected beneficial impact on future offending). This chapter will focus on research that has examined these issues. Studies that have addressed case outcomes and recidivism of transferred youths will be considered. Through this review, it will become clear that this research has been of uneven quality, and there is a need to add to the existing body of literature. Based on the key elements of deterrence theory, this chapter will be organized around the topics of the certainty, severity, and swiftness of punishment received by waived youths and the deterrent effects associated with this policy.

Punishment Certainty Researchers who have examined the certainty of punishment for juveniles transferred to adult court have focused on conviction rates. Not surprisingly, most studies find high conviction rates among transferred youths, generally in the range of 75% to 95% (Bishop & Frazier, 1991; Bishop et al., 1989; Bortner, 1986; Champion, 1989; Clarke, 1996; Eigen, 1981a, 1981b; Fagan, 1990; Gillespie & Norman, 1984; Hamparian et al., 1982; Houghtalin & Mays, 1991; Lemmon et al., 1991; Podkopacz & Feld, 1996; Rudman et al., 1986; Thomas & Bilchik, 1985; for exceptions, see Kinder et al., 1995; Sagatun et al., 1985; Singer, 1996). However, an important question is whether these conviction rates in criminal court are significantly different from those of comparable offenders in juvenile court. Research that has addressed this issue is limited, as most studies on this topic have been descriptive and did not include any comparison group of youthful offenders retained in juvenile court. Furthermore, as the following discussion reveals, studies that have employed comparison groups have produced equivocal results. Research by Eigen (1981a, 1981b) focused on homicide and robbery offenders in Philadelphia. Juveniles judicially waived to adult court in 1970 and 1973 were compared to youths retained in juvenile court, and transferred juveniles were also compared to a sample of adults originally charged in criminal court. Of the 75 youths waived on

Punishment and Deterrence

59

homicide charges, 89% were convicted in adult court, and 90% of those convicted were found guilty of first or second degree murder or voluntary manslaughter. Of the 79 youths retained in juvenile court, only 77% were adjudicated delinquent. Finally, 70% of the 200 adult homicide defendants were convicted in criminal court, and 89% of those convicted were found guilty of either first or second degree murder or voluntary homicide. Very similar results were obtained with regard to robbery offenders. The evidence from Eigen’s (1981a, 1981b) research would seem to indicate that violent youthful offenders are held to a higher degree of accountability in adult court than in juvenile court, and in criminal court they are even more likely to be convicted than similar adult defendants. However, there is good reason to doubt this conclusion. The problem of selection bias, discussed previously in Chapter 3, would appear to have a strong influence on the findings. Eigen also found that youths who allegedly killed a white victim and those with a higher degree of participation in a homicide offense were more likely to be waived. Also, the most serious and aggravated robberies were found in the certified sample. These factors were not controlled for in making the comparisons of conviction rates, meaning that the higher conviction rates for transferred youths could be a product of offense seriousness (or perceived offense seriousness) rather than court of jurisdiction. Research conducted as part of the effort to evaluate the Violent Juvenile Offender (VJO) Program of the Office of Juvenile Justice and Delinquency Prevention improved upon the work of Eigen (1981a, 1981b). In an early report, Rudman and colleagues (1986) examined 138 youths charged with violent offenses and considered for transfer in Boston, Newark, and Phoenix between 1981 and 1984. Of the 138 juveniles, 71 were retained in juvenile court and 67 were waived to criminal court. An important finding was that the rate of target crime conviction was slightly greater for youths retained in juvenile court (95.5%) than for those transferred to criminal court (92.2%). The key point here may be that all of the juveniles were actually considered for transfer, as the data suggested that “once a juvenile charged with a violent crime (with a prior violent or serious offense) is considered for transfer to adult court, he or she very likely will be convicted of a target crime irrespective of which court has jurisdiction” (Rudman et al., 1986, pp. 86-87).

60

Excluding Violent Youths From Juvenile Court

Later reports on the same program expanded the sample to include 201 violent youths considered for transfer in Boston, Phoenix, Newark, and Detroit during the same time period (Fagan, 1990; see also Fagan & Deschenes, 1990; Fagan et. al., 1987). Of these juveniles, 125 were retained in juvenile court and 76 were transferred to adult court. Again, the certainty of punishment was found to be comparable in both court systems, although conviction rates were now higher in criminal court (93%) than in juvenile court (86%). On the other hand, charge reduction was more common in criminal court, as the conviction rate for target offenses was slightly greater in juvenile court. Overall, “the results showed that sanctions are fairly certain for both waived and retained youth charged with violent felonies. There is little doubt that youths are held accountable for violent crimes, irrespective of the judicial forum in which the case is adjudicated” (Fagan, 1990, p. 114). While the research discussed to this point has focused on violent youthful offenders, a few other studies have considered the certainty of punishment for both violent and property offenders in the juvenile and adult systems. A study by Kinder and her colleagues (1995), of young offenders from St. Louis, compared 111 males certified as adults on a variety of charges in 1993 with the same number of randomly selected males who were adjudicated delinquent for felonies in juvenile court. A major limitation of this study was that 66% of the cases transferred to criminal court were not followed through to a determination of guilt. With this in mind, the authors reported that only 23% of the waived cases had resulted in conviction, while 74% of the cases retained in juvenile court resulted in an adjudication of delinquency. They also stated that of the transferred cases which were still pending, “many will most likely be dismissed” (Kinder et al., 1995, p. 40). Two other methodologically stronger studies have also considered both violent and property offenders. The first, by Fagan (1995), examined 15 and 16 year old robbery and burglary defendants from New York and New Jersey who were arrested during 1981 and 1982. Four hundred youths from two counties in New York, who were processed in criminal court, were compared to 400 youths from two matched counties in New Jersey, who were handled in juvenile court. Based on New York law (see also Singer, 1996), the cases from the New York sample originated in adult criminal court, while New Jersey law allowed for the comparable cases to be processed in juvenile court.

Punishment and Deterrence

61

Fagan (1995) found that for both samples, the rate of conviction was higher for burglary cases than for robbery cases. The burglary conviction rate in New Jersey’s juvenile courts (66%) was insignificantly greater than the burglary conviction rate in New York’s criminal courts (63%). In contrast, robbery cases in New Jersey’s juvenile courts were significantly less likely (46% conviction rate) to result in conviction than were robbery cases in New York’s criminal courts (56% conviction rate). These mixed findings led Fagan to conclude, “Accountability for adolescent offenders in criminal courts was no greater than for those in the juvenile courts” (p. 253). Finally, recent research by Podkopacz and Feld (1995, 1996) examined 330 cases processed in Hennepin County, Minnesota, from 1986-1992. In these cases, which involved both violent and property offenses, a prosecutor filed a reference motion for the juvenile court to consider transfer. This resulted in 215 cases being waived to criminal court, while the other 115 cases were retained in juvenile court. Although conviction rates were not emphasized in their study, the authors reported that 97% of the transferred cases resulted in conviction in criminal court, while only 86% of the retained cases resulted in an 25 adjudication of delinquency. As a whole, while recognizing the potential difficulty of selection bias, the evidence from studies that have compared conviction rates of transferred juveniles with those of youths retained in the juvenile system does not suggest that the certainty of punishment for similar offenders is much greater in adult criminal court than in juvenile court. A few studies indicate that conviction rates are higher in adult court, but others show little difference or even that conviction rates are greater in juvenile court. In any case, there is little or no evidence that juvenile courts dismiss a high rate of the serious cases that come before them. 25

According to Podkopacz and Feld (1996), "The courts dismissed charges against 14% of the youths retained in the juvenile system, but only 3% of those referred to adult criminal court. Clearly, prosecutors experienced less difficulty establishing the guilt of those youths referred to criminal court. This suggests either that juvenile courts implicitly may screen waiver cases for their prosecutive merits, or that prosecutors may pursue less vigorously the juvenile cases in which they anticipate less penal pay-off for their efforts" (p. 485, note 98).

62

Excluding Violent Youths From Juvenile Court

This is supported by the fact that juvenile court conviction rates revealed in the comparison studies just discussed also generally fall into the range of conviction rates presented in descriptive studies that have examined waived youths (Bishop & Frazier, 1991; Bishop et al., 1989; Bortner, 1996; Champion, 1989; Clarke, 1996; Gillespie & Norman, 1984; Hamparian et al., 1982; Houghtalin & Mays, 1991; Lemmon et al., 1991; Thomas & Bilchik, 1985).

Punishment Severity A major expectation in the move toward transferring more juvenile offenders to adult court is that these youths will receive harsher punishment than those retained in the juvenile system. This severe treatment can be justified on the grounds of retribution, incapacitation, deterrence, or all three. A question of great importance, then, is whether or not transferred youths do receive more severe punishment than they would have experienced in juvenile court. Contrary to what one might expect, the evidence indicates that harsher sanctioning in adult criminal court is not guaranteed. Some researchers have argued that a “leniency gap” exists for juveniles waived to criminal court, as these youths appear to receive less severe sentencing than they would have in juvenile court (Bortner, 1986; Champion, 1989; Emerson, 1981; Gillespie & Norman, 1984; Hamparian et al., 1982; Royscher & Edelman, 1981; Sagatun et al., 1985). These studies suggested that youths in criminal court are not seen as serious offenders, because they appear younger and less experienced than their adult counterparts. However, recent research has begun to clarify this situation. One important issue is the operationalization of sanction severity. Following the suggestions of Hagan and Bumiller (1983), various studies have examined two dimensions of sanction severity: sentence type and sentence length. Concerning the type of sentence imposed, studies show that youthful property offenders tend to be treated leniently in criminal court, often receiving sentences of probation in lieu of incarceration (Barnes & Franz, 1989; Bishop & Frazier, 1991; Bishop et al., 1989; Bortner, 1986; Champion, 1989; Hamparian et al., 1982; Podkopacz & Feld, 1996). On the other hand, juveniles convicted of violent offenses appear to be treated harshly in criminal court, where a jail or prison term is often imposed (Barnes & Franz, 1989; Clarke, 1996; Eigen,

Punishment and Deterrence

63

1981a, 1981b; Fagan, 1990, 1995; Houghtalin & Mays, 1991; Podkopacz & Feld, 1996; Rudman et al., 1986). However, as with conviction rates, it is still somewhat unclear if these incarceration rates in criminal court are very different from those of similar offenders in juvenile court, because few studies have employed comparison groups. After making some basic observations and comparative arguments, early researchers who examined the possibility of a leniency gap concluded that less severe sanctioning existed for waived juveniles in criminal court. In contrast, other research, using stronger comparison groups and examining specific types of offenders, has produced different findings. The previously discussed studies of Eigen (1981a, 1981b), Rudman et al. (1986), and Fagan (1990), which focused on violent youthful offenders, all found higher incarceration rates among juveniles transferred to criminal court. Eigen reported that 87% of the juveniles convicted of murder in adult court were incarcerated, as compared to 76% of the convicted adults and 49% of the youth adjudicated delinquent in juvenile court. Rudman and colleagues found that 90% of the waived juveniles who were convicted were also incarcerated, while only 77% of those who were adjudicated delinquent received a dispositional placement. Finally, Fagan reported that 89% of the juveniles convicted in criminal court were incarcerated, as compared to 84% of those adjudicated in juvenile court. A few other studies have compared sentence type for both violent and property offenders in juvenile and adult court. Barnes and Franz (1989) examined data on all 206 youth considered for transfer between 1978 and 1983 in a northern California metropolitan area. Almost half (47%) of the juveniles were transferred, and the rest remained in juvenile court. The analysis revealed that the effect of waiver varied with type of offense. Violent offenders were more likely to be incarcerated in adult court than were comparable offenders in juvenile court, while property offenders were less likely to be incarcerated in criminal court than were similar youths in juvenile court. More recent research has failed to find the same effect of type of offense. In Fagan’s (1995) study, it was revealed that 46% of the youths convicted of robbery in New York criminal courts were incarcerated, while only 18% of the offenders adjudicated delinquent on robbery charges in New Jersey juvenile courts were similarly treated. The findings for burglary defendants were nearly identical, as 46% of the youths convicted in criminal court and 24% of those adjudicated

64

Excluding Violent Youths From Juvenile Court

delinquent in juvenile court were incarcerated. Podkopacz and Feld (1996) reported similar findings, although the percentage of youths incarcerated was much higher in both the juvenile and adult criminal 26 courts that were studied. Bishop and her colleagues (1996) conducted another recent study that considered sentence type for youths in juvenile and adult court. The authors compared 2,738 certified offenders in Florida, who were transferred to criminal court on a variety of charges in 1987, with the same number of individually matched delinquents who were retained in the juvenile system. Although sentence type was not a major focus of the study (recidivism was the main outcome variable examined), the authors did report that transferred juveniles were more likely to be incarcerated than youths who remained in the juvenile system. In sum, research that has examined the type of sentence imposed on juveniles certified as adults has produced mixed results. Some studies have found evidence that juvenile offenders receive more lenient treatment in criminal courts, often receiving sentences of probation. Recent research indicates a change in this pattern, especially for violent offenders. It would appear that the increasing concern over youth violence during the past decade has not only resulted in a larger percentage of violent juveniles being transferred to adult court, but also in a larger percentage of those youths being incarcerated. However, the lingering problem of selection bias again needs to be noted. Many of the above studies employed weak or no controls for offense seriousness and prior record, meaning that the transferred youths may have been more likely to be incarcerated simply due to being more serious offenders. Still, as stated by Podkopacz and Feld (1996), “although the 26

"The adult criminal courts incarcerated 93% of the youths convicted of a presumptive [violent] offense, while the juvenile court imposed long-term confinement on 65% of the youths retained in juvenile court who were adjudicated for a presumptive offense. The adult criminal courts confined 78% of youths convicted of non-presumptive [property] offenses in jail or prison, whereas the juvenile court committed only 61% of non-presumptive delinquents to a correctional facility. Thus, for both presumptive and nonpresumptive offenses, criminal courts incarcerated youths convicted as adults significantly more often than did the juvenile court" (Podkopacz & Feld, 1996, p. 487).

Punishment and Deterrence

65

waiver process selects youths on the basis of seriousness, the differences in rates of dismissal, conviction, and incarceration between the two systems are striking” (p. 487). Research on the second dimension of sanction severity, length of incarceration, has produced similar results. Various studies have shown that for those transferred youths who are incarcerated, lengthy sentences are common (Bishop & Frazier, 1991; Bishop et al., 1989; Lemmon et al., 1991; Singer, 1996; Thomas & Bilchik, 1985). These studies found average jail and prison sentences ranging from 1 to 4 years or more, with the longest sentences imposed on violent offenders. However, the next question would be whether or not the periods of incarceration issued in criminal court are longer than those prescribed for similar offenders in juvenile court. Numerous studies have found evidence that lengthier sentences are imposed in adult criminal court. Eigen’s (1981a, 1981b) research revealed that juveniles convicted in adult court on either homicide or robbery charges received longer periods of incarceration than did convicted adult defendants and youths adjudicated in juvenile court. In the study by Rudman et al. (1986), youths incarcerated by criminal courts received sentences five times longer than those imposed on offenders retained in the juvenile system. Similarly, Fagan’s (1990) analysis revealed that sentences of incarceration for juveniles in criminal court were four times longer than for youths adjudicated in juvenile court. Finally, Bishop and her colleagues (1996) found a mean sentence length of 245 days for waived juveniles who were incarcerated, compared to only 90 days for youths retained and placed by the juvenile court. Jurisdictional age limits of juvenile courts vary among states, but the maximum age for supervision and confinement is generally between 18 and 21 (DeFrances & Strom, 1997; Fritsch et al., 1996; Krisberg & Austin, 1993). The results discussed above suggest that because criminal courts are not limited by the jurisdictional age restrictions present in the juvenile system, youths incarcerated by adult courts experience substantially longer sentences than those imposed by juvenile courts. However, this conclusion must again be tempered by the recognition of possible selection bias, as well as the inconsistent findings of several other studies. For example, Podkopacz and Feld (1996) reported that youths convicted as adults on violent offenses experienced longer sentences of

66

Excluding Violent Youths From Juvenile Court

incarceration than those imposed on juveniles adjudicated delinquent on similar charges. On the other hand, youths convicted in adult court on property offenses received shorter sentences of incarceration than comparable offenders in the juvenile system. Even more interesting, Fagan’s (1995) study of robbery and burglary offenders found no difference in sentence lengths for youths charged with either offense and incarcerated by either the criminal or juvenile court. Irrespective of type of offense and court of jurisdiction, average minimum sentences of incarceration were approximately 11 months, and average maximum sentences were 31 to 34 months. It should be noted that the amount of time an offender actually serves in confinement might be considerably different than the length of sentence imposed by a criminal court. A recent study on this issue examined the imprisonment of waived youths in Texas (Fritsch et al., 1996). The researchers examined 946 youths transferred to adult court from 1981 to 1993 and sentenced to prison. The key finding was that these juveniles consistently received longer sentences than were available in juvenile court, but they served an average of only 27% of their original sentence. When actual time served was taken into account, the waived youths rarely served longer sentences than those that were available in juvenile court. Overall, it appears that transferred juveniles, particularly those waived on violent offenses, receive longer sentences of incarceration than those issued to comparable youths in juvenile court. However, as indicated by Fagan’s (1995) study, this increase in sentence length is by no means assured. Furthermore, even when longer sentences are imposed, the actual time served by youths in the adult system may be the same or even shorter than that of similar offenders in the juvenile system.

Punishment Celerity As compared to the certainty and severity of punishment for juveniles transferred to adult court, the swiftness of their punishment has been much less studied. Recently, questions have been raised regarding the speed at which cases are processed within the juvenile justice system (Butts, 1996, 1997; Butts & Halemba, 1994). While the th 6 Amendment to the Constitution guarantees a “speedy and public trial” for anyone charged with a criminal offense, the definition of

Punishment and Deterrence

67

“speedy” is open to debate. Furthermore, despite the “due process revolution” in juvenile justice that extended various procedural rights to youthful offenders, the United States Supreme Court never addressed the issue of speedy trial rights for juveniles. In general, case-processing time is becoming an emerging point of emphasis in juvenile justice, as “immediate interventions” are being stressed as a response to delinquent behavior (Howell, 1997; Krisberg et al., 1995). Therefore, it would seem important to know how rapidly comparable cases are processed in the juvenile and adult systems. Recent research by Butts (1997) focused on the handling of delinquency cases during 1991 and 1992 by juvenile courts in 16 states. The median time from referral to disposition was 40 days, while 26% of all cases had disposition times exceeding 90 days. However, formally petitioned cases had substantially longer disposition times than cases handled informally, as almost 40% of all petitioned cases required more than 90 days to reach disposition. In larger jurisdictions, nearly half of all petitioned cases had disposition times in excess of 90 days, and 22% required more than 120 days. Finally, nearly 40% of all cases that resulted in an out-of-home placement or formal probation supervision required more than 90 days to reach disposition. The 90-day time period is important because it is the maximum disposition time recommended by various national standards (Butts, 1997, pp. 6-7). Based on this research, it seems that the most serious cases take the longest time to process in juvenile court, often requiring 3 to 4 months or more to reach disposition. In the context of the current study, the next question would concern the case processing time of comparable offenders in the juvenile and adult systems. Although disposition times for serious offenders in juvenile court may not meet national standards, juveniles waived to adult court may be subjected to even longer periods of case processing. A review of the literature finds that only a few studies have addressed this issue. Lemmon and his colleagues (1991) examined all youths judicially waived to criminal court in Pennsylvania in 1986. For these 221 cases, the mean time from transfer to sentencing for convicted offenders was 8.5 months, while the median time was 7 months. Although no comparison group was employed, the evidence indicates caseprocessing times in excess of those discussed by Butts (1997) with regard to serious offenders in the juvenile system.

68

Excluding Violent Youths From Juvenile Court

Three comparison studies have also found evidence of swifter punishment in juvenile court. Kinder et al. (1995) reported that at the time their research ended, all of the cases that had been retained in juvenile court had completed disposition, while 66% of the transferred cases were still pending. In a methodologically stronger study, Rudman and colleagues (1986) found that it took 2.5 times as long (246 versus 98 days) for a violent youth to be transferred, convicted, and sentenced than for a similar youth to be retained, adjudicated, and disposed in juvenile court. Finally, Fagan’s (1995) recent study also reported swifter action in juvenile court, as cases took 100 days on average to be disposed in New Jersey juvenile courts and 145 days to reach 27 sentencing in New York’s criminal courts. In sum, while the case processing times of serious cases in juvenile court may be questioned and criticized, it seems that comparable cases in adult criminal court take even longer to reach completion. However, research in this area is limited, as only two studies have made strong comparisons between the two systems. If there is any benefit to be gained from a more rapid response to offending, then more knowledge needs to be generated concerning differences in case processing times between the juvenile and adult systems.

Deterrent Effects Along with retribution and incapacitation, the principle of deterrence has been used as a primary justification for transferring increased numbers of youthful offenders to adult court (Bishop et al., 1996; Fagan, 1995; Singer, 1996). It is assumed by policymakers and the general public that treating juveniles as adults will reduce overall juvenile crime (i.e., a general deterrent effect) and also reduce or eliminate the future offending of those transferred to criminal court (i.e., a specific deterrent effect). This expectation of greater deterrence is based on the perception that juvenile courts are more lenient, and that criminal courts can provide increased accountability and stronger punishment, which will reduce youthful offending. 27

An obvious question concerning Fagan's (1995) study is whether any observed differences (such as the difference in case processing time) are due to a "jurisdictional effect" rather than a "transfer effect." This point is further discussed at the end of this chapter.

Punishment and Deterrence

69

Based on research concerning the certainty, severity, and swiftness of punishment for juveniles transferred to adult court, it would appear that if criminal courts have an advantage, it is in terms of punishment severity. Although there is some debate in the literature, adult criminal courts do appear to provide harsher punishment, particularly for violent offenders. In contrast, the studies on punishment certainty do not reveal consistently higher conviction rates in adult court than in juvenile court. Additionally, based on very limited evidence, juvenile courts seem to 28 provide speedier processing of serious cases. The research discussed in Chapter 3 provided limited support for the deterrent effects of increasingly severe punishment. While punishment certainty appeared more important than punishment severity, there was evidence that for the most serious offenders, harsh punishment may be the only way to generate deterrence. Because violent youths are currently being targeted for transfer to criminal court, and adult courts do appear to be imposing harsh sentences on these offenders, it would seem logical to expect a beneficial impact on juvenile crime and on the future criminal behavior of those juveniles who are transferred. Unfortunately, the minimal research that has addressed these issues has not found the expected deterrent effects. In terms of general deterrence, only two studies have looked at the impact of expanded juvenile transfer laws on aggregate adolescent crime rates. Singer and McDowall (1988; see also Singer 1996) evaluated the effect of New York’s legislative waiver law that became effective in 1978. The law lowered the eligible age of criminal justice handling of juveniles to 13 for murder and 14 for other violent offenses. Using Philadelphia as a control series, the authors conducted a time series analysis of monthly juvenile violent crime arrest rates during the period of 1974 to 1984. The results showed that the law had little

28

As noted in Chapter 3, there is ambiguity surrounding the potential effect of swift punishment. Speedier case processing in juvenile court may produce a beneficial impact on future offending. Consider, though, that delayed punishments could be felt as more painful, due to the time spent thinking about the impending punishment. If more severe punishments provide greater deterrence, lengthier case processing in adult criminal court may also be advantageous.

70

Excluding Violent Youths From Juvenile Court

measurable effect on juvenile crime and thus did not produce the deterrent effect expected by policymakers. Similar findings were obtained by Jensen and Metsger (1994), who also used a time series design to assess Idaho’s 1981 legislative waiver law that required violent juvenile offenders aged 14 to 18 to be sent to adult court. Using Montana and Wyoming as comparison states, where similar transfer procedures as those originally existing in Idaho were being employed, yearly juvenile violent crime arrest rates were examined during the periods of 1976 to 1980 and 1982 to 1986. The analysis revealed that violent juvenile crime increased significantly in Idaho following the enactment of the law, but decreased in the comparison states, suggesting that the new law had no deterrent effect. Most important for the current research, only a few studies have assessed the specific deterrent effect of transferring juveniles to adult court. Fagan’s (1995) research on New York and New Jersey youthful offenders revealed that robbery defendants processed in New York criminal courts were rearrested and reincarcerated at a higher rate than comparable youths in New Jersey juvenile courts, and the New York offenders were also rearrested in a shorter amount of time. In contrast, there were no significant differences for burglary offenders in terms of rearrest, reincarceration, and time until rearrest in juvenile versus adult criminal court. In their study of matched youths in Florida, Bishop and her colleagues (1996) found that 30% of the juveniles transferred to adult court were rearrested during a 1-year follow-up period, while only 19% of the youths retained in juvenile court were rearrested. Furthermore, the mean time to failure was shorter for waived juveniles (135 days) than for retained youths (227 days). Finally, there was evidence that recidivism was more serious among youths treated as adults, as 93% of the transferred juveniles who were rearrested were charged with a felony, while 85% of the retained youths who were rearrested were charged with a felony. In a follow-up to the study by Bishop et al. (1996), the same authors (Winner et al., 1997) sought to determine whether the shortterm differences in recidivism between transferred and retained youths persisted over time for all types of offenders. Approximately six additional years of tracking was conducted to see how the transfer decision affected recidivism over the long term. The analysis showed that youths retained in juvenile court eventually caught up with those

Punishment and Deterrence

71

transferred to adult court in terms of the prevalence of rearrest, but this was due completely to an increased probability of rearrest over the long 29 term for retained youths processed on felony property offenses. Once the effect of offense type was controlled, the multivariate analysis indicated that transfer led to more recidivism. Finally, even though transferred property felons were less likely to reoffend, when they did so they reoffended more frequently and more quickly than did the comparable retained youths. Lastly, in their study of youths considered for waiver in Hennepin County, Minnesota, Podkopacz and Feld (1996) also failed to find evidence of greater specific deterrence in adult criminal court. During a 2-year follow-up period, 58% of the transferred juveniles were convicted of a new crime, in contrast to only 42% of the retained youths. The authors offered three possible explanations for the lower juvenile court recidivism rate: (a) through an emphasis on prior offending, the juvenile court succeeded in transferring the most serious and frequent offenders who had a greater probability of recidivism; (b) treatment services were more effective in the juvenile correctional system; or (c) the adult criminal system better trained (rather than deterred) further criminality than did the juvenile system. The second and third explanations presented above by Podkopacz and Feld (1996) both suggest greater effectiveness in the juvenile court system with regard to handling serious youthful offenders. The first explanation points to the problem of selection bias in this type of 29

Fagan (1995) produced somewhat similar findings, revealing that transfer affected burglars differently than it did robbers. Together, these results seem to indicate an offense-specific transfer effect. Juveniles transferred on violent offenses appear to exhibit greater recidivism than that of similar youths retained in the juvenile system, while the recidivism of juveniles transferred on property offenses seems comparable to that of similar youths retained in juvenile court. However, as stated by Winner et al. (1997), "We know of no theory that would predict offense-specific effects for transfer. . . . Prior research does not shed light on the issue either. . . . Our findings of a higher prevalence of recidivism for transfers except for those transferred for property offenses, of a quicker time to rearrest for transfers generally, and of a higher frequency of rearrest among those transfers who reoffend suggest avenues for future research" (pp. 560-561).

72

Excluding Violent Youths From Juvenile Court

research, or that greater recidivism could be found among transferred juveniles simply because they are more serious and persistent offenders. However, it is extremely important to consider the fact that none of these explanations would favor a dramatic increase in the number of juveniles who are sent to adult criminal court. Obviously, if the juvenile court provides better treatment, or if the adult system trains (rather than deters) further criminality, it would be prudent to keep all but the most serious and violent youthful offenders in the juvenile system. Furthermore, if transferred juveniles exhibit greater recidivism simply because they are substantially different from youths retained in the juvenile system, then it would not seem wise or just to treat broad categories of youthful offenders in a similar manner and send them in mass to adult court. Yet this is what the currently most popular method of transfer, legislative waiver, seeks to do.

Punishment and Deterrence: Is Transfer Effective? As revealed in the research discussed in this chapter, transfer is an extreme response to juvenile offending, with potentially severe consequences. While the evidence is far from unambiguous, youths who are treated as adults may be subjected to a drawn-out adjudicatory process (possibly involving an extended stay in jail), followed by a criminal conviction and lengthy prison sentence. With regard to the conditions of confinement, there is also evidence that juveniles in adult jails and prisons receive less adequate treatment services and are far more likely to be violently victimized than youths in juvenile correctional facilities (Forst, Fagan, & Vivona, 1989; Howell, 1997; Reddington & Sapp, 1997). Subjecting juveniles to this potentially harsh punishment is supported by the expectation of enhanced public safety. With the current popularity of transferring juveniles to adult criminal court, it is surprising how limited the research is on the effectiveness of this policy. Most studies have provided only descriptive statistics on offense and offender characteristics of waived youths and their case outcomes in adult court. Furthermore, studies that have employed comparison groups present various methodological weaknesses, particularly in controlling for offense seriousness and prior record. Finally, available research has focused on judicial waiver, even though 90% of all juveniles in the criminal system are transferred by

Punishment and Deterrence

73

other means (Howell, 1997, p. 108). The lack of knowledge about legislative waiver is especially problematic, as this method is currently most popular and seeks to exclude large categories of youthful offenders from juvenile court jurisdiction. In short, past research has produced equivocal results and left many questions unanswered. Nevertheless, studies conducted to this point do suggest several generalizations. Overall, studies have not definitively established more certain punishments in adult court. However, several quality studies using recent data have found that violent youths experience an extremely strong likelihood of conviction in adult court, and this probability of conviction is at least as great (if not greater) than that of comparable offenders in juvenile court. Therefore, there is little doubt that violent youths are held accountable for their actions in adult court, and the basic goal of a reliable response to youth violence can be achieved through juvenile transfer. On the other hand, based on limited evidence, the juvenile court appears to operate at a faster pace for serious and violent offenders. Speedier case processing in the juvenile system may lead to a variety of beneficial consequences, such as immediate placement in appropriate treatment facilities and programs, an individual perception of fair procedures, and reduced contact with adult criminals. Furthermore, rapid case processing may decrease or eliminate the time a youth spends “on the street” awaiting adjudication and disposition of his or her case. Although ignored by past researchers, it is entirely possible that transferred youths are often released on bail prior to trial or sentencing in criminal court, and as a result, they are free to commit more crime in the community. Even if no new crime is committed, release on bail may lead to a failure to appear for future court proceedings or prohibit the administration of treatment and supervision services. Finally, if the criminal court has a distinct advantage, it is in providing more severe punishments through incarceration. Particularly for violent offenders, recent research has established that adult courts confine youths at a higher rate and for longer periods of time than do juvenile courts. Although selection bias is a major issue, these findings support an argument of greater retribution and incapacitation in adult criminal court. At question, however, is whether this increase in punishment severity is worthwhile. A major justification for providing this harsher treatment is to deter future juvenile crime, but the few

74

Excluding Violent Youths From Juvenile Court

studies that have addressed this issue have failed to find the expected general or specific deterrent effects. Moreover, as Bishop and her colleagues (1996) suggested in their Florida study, short-term gains achieved by incapacitating transferred youths can be offset by greater recidivism once they are released from confinement. Overall, if the goal of greater deterrence is not being met, even longer periods of incarceration may be required to protect the public, resulting in substantial monetary costs.

The Present Study It is noteworthy that only one comparison study covered the full spectrum of the certainty, severity, and swiftness of punishment and subsequent criminal behavior of offenders in the juvenile and adult justice systems. Fagan’s (1995) research on New York and New Jersey youths charged with robbery and burglary addressed these issues, but the study poses several limitations. First, the samples consisted of offenders processed during 1981 and 1982. Aside from the fact that these cases are now over 15 years old, the data (including much of the information on recidivism) predated the surge in youth violence that began during the latter half of the 1980s. Therefore, there is a need to analyze more recent data, particularly with regard to violent youths. Second, the study seemingly employed a weak control for prior offending, as “juvenile court records were not available for the criminal court cases” (Fagan, 1995, p. 247). This allows for the possibility that youths in adult court were just more serious and frequent offenders, which could explain their greater recidivism. Finally, the samples were of youths from two separate jurisdictions with completely different juvenile offender laws. Although Fagan argued that this design minimized selection bias, it is hard to ignore probable differences in justice system operations and contexts of offending between the two jurisdictions. For example, police and correctional practices with youths could differ greatly across states, and the 1980s drug crisis in New York City (Johnson, Williams, Dei, & Sanabria, 1990) could further complicate comparisons with other jurisdictions. In sum, there is a need to expand and improve upon the existing body of literature regarding the policy of transferring juveniles to adult criminal court. Given the recent concern over youth violence and the emphasis on excluding violent young offenders from juvenile court

Punishment and Deterrence

75

jurisdiction, a study is needed that examines the overall effectiveness of treating violent juveniles as adults. Similar to the research by Fagan (1995), the current study will address the issues of punishment certainty, severity, and swiftness as well as the future offending of comparable youths in the juvenile and adult justice systems. However, data will be used which postdates the rapid rise in violent juvenile crime that occurred between the mid-1980s and mid-1990s. Also, the study will focus on a single state (Pennsylvania) that recently adopted a policy of legislatively excluding certain violent youths from juvenile court. Finally, in addition to a quantitative analysis, findings from interviews with various juvenile and criminal justice system officials will be discussed, in order to assess how effective the new legislation has been in dealing with violent juvenile offenders.

This page intentionally left blank

CHAPTER 4

Research Agenda

The purpose of this study is to examine the consequences of transferring violent juvenile offenders to adult criminal court. To assess these consequences, we will focus on two impact points: (1) The punishment imposed on comparable violent youths in the juvenile and adult systems, and (2) The subsequent criminal behavior of these offenders. More specifically, a quasi-experimental design will be employed to investigate whether waived juveniles are treated more harshly than comparable youths in the juvenile system, and whether transfer to adult court generates a greater specific deterrent effect than is produced through juvenile court processing. Using deterrence theory as a guide, the current study will focus on the certainty, severity, and swiftness of prescribed sanctions and the recidivism of violent youthful offenders who are handled in the juvenile and adult justice systems.

Hypotheses

30

Efforts to increase the number of juveniles transferred to adult court have been fueled by the perception that juvenile courts are overly lenient, and that adult criminal courts can and will provide increased accountability through harsher punishment (Bishop et al., 1996; Fagan, 30

For each hypothesis test, the null hypothesis (Ho) will be that there is no significant difference in the dependent variable for violent youths in juvenile versus adult court. This section presents alternative hypotheses (Ha) suggested by prior research. When warranted, the alternative hypothesis will indicate a one-tailed test. Otherwise, a two-tailed test will be suggested. P-values will be used to estimate the effect of transfer on the dependent variables.

77

78

Excluding Violent Youths From Juvenile Court

1995; Greenwood, 1995). In general, it is expected that youths in the criminal system will experience a greater chance of conviction, a higher probability of incarceration, longer terms of confinement, and a rapid response to their offending behavior. In turn, this harsh treatment is expected to reduce future inclinations toward offending. However, the research reviewed in Chapter 4 raises questions about these expectations. Before we examine the hypotheses suggested by the review of the literature, it is important to consider a potential area of study that has been ignored by previous researchers. Past studies have failed to consider the detainment of youthful offenders prior to disposition, or whether transferred juveniles are more or less likely than similar youths retained in juvenile court to be released from custody prior to a 31 determination of guilt and subsequent sentencing. This period of detention is not only relevant to the nature of punishment imposed on juvenile offenders (it would appear to be an aspect of punishment certainty, severity, swiftness, or all three), but also it seems pertinent to public safety. Overall, a finding that waived youths are more likely to remain in custody prior to their trial and sentencing would seem, from an accountability and public safety point of view, to provide support for treating serious and violent juvenile offenders as adults. Although the United States Supreme Court has extended various due process rights to juveniles, the right to bail has not been granted (see Schall v. Martin, 1984). This suggests that the juvenile court might be better able to detain offenders from the time of arrest through disposition. On the other hand, supporters of a policy of treating juveniles as adults might argue that criminal courts can set bail at a sufficiently high amount to keep transferred youths in custody until trial and sentencing. Consequently, this study will seek to determine if:

31

Offenders may be released from custody either prior to trial (determination of guilt in juvenile court) or, if they are detained and convicted, prior to sentencing (disposition in juvenile court). These are two distinct time periods, but for this study they will be combined into one pre-dispositional period. For offenders who are acquitted or have their case dismissed, the pre-dispositional period will end on the date of acquittal or dismissal. For offenders who are convicted, the pre-dispositional period will continue until the date of sentence.

Research Agenda

79

Ha (1): There is a significant difference in the likelihood of release from pre-dispositional secure custody for violent youths in juvenile versus adult court. As discussed in Chapter 4, researchers who have examined the certainty of punishment for juveniles waived to adult criminal court have focused on conviction rates. While these rates have been generally found to be high, it is questionable if conviction is more likely in adult court than in juvenile court for comparable offenders. Some studies have found evidence of higher conviction rates in adult criminal courts (Eigen, 1981a, 1981b; Fagan, 1995; Podkopacz & Feld, 1996), while others have found juvenile court conviction rates to be higher (Kinder et al., 1995), or that there is little difference between the two systems (Fagan, 1990; Rudman et al., 1986). Furthermore, conviction rates may be based on conviction for any offense or conviction on a “target offense” (i.e., the most serious offense for which the youth was arrested; see Fagan, 1990; Rudman et al., 1986). It is possible that one court system could produce a greater likelihood of conviction, but also commonly convict on reduced charges as part of a plea bargaining process. Accordingly, we will investigate whether: Ha (2): There is a significant difference in the likelihood of conviction on any offense for violent youths in juvenile versus adult court. And if, for those youths who are convicted: Ha (3): There is a significant difference in the likelihood of 32 conviction on a target offense for violent youths in juvenile versus adult court. A popular argument for treating increased numbers of juvenile offenders as adults is that criminal courts will provide more severe sanctioning than that imposed in juvenile courts. Although some early research challenged this assertion (see, e.g., Bortner, 1986; Emerson, 1981; Hamparian et al., 1982; Royscher & Edelman, 1981), more recent studies tend to provide supportive evidence, particularly for violent offenders. Concerning sentence type, various comparison studies have found higher incarceration rates for violent youths 32

In this study, a target offense will be robbery or aggravated assault.

80

Excluding Violent Youths From Juvenile Court

transferred and convicted in criminal court (Barnes & Franz, 1989; Eigen, 1981a, 1981b; Fagan, 1990, 1995; Rudman et al., 1986; Podkopacz & Feld, 1996). Therefore, we will test the following: Ha (4): Violent youths convicted in adult court are significantly more likely to be incarcerated than are comparable offenders adjudicated delinquent in juvenile court. Similarly, research on the second dimension of sanction severity, incarceration length, suggests that more severe treatment is experienced by violent youths remanded to adult criminal court. Although there is some debate in the literature (see Fagan, 1995; Fritsch et al., 1996), the weight of the evidence indicates that for comparable violent offenders in the juvenile and adult systems, lengthier sentences are imposed in criminal courts (Bishop et al., 1996; Eigen, 1981a, 1981b; Fagan, 1990; Podkopacz & Feld, 1996; Rudman et al., 1986). This suggests the following: Ha (5): Violent youths convicted and incarcerated by adult courts receive significantly longer sentences than do comparable offenders adjudicated delinquent and incarcerated by juvenile courts. While the speed at which cases are processed within the juvenile justice system is a current concern (Butts, 1996, 1997; Butts & Halemba, 1994), as it appears that dispositional times for serious cases often fail to meet national standards, it seems likely that juveniles waived to adult court are subjected to even longer periods of case processing. Although the research is limited, the few comparison studies that have been conducted have consistently found evidence of swifter case processing in juvenile courts (Fagan, 1995; Kinder et al., 1995; Rudman et al., 1986). Consequently, we will test the following: Ha (6): Violent youths in adult court experience longer case processing time than do comparable offenders in juvenile court. The first six hypotheses address the nature of punishment received by similar violent youths in the juvenile and adult justice systems. A second and related issue is the future criminal behavior of transferred youths. If, in fact, waived juveniles are treated more harshly in adult criminal court, this punishment is expected to provide a specific

Research Agenda

81

deterrent effect that reduces or eliminates future inclinations toward offending. To a certain extent, deterrence research could be used to support this expectation (see Gottfredson & Barton, 1993; Murray & Cox, 1979). Accordingly, directional hypotheses could be formed around the idea that juveniles treated as adults will exhibit lesser recidivism than will comparable youths retained in the juvenile system. Unfortunately, the few studies that have examined the specific deterrent effect of transferring juveniles to adult court do not provide evidence to support directional hypotheses of this nature. Instead, the evidence suggests greater, more serious, and more rapid recidivism among youths waived to criminal court (Bishop et al., 1996; Fagan, 1995; Podkopacz & Feld, 1996; Winner et al., 1997). Furthermore, predictions taken from labeling theory stand in stark contrast to those derived from deterrence theory. Based on this empirical and theoretical knowledge, an argument could be made for directional hypotheses that assert increased recidivism among transferred youths. Adding to this dilemma is the potential difficulty of selection bias. As suggested by Podkopacz and Feld (1996, p. 491), greater recidivism among waived youths could be due solely to the juvenile court’s ability to transfer the most serious and frequent offenders who have a greater probability of future criminal behavior. This type of problem has plagued past research on labeling theory, most notably in studies that have examined the prospect of a deviance amplification effect from formal juvenile court processing (Smith & Paternoster, 1990). Because of the possibility that a positive association between being referred to juvenile court and future offending could arise due to a selection artifact, Smith and Paternoster suggested the following: In examining whether formal processing by the juvenile justice system increases future offending, the most appropriate null hypothesis is that juvenile justice processing has no causal effect on future offending. This does not mean that we believe this hypothesis is true. It does mean that the burden of proof rests with those who claim that a causal effect exists. Such claims are strengthened to the degree that rival explanations, such as selection bias, can be ruled out. (pp. 1128-1129) Therefore, after making a concerted effort to control for offense and offender characteristics, demographic variables, and other social

82

Excluding Violent Youths From Juvenile Court

factors, we will assume that there is no difference in the specific deterrent effect of juvenile versus adult court sanctions. We will then examine the following alternative hypotheses: Ha (7): For violent youths released from secure custody prior to final disposition, there is a significant difference in the likelihood of pre-dispositional recidivism for those in juvenile versus adult court. Ha (8): For violent youths released from secure custody prior to final disposition, there is a significant difference in the seriousness of pre-dispositional recidivism for those in juvenile versus adult court. Ha (9): For violent youths released from confinement following final disposition, there is a significant difference in the likelihood of post-dispositional recidivism for those in juvenile versus adult court. Ha (10): For violent youths released from confinement following final disposition, there is a significant difference in the seriousness of post-dispositional recidivism for those in juvenile versus adult court. Ha (11): For violent youths released from confinement following final disposition, there is a significant difference in the time to post-dispositional recidivism for those in juvenile 33 versus adult court. In sum, this study attempts to provide a better understanding of the effects of treating violent juvenile offenders as adults. The hypotheses listed above cover a broad range of key issues pertaining to this policy. In addition to empirically testing these hypotheses through quantitative analyses, qualitative research will also be conducted that seeks to expand on what is known about the statutory exclusion of violent youths from juvenile court. Information obtained through interviews 33

Analyses of recidivism require consideration of time at risk, to account for differences in the time that offenders have an opportunity to commit further crime. Due to a lack of reliable dates of release on bail for transferred youths, time to recidivism could only be analyzed during the post-dispositional period.

Research Agenda

83

with various juvenile and criminal justice system officials will be assessed in an effort to answer the following questions: (1) In practice, is legislative waiver an effective response to violent juvenile offending? (2) As compared to judicial waiver, what are the strengths and weaknesses associated with legislative waiver? (3) What type of case is likely to be “reverse waived” to juvenile court? (4) Is there evidence of any general or specific deterrent effects associated with the legislative waiver of violent juvenile offenders? (5) What are the alternatives, if any, to treating violent juvenile offenders as adults?

Data and Methods The current research is based on a combined quantitative and qualitative design. More specifically, we will utilize a model identified by Creswell (1994) as the “dominant-less dominant” design: The researcher presents the study within a single, dominant paradigm with one small component of the overall study drawn from the alternative paradigm. A classic example of this approach is a quantitative study based on testing a theory in an experiment with a small qualitative interview component in the data collection phase. . . . The advantage of this approach is that it presents a consistent paradigm picture in the study and still gathers limited information to probe in detail one aspect of the study. (p. 177) The primary goal of this study is to quantitatively test the 11 hypotheses mentioned earlier in this chapter. To do so, a quasiexperimental design will be employed, with transfer to adult criminal court acting as the central independent variable. This quantitative research will then be supplemented by qualitative interviews with justice system officials, with the expectation that these interviews will

84

Excluding Violent Youths From Juvenile Court

provide rich descriptive information to better inform and understand the quantitative findings. Subjects As discussed in Chapter 2, Pennsylvania recently enacted legislation that statutorily excludes certain violent youths from juvenile court jurisdiction. The legislative waiver law, which became effective in March 1996, targets two types of juveniles (between the ages of 15 and 18): those who commit a violent offense with a deadly weapon and those who commit a violent offense after previously being adjudicated 34 delinquent on a violent offense. The current research will examine those offenders who were formally processed in Pennsylvania in 1994 and would have been excluded from juvenile court jurisdiction, had the 35 recent legislation been in effect at that time. Specifically, this study analyzes data pertaining to a cohort of 557 male juvenile offenders who were arrested for robbery, aggravated 36 assault, or both, and a deadly weapon was involved in their offense. These youths were between the ages of 15 and 18 at the time of the offense, and they received a juvenile court disposition sometime during 1994. Of the 557 offenders, 138 were transferred to adult criminal court

34

As discussed in Chapter 2, aggravated assault with a deadly weapon is statutorily excluded, but repeat aggravated assault (without a deadly weapon) is not excluded from juvenile court jurisdiction. 35 It might be asked why actual excluded cases were not used for this study. First, many of these cases (even those from 1996) were still being processed by the criminal justice system. This greatly limited the analyses that could be conducted, particularly with regard to recidivism. Second, data were not yet available for these cases, but solid data were available for the 1994 cases. 36 The study is limited to males because of the very small number of female offenders who met these offense criteria. Only 2 transferred female offenders from 1994 were identified, and less than 15 were identified that were retained in juvenile court. The small number of transferred females is consistent with research that shows females make up less than 5% of all transfers (see, e.g., Clarke, 1996; DeFrances & Strom, 1997; Kinder et al., 1995).

85

Research Agenda 37

by judicial waiver and 419 were retained in juvenile court. The essence of the study is to compare those juveniles transferred to adult court with those youths retained in juvenile court, according to the hypotheses presented earlier in this chapter. This study confines itself to robberies and aggravated assaults, involving a deadly weapon, for several reasons. First, these offenses comprise the typical violent juvenile crimes that have evoked concern, fear, and legislative action over the past 10 to 15 years. Second, a preliminary analysis of the data uncovered an extremely small number (less than 10) of other violent offenses with a weapon that would have been excluded from juvenile court jurisdiction under Pennsylvania’s recent legislation. Third, aggravated assault is not included under the provisions of the legislative waiver law that excludes repeat violent offenses that do not involve a deadly weapon. Finally, preliminary contacts with justice system officials revealed that a very high percentage (roughly 95%) of all cases actually excluded since the new law went into effect in 1996 consist of robberies and aggravated assaults with a deadly weapon. It should be noted at the outset that due to a lack of random assignment, this study couldn’t directly confront the problem of 38 selection bias. The cohort of male offenders was selected on the basis of meeting the current criteria for exclusion from juvenile court

37

As discussed in Chapter 2, judicial waiver was the dominant method of transfer that was used in Pennsylvania until March 1996. In 1994, for those juveniles who were never previously found guilty in adult court, the act of murder represented the only crime for which they could be statutorily excluded from juvenile court jurisdiction. Prosecutorial waiver was and is not employed in Pennsylvania. 38 Based on the fact that the offenders in this study were all charged with a violent felony involving a deadly weapon, police selection bias at the time of arrest would not appear to be a major problem. Data limitations preclude an adequate assessment of selection bias at other stages of case processing (see Smith & Paternoster, 1990), and, therefore, results from the statistical analyses should be interpreted with caution.

86

Excluding Violent Youths From Juvenile Court 39

jurisdiction. Those who were transferred to adult court were certified by a juvenile court judge as no longer being amenable to treatment in the juvenile justice system, generally meaning that these cases were seen as being more serious than those retained in juvenile court. Unfortunately, this type of selection bias is nearly unavoidable in transfer research, as random experiments would appear to be neither ethical nor feasible. However, a strong effort will be made to control for variables that may possibly influence the decision to transfer, case outcomes, and the future offending behavior of youths in this study (e.g., prior record, type of weapon involved in the offense, demographic variables, family and school status). In spite of this limitation, it should also be noted that if significant differences do exist between waived and retained youths in terms of case outcomes and subsequent criminal behavior, an argument that these differences are due to selection bias rather than a “transfer effect” may actually be used as reason for caution with regard to the use of legislative waiver. If significant differences arise due to transferred juveniles being more serious and frequent offenders (rather than a differential effect from juvenile versus adult court sanctions), one would have to question whether treating broad categories of youthful offenders as adults is a wise or just policy. In other words, if the juvenile court was previously successful in waiving the “worst” offenders (who may have a greater likelihood of conviction,

39

Cases that were retained in juvenile court and selected for this study were those involving at least one of the two specified violent offenses (robbery and aggravated assault) and a deadly weapons charge (e.g., Prohibited Offensive Weapons, Weapon on School Property, Crimes Committed with a Firearm, Firearm without a License, etc.). Cases that were transferred to adult court and selected for this study were those involving at least one of the two violent offenses and evidence that a deadly weapon was involved in the offense. More extensive data was available for the transferred cases, allowing for a small number of cases (less than 20) to be identified that involved a deadly weapon, but in which no deadly weapons charge was filed. Due to data limitations, these types of cases could not be identified if they were retained in juvenile court.

87

Research Agenda 40

incarceration, and recidivism), a policy of statutorily excluding many more youths (who are not the worst offenders) does not seem prudent. Past research provides little or no support for treating mass quantities of juveniles as adults, and there is reason to believe that this practice could backfire. Specifically, these adolescents may receive inadequate treatment services, be “trained” by adult criminals, perceive they are being treated unjustly, or all three. Consequently, a policy of legislative waiver could create greater numbers of youths who are more likely to offend upon release. In general, then, the subjects for the quantitative portion of the current research represent those violent offenders who would have been excluded from juvenile court jurisdiction in Pennsylvania in 1994, had the more recently enacted legislative waiver law been in effect at that time. In an additional qualitative component of the study, results from interviews with various judges, prosecutors, public defenders, and juvenile probation officials are presented. The explicit purpose of these interviews was to tap the insights of those who are familiar with justice system handling of violent juvenile offenders, and in doing so further assess the effectiveness of the recent legislation. In March, May, and July of 1998, face-to-face interviews were conducted with individuals in three diverse Pennsylvania counties. The first county consisted of a large metropolitan area, the second contained a smaller urban area and rural surroundings, and the third was a suburban county with an additional rural vicinity. Prior to the new legislation becoming effective in 1996, the large metropolitan county judicially waived approximately 150 cases per year (out of roughly 7,500 total dispositions), while the smaller urban county certified approximately 50 cases per year (out of about 1,200 total dispositions), and the suburban county rarely transferred a juvenile case to adult court (out of approximately 500 total dispositions). These counties were selected based on their diverse past use of juvenile transfer, with the expectation that the new law might have a varying impact on different types of jurisdictions. In each county, a juvenile court judge, a prosecutor, a public defender, and 2 to 4 juvenile 40

This issue was brought up in Chapter 4, in a discussion of research by Podkopacz and Feld (1996) that found evidence of greater recidivism among transferred juveniles as compared to similar youths retained in juvenile court.

88

Excluding Violent Youths From Juvenile Court

probation officials were asked a series of open-ended questions pertaining to the use, advantages, and disadvantages of legislative waiver. A total of 18 justice system officials, all possessing extensive experience working with violent youths, were interviewed. Data Source Case information pertaining to the juvenile offenders discussed above was obtained through The Center for Juvenile Justice Training and Research (CJJT&R), located at Shippensburg University (PA). The Center was established and is managed by the Juvenile Court Judges’ Commission (JCJC), a Commonwealth of Pennsylvania agency within the Governor’s Office and it’s Office of General Counsel. In addition to providing a number of training and educational programs to juvenile probation officers from across the state, CJJT&R also operates a Statistical Analysis Center that compiles data and publishes an annual report on the activities of all juvenile courts in Pennsylvania. Because of this latter responsibility, CJJT&R maintains individual county data that enables the current study to be completed. In order to receive funding from JCJC, each county in Pennsylvania must submit offender and offense information pertaining to every juvenile court disposition handled within its jurisdiction. While providing extensive data on all youths handled by county juvenile courts, the reporting system also includes all cases in which “transfer to 41 criminal court” is identified as the final disposition. Data for the current study were taken directly from the database maintained by CJJT&R and are limited to the information that the agency considers important for its purposes. Consequently, some variables that may be relevant to a discussion of juvenile justice processing and transfer to 42 adult court were not available. Finally, missing data pertaining to 41

CJJT&R tracked all cases judicially waived to adult court in 1994 (N=453). For the current study, data were used pertaining to the 138 certified cases that involved robbery or aggravated assault charges and use of a deadly weapon. 42 For example, there are no socio-economic indicators. Socio-economic status has been found to influence processing decisions (Blumstein et al., 1983; Frazier & Bishop, 1995; Sampson, 1986; Wordes & Bynum, 1995) and may be important in the decision to transfer a case to adult court.

Research Agenda

89

variables employed in this study were obtained though direct contacts with various county offices (i.e., district attorney, adult probation, and 43 juvenile probation offices). On the following pages, Table 1 presents statistical descriptions (i.e., means, standard deviations, ranges) of all the variables employed in this study, based on the entire cohort of offenders. Table 2 then provides comparative, descriptive information for the separate transferred and nontransferred offenders. Independent Variable The key independent variable in this study is transfer to adult criminal court. Based on the research design, the independent variable does not represent a true manipulation; Instead, it is a “treatment” given to an assigned group that may differ from the comparison group in terms of criminal activity and demographic characteristics (see Campbell & Stanley, 1966; Cook & Campbell, 1979). This variable (TRANSFER) was coded as 0 if the offender was retained in juvenile court and 1 if the offender was waived to adult criminal court. Of the entire cohort of 557 offenders, 25% were transferred to adult criminal court. Control Variables Based on the quasi-experimental design employed in this study, numerous variables will be utilized to control for any influence they may have on the decision to transfer, case outcomes, future criminal behavior, or all three. First is age. Harsher penalties tend to be associated with older offenders than with younger offenders who are beginning their delinquency career (Wolfgang et al., 1972), and an older age at the time of offense has been found to predict judicial certification to adult criminal court (Eigen, 1981a, 1981b; Fagan & Deschenes, 1990; Podkopacz & Feld, 1996; Poulos & Orchowsky, 1994). To control for these factors, age at referral (AGEREF) was included as a continuous variable. The mean age at referral for the 43

Consisting of phone calls, letters, and visits to county offices by the author.

Table 1: Descriptive Statistics For All Variables Variable

Mean

SD

Min

Max

N

TRANSFER AGEREF RACE COUNTY SCHOOL FAMILY WEAPON AGEFIRST PRIORREC PRIORVIO RELEASE CONVICT TARCON INCARCE INLENGTH PROCESS PREARR PREVIOL RISKTIME POSTARR POSTVIOL TIMEARR

0.25 16.20 0.81 0.86 0.71 0.76 0.86 14.76 3.23 0.15 0.40 0.68 0.78 0.74 2.80 4.21 0.25 0.13 17.88 0.31 0.17 6.74

0.43 0.85 0.39 0.34 0.45 0.43 0.34 1.72 4.00 0.35 0.49 0.47 0.41 0.44 0.90 1.19 0.43 0.34 0.80 0.46 0.38 4.34

0.00 15.00 0.00 0.00 0.00 0.00 0.00 10.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.00 0.00 0.00 1.00

1.00 18.00 1.00 1.00 1.00 1.00 1.00 18.00 29.00 1.00 1.00 1.00 1.00 1.00 5.26 6.72 1.00 1.00 18.00 1.00 1.00 18.00

557 557 557 557 557 557 557 557 557 557 557 557 378 378 280 557 224 224 494 494 494 152

90

Table 2: Descriptive Statistics By Transfer Status Variable AGEREF RACE COUNTY SCHOOL FAMILY WEAPON AGEFIRST PRIORREC PRIORVIO RELEASE CONVICT TARCON INCARCE INLENGTH PROCESS PREARR PREVIOL RISKTIME POSTARR POSTVIOL TIMEARR

Transferred Offenders Mean SD Min 16.70 0.67 15.00 0.78 0.42 0.00 0.79 0.41 0.00 0.60 0.49 0.00 0.78 0.41 0.00 0.75 0.43 0.00 14.13 2.10 10.00 6.38 5.45 0.00 0.30 0.46 0.00 0.55 0.50 0.00 0.87 0.34 0.00 0.85 0.36 0.00 0.96 0.20 0.00 3.31 0.95 0.00 5.32 0.97 0.00 0.46 0.50 0.00 0.29 0.46 0.00 17.38 1.68 9.00 0.38 0.49 0.00 0.24 0.43 0.00 5.77 4.61 1.00

Max 18.00 1.00 1.00 1.00 1.00 1.00 17.00 29.00 1.00 1.00 1.00 1.00 1.00 5.26 6.72 1.00 1.00 18.00 1.00 1.00 18.00

N 138 138 138 138 138 138 138 138 138 138 138 120 120 115 138 76 76 79 79 79 30

(Continued . . .)

91

92

Excluding Violent Youths From Juvenile Court

Table 2 (continued)

Variable AGEREF RACE COUNTY SCHOOL FAMILY WEAPON AGEFIRST PRIORREC PRIORVIO RELEASE CONVICT TARCON INCARCE INLENGTH PROCESS PREARR PREVIOL RISKTIME POSTARR POSTVIOL TIMEARR

Nontransferred Offenders Mean SD Min 16.04 0.83 15.00 0.82 0.38 0.00 0.89 0.32 0.00 0.74 0.44 0.00 0.75 0.43 0.00 0.90 0.30 0.00 14.96 1.52 10.00 2.19 2.69 0.00 0.10 0.29 0.00 0.35 0.48 0.00 0.62 0.49 0.00 0.75 0.43 0.00 0.64 0.48 0.00 2.45 0.67 0.00 3.84 1.01 0.00 0.14 0.34 0.00 0.05 0.23 0.00 17.98 0.40 10.00 0.29 0.46 0.00 0.16 0.36 0.00 6.98 4.26 1.00

Max 18.00 1.00 1.00 1.00 1.00 1.00 18.00 17.00 1.00 1.00 1.00 1.00 1.00 3.89 6.66 1.00 1.00 18.00 1.00 1.00 18.00

N 419 419 419 419 419 419 419 419 419 419 419 258 258 165 419 148 148 415 415 415 122

entire cohort was 16.2 years. The comparative breakdown shows that the transferred offenders were older (16.7 years), on average, than the youths retained in juvenile court (16.04 years), which corresponds with past research findings that older youths are more likely to be waived to adult court. Race will also be used as an offender characteristic. Although recent research has failed to find direct racial bias in the transfer process (Fagan & Deschenes, 1990; Fagan et al., 1987; Podkopacz & Feld, 1996; Poulos & Orchowsky, 1994), racial characteristics are hypothesized by labeling theory to influence justice system outcomes (Paternoster & Iovanni, 1989). For this study, race (RACE) was coded

93

Research Agenda 44

as 0 for white and 1 for nonwhite. Overall, the cohort was 81% nonwhite. The comparison of transferred and nontransferred youths indicates a slight difference by race, as 82% of the nontransferred and 78% of the transferred offenders were nonwhite. Several social factors will also be measured and utilized as control variables. Research shows that juvenile justice processing and case outcomes often depend on whether the offender was handled in a rural, suburban, or urban setting (Feld, 1993; Smith & Paternoster, 1990), and location has been found to have a significant influence on the transfer decision (Feld, 1989; Hamparian et al., 1982; Poulos & Orchowsky, 1994). With this in mind, county of jurisdiction (COUNTY) was coded 45 as 1 for urban and 0 for suburban/rural. Of the entire cohort, 86% of the youths were processed in urban counties. Seventy-nine percent of the transferred offenders were processed in urban counties, compared to 89% of the youths retained in juvenile court. This again corresponds with past research indicating that urban youths are less likely to be waived to adult court. Two other important social factors pertain to the youth’s family and school situations. Traditionally, home and school environments have received strong consideration by the juvenile court (Empey & Stafford, 1991; Bernard, 1992), and both may have a significant 44

Because of the way race was coded in the data we employed, only whites and nonwhites could be distinguished. However, the nonwhite category is made up almost entirely of African Americans. 45 This variable was measured based on a county classification system existing in Pennsylvania, which was obtained from The Center for Juvenile Justice Training and Research. Counties in the state are classified based on population, on a scale of 1 to 8, with 1 representing the largest population. Philadelphia County is the only Class 1 county, and Allegheny County (Pittsburgh) is the only Class 2 county. Both were coded as urban. All other counties are classified from 2A through 8. Class 2A and Class 3 counties were coded as suburban if they were adjacent to Philadelphia or Allegheny County; Otherwise, they were coded as urban. Class 4 counties were coded as suburban if they were adjacent to an urban county; otherwise, they were coded as rural. Class 5, Class 6, Class 7, and Class 8 counties were coded as rural. The small number (n=22) of offenders processed in rural counties precluded a separation of rural and suburban counties.

94

Excluding Violent Youths From Juvenile Court

influence on the transfer decision (see Podkopacz & Feld, 1996; Singer, 1993), case outcomes, future criminal behavior, or all three. To control for these factors, school status (SCHOOL), at the time of referral, was coded as 0 for not enrolled and 1 for enrolled, graduated, or GED. Family status (FAMILY), also at the time of referral, was coded as 0 for living with two parents and 1 for other living arrangements (e.g., one parent, relative, in placement, living independently). While 71% of the entire cohort was enrolled in school, graduated, or obtained a GED, only 60% of the transferred offenders had this same school status, compared to 74% of the nontransferred youths. Less difference can be seen in terms of family status. Overall, only 24% of the offenders in the cohort were living with two parents, a figure that is similar to that of both the transferred (22%) and nontransferred (25%) youths. While the current research is limited to those offenders charged with robbery, aggravated assault, or both, use of a deadly weapon was also present as an offense characteristic. According to Pennsylvania 46 law, deadly weapons can encompass a wide variety of devices. Furthermore, the type of weapon employed during the offense may have an influence on the decision to transfer and subsequent case outcomes (e.g., use of a firearm could increase the probability of transfer as compared to other weapons and could also influence the way cases are prosecuted). Consequently, weapon type (WEAPON) will be used as a control variable, coded as 1 for firearm and 0 for other deadly weapon. Eighty-six percent of the cohort of offenders used a firearm during the commission of the alleged offense. Somewhat surprisingly, 90% of the nontransferred youths employed a firearm, compared to only 75% of the transferred offenders. Finally, it will be imperative to consider the prior offense history of the offenders in this study, in order to account for differences in past delinquent behavior between those youths who were transferred to adult court and those who remained in juvenile court. Prior offense history has been found to be a strong predictor of dispositional outcome (Blumstein et al., 1983; Blumstein, Cohen, Roth, & Visher, 1986; Thornberry & Christenson, 1984), meaning offenders with lengthy criminal histories may be treated in a retributive manner, while those with less extensive records may be treated more leniently. Also, prior 46

For example, firearms, clubs, knives, etc. (see footnote 16 in Chapter 1).

Research Agenda

95

record is an important consideration in research on recidivism, especially in studies that examine the effects of different justice system policies and programs. In general, high risk youths with the most extensive criminal histories would be expected to receive the most severe dispositions, but they would also be more likely to commit new offenses regardless of any relationship between dispositions and future offending (Smith & Paternoster, 1990). Based on these concerns, various measures will be utilized to control for prior offending. First, because of the strong relationship that exists between early onset and subsequent serious, violent, and chronic offending (Blumstein et al., 1986; Farrington, 1986, 1998; Loeber & LeBlanc, 1990; Thornberry, Huizinga, & Loeber, 1995), age at first referral (AGEFIRST) to the juvenile court will be included as a continuous variable. The mean age at first referral for the entire cohort was 14.76 years. The comparative breakdown shows nontransferred youths to have an older age at first referral (14.96 years), on average, than that of the transferred offenders (14.13 years). Second, three variables (prior referrals, prior adjudications, and prior placements) were combined into a prior record composite scale, which measures the extent of each offender’s delinquent history. Prior referrals represent the total number of times a youth was previously referred to the juvenile court. Prior adjudications represent the total number of times a youth was previously adjudicated delinquent. Prior placements represent the total number of times a youth was placed in a juvenile correctional facility as a result of a juvenile court disposition. After adding these three measures together, the mean score on the prior record scale (PRIORREC) for the entire cohort was 3.23. As expected, the mean score for the transferred offenders was much higher (6.38) than that of the nontransferred youths (2.19). Finally, to measure the seriousness of prior offending, a dummy variable will be used to identify those juveniles who were previously adjudicated delinquent on one of the violent felonies targeted in the recently enacted legislative waiver law in Pennsylvania. This prior violent felony variable (PRIORVIO) was coded as 1 if a prior violent felony was substantiated and 0 if no prior adjudication of delinquency on a violent felony occurred. Of the entire cohort, 15% had a prior violent felony substantiated. Of the nontransferred youths, only 10% had a prior violent felony substantiated, compared to 30% of the transferred offenders.

96

Excluding Violent Youths From Juvenile Court

Dependent Variables Eleven dependent variables will be evaluated. Six are case outcome variables, while the remaining five pertain to future criminal behavior. The case outcome variables are (a) release from custody prior to final disposition, (b) conviction, (c) conviction on a target offense, (d) incarceration, (e) length of incarceration, and (f) case processing time. The recidivism variables are (a) rearrest prior to final disposition, (b) violent felony rearrest prior to final disposition, (c) rearrest following final disposition, (d) violent felony rearrest following final disposition, and (e) time to rearrest following final disposition. Release from custody prior to final disposition pertains to whether the offender was released from detention or secure custody prior to sentencing (if the case resulted in a finding of guilt in either juvenile or adult court) or adjudication (if the case resulted in a dismissal or 47 acquittal in either juvenile or adult court). Release (RELEASE) was coded as 0 if the offender remained in custody during this time period and 1 if the youth was released. Of the entire cohort, 40% were released from custody sometime prior to final disposition. An examination of the transferred and nontransferred youths finds that a larger percentage of the waived offenders (55%) were released than of the youths retained in juvenile court (35%). Two aspects of punishment certainty will be examined, conviction and conviction on a target offense of robbery or aggravated assault. Conviction (CONVICT) was coded as 0 if the case resulted in a dismissal or acquittal and 1 if any charges were substantiated in juvenile or adult court. Sixty-eight percent of the cohort of offenders was convicted in one court or the other. A larger percentage of the transferred youths (87%) were convicted than of the nontransferred juveniles (62%). Target conviction (TARCON) was coded as 1 if the offender was convicted of robbery or aggravated assault and 0 if neither of these charges were substantiated. Of those youths who were convicted in either juvenile or adult court, 78% were convicted on a target offense. 47

For this study, secure custody was defined as an adult jail, a juvenile detention center, or some other juvenile correctional facility. Other types of monitoring programs (e.g., electronic monitoring, day treatment facilities, house arrest) were not considered to be secure custody.

Research Agenda

97

Of those transferred and convicted offenders, 85% were convicted of robbery or aggravated assault, compared to 75% of the youths retained and adjudicated delinquent in juvenile court. Two aspects of punishment severity will also be examined, incarceration and length of incarceration. As with target conviction, the incarceration variable pertains only to those offenders who had charges substantiated against them in juvenile or adult court, while the incarceration length variable pertains only to those offenders who were incarcerated. Incarceration (INCARCE) was coded as 1 if a sentence of secure confinement was imposed and 0 if the sentence did not involve 48 incarceration. Of those offenders who were convicted, 74% received a sentence of incarceration. A much larger percentage of the convicted, transferred youth (96%) were incarcerated than of the offenders retained and convicted in juvenile court (64%). Incarceration length is the length of time ordered by the court for secure confinement, measured in months. Due to this variable being positively skewed, the natural log of incarceration length will be used as the dependent variable. For youths retained in juvenile court, this variable represents the actual time served in a secure correctional 49 facility. For juveniles transferred to adult court, this variable

48

Incarceration was defined as a state or county prison, a state-run Youth Development Center, or a privately run juvenile correctional facility (e.g., Glen Mills, near Philadelphia, PA; George Junior Republic, near Erie, PA). Sentences to or placements in residential drug and alcohol facilities, wilderness programs, and intensive supervision programs were not considered to be incarceration. 49 Juvenile courts in Pennsylvania employ indeterminate sentencing, meaning that youths placed in secure correctional facilities remain incarcerated until they are deemed suitable for release by a juvenile court judge (the juvenile court can maintain jurisdiction up to the age of 21). Incarceration length for youths in juvenile court was measured as the total number of months between the date of disposition and the date of release from placement. At the time this study was conducted, four youths who had been placed in a correctional facility had not yet been released. For these offenders, it was assumed that they would remain incarcerated until the age of 21, which, in all four cases, amounted to less than an additional 12 months.

98

Excluding Violent Youths From Juvenile Court 50

represents the minimum prison sentence imposed. Due to data limitations, actual time served in prison could not be computed (see Fritsch et al., 1996). However, because of Pennsylvania’s sentencing procedures, minimum sentences and actual time served will tend to be 51 similar. For the entire cohort of offenders, the mean of the natural log of incarceration length (INLENGTH) was 2.80 (16.4 months). The comparative breakdown suggests that the transferred youths who were incarcerated received longer sentences (natural log of 3.31, or 27.4 months), on average, than did the offenders retained and placed by juvenile courts (natural log of 2.45, or 11.6 months). Case processing time represents the swiftness of punishment imposed on offenders in juvenile and adult court. Processing time was measured in days, as a continuous variable. Again, due to this variable being positively skewed, the natural log of processing time will be used as the dependent variable. For those offenders retained in juvenile court, this variable amounted to the total number of days from initial 52 referral to final disposition. For those youths waived to adult court, 50

Adult criminal courts in Pennsylvania use an indeterminate sentencing scheme for prison sentences. The sentencing judge imposes both a minimum and maximum period of incarceration, with the minimum term being no more than one-half of the maximum. Sentences with a maximum term of less than 2 years are considered county-level sentences and are served in a county prison. Sentences with a maximum term of 2 years or more are considered state-level sentences and are generally served in a state prison, although the sentencing judge can order that the defendant remain in the county prison. 51 For state-level prison sentences (i.e., those with a maximum term of 2 years or more), the defendant must serve at least the minimum term imposed prior to being released by the state parole board. There is no "earned time" or "good time" available for those offenders serving a state sentence, even if the sentence is being served in a county prison. However, individual counties may establish early release programs for those defendants serving a county-level prison sentence (i.e., those with a maximum term of less than 2 years), and county judges do have the authority to order the parole of an inmate prior to the completion of the minimum term of a county sentence. 52 For offenders in either juvenile or adult court, the date of final disposition is the date of sentencing (if the youth was previously convicted) or the date of case dismissal or acquittal.

Research Agenda

99

processing time was measured as the total number of days from the date of transfer to final disposition. The mean of the natural log of processing time (PROCESS) for the entire cohort was 4.21 (67.4 days). A comparison of the transferred and nontransferred offenders implies that the youths waived to adult court were subjected to longer case processing times (natural log of 5.32, or 204.4 days), on average, than those of the offenders retained in juvenile court (natural log of 3.84, or 46.5 days). Recidivism measures were constructed from data on official arrests that occurred both prior to and following final disposition. Official arrest information was also acquired through The Center for Juvenile Justice Training and Research and was based on state juvenile records, county criminal records (obtained through direct contacts with district attorney and adult probation offices), and a statewide criminal records 53 check conducted by the Pennsylvania State Police. Although the rearrest data was restricted to arrests occurring in Pennsylvania, the youthfulness of the offenders in this study would appear to limit both their mobility and the likelihood that they would be rearrested in 54 another state. The recidivism variables reflect multiple measures of official criminal activity: rearrest prior to and following final disposition, violent felony rearrest prior to and following final disposition, and time 53

In a discussion of the limitations of rap sheets in criminological research, Geerken (1994) noted, "The problem of incompleteness in the list of arrests because of non-submission/rejection can be addressed only by combining databases from different sources. Whenever possible, multiple official databases should be merged and their inconsistencies carefully resolved" (p. 19). 54 Of biggest concern might be that the youths from Philadelphia could be rearrested in an adjacent state (i.e., New Jersey or Delaware), especially since a substantial percentage of the offenders in this study were from Philadelphia County. However, contacts with juvenile probation officials in Philadelphia County revealed that youths under supervision there are very rarely arrested in another jurisdiction. Furthermore, in their follow-up study of the Philadelphia birth cohort of 1945, Wolfgang and his colleagues (1987, p. 10) found that of all the arrests uncovered for each subject up to the age of 30, approximately 93% occurred in Philadelphia.

100

Excluding Violent Youths From Juvenile Court

to rearrest following final disposition. Although the use of arrest records in measuring recidivism is inherently problematic, in that not all future offenses result in detection (see Maltz, 1984), multiple measures of officially recorded contacts with the law have been used successfully in various studies that have examined the specific deterrent effect of formal sanctions (see, e.g., Bishop et al., 1996; Fagan, 1995; Podkopacz & Feld, 1996; Smith & Gartin, 1989). The current study, unlike prior transfer research, will examine the future criminal behavior of violent youths both prior to and following final disposition. To do so requires a consideration of time at risk, or time that each offender was “on the street” and capable of committing further crimes. Unfortunately, exact dates of release on bail were not always available for youths who were transferred to adult court. To deal with this problem, analyses of recidivism occurring prior to final disposition will be limited to those offenders who were known to be released from secure custody either prior to sentencing (if the case resulted in a finding of guilt in juvenile or adult court) or adjudication (if the case resulted in a dismissal or acquittal in juvenile or adult court). Additionally, case-processing time will be employed as a control variable. Although inexact, this variable will account for heterogeneity in time at risk among those who were released. All else being equal, the probability of subsequent criminal behavior would be expected to increase with the length of the pre-dispositional period (see Smith & Paternoster, 1990). For youths released prior to final disposition, pre-dispositional arrest (PREARR) was coded as 1 if the offender was arrested during this time period and 0 if no arrest occurred. For the entire cohort, 25% of the released offenders were rearrested prior to final disposition. A comparison of the transferred and nontransferred youths suggests greater recidivism among those treated as adults. Of the waived and released youths, 46% were rearrested prior to final disposition, in contrast to only 14% of the offenders retained and released by a juvenile court. Violent felony arrests will be used to measure the seriousness of recidivism prior to final disposition. This variable will identify those youths who were rearrested on one of the violent felonies targeted in the recent Pennsylvania legislation. Pre-dispositional violent felony (PREVIOL) was coded as 1 if a felony arrest occurred and 0 if no felony arrest took place. Overall, 13% of the released offenders were

Research Agenda

101

rearrested for a violent felony offense prior to final disposition. Again, a comparison of the transferred and nontransferred youths implies more serious recidivism among those treated as adults. Only 5% of the offenders retained and released by a juvenile court were rearrested on a violent felony charge prior to final disposition, compared to 29% of those waived and subsequently released by an adult court. Post-dispositional analyses of recidivism will be limited to those offenders who were either not incarcerated or were released from incarceration sometime prior to December 31, 1997. Of the 557 youths in the cohort, 494 received a disposition and were released back into the community prior to that date. Subsequent recidivism of these 55 offenders will be examined over an 18-month follow-up period. Of the 494 youths identified, 480 had experienced at least an 18-month exposure to risk of rearrest by June 30, 1998 (the recidivism cut-off date for this study). For the remaining 14 offenders, exposure risk ranged from 9 to 17 months. To address these differential exposure times, time at risk 56 (RISKTIME) will be used as a control variable. Overall, the entire cohort experienced a mean time at risk of 17.88 months. The comparative breakdown indicates that the youths retained in juvenile court experienced a slightly greater time at risk (17.98 months), on average, than did the waived youths (17.38 months), but it should be remembered that the vast majority of offenders were in the follow-up period for the full 18 months. With regard to recidivism following final disposition, three outcome variables will be examined. For these youths, postdispositional arrest (POSTARR) was coded a 1 if the offender was arrested during this time period and 0 if no arrest occurred. Of the 55

Although the follow-up period is relatively short, youths in this study were at peak ages for offending. As discussed by Bishop et al. (1996, p. 177), if the "get tough" philosophy has merit, one would expect a short-term follow-up to favor transfers over nontransfers in comparisons of recidivism. This would be due to an expected short-term specific deterrent effect resulting from transfer to adult court. 56 For examples of other studies employing time at risk as a control variable, see Smith and Paternoster (1990), Smith and Polsenberg (1992), and Smith, Wish, and Jarjoura (1988).

102

Excluding Violent Youths From Juvenile Court

entire 494 youths, 31% were rearrested following final disposition. A comparison of the transferred and nontransferred youths again suggests greater recidivism among those treated as adults. Of the waived youths, 38% were rearrested following final disposition, as compared to 29% of 57 the youths retained in juvenile court. Violent felony arrests will again be used to measure the seriousness of recidivism following final disposition. Post-dispositional violent felony (POSTVIOL) was coded as 1 if a violent felony arrest occurred and 0 if no violent felony arrest took place. Overall, 17% of the offenders were rearrested for a violent felony offense following final disposition. Of the youths retained in juvenile court, 16% were rearrested on a violent felony charge following final disposition, compared to 24% of those waived to adult court. Finally, in order to analyze time to failure (i.e., time between postdispositional release and rearrest), time to rearrest (TIMEARR) was measured in months. Overall, of those 152 youths who were rearrested, 58 the mean time to failure was 6.74 months. The comparative breakdown suggests that the transferred youths experienced a shorter time to rearrest (5.77 months), on average, than did the offenders retained in juvenile court (6.98 months). Statistical Analyses Following a presentation and discussion of the zero-order correlations among the variables used in this study, several multivariate techniques will be utilized to test the hypotheses presented at the outset of this chapter. Multivariate analysis is necessary to account for the control variables discussed above, and the specific technique employed 57

These percentages are somewhat similar to those revealed by Bishop et al. (1996) in their initial Florida study. Their findings indicated that during a postdispositional follow-up period of less than 2 years, 30% of the transferred juveniles were rearrested, compared to 19% of the youths retained in juvenile court. It should be noted that the offenders included in their study were charged with a wide variety of offenses (i.e., both violent and property offenses). 58 These descriptive failure times are based only on those youths who were rearrested following final disposition. As discussed shortly, we will conduct a survival analysis to further assess time to rearrest, which will consider both the occurrence and the timing of recidivism.

Research Agenda

103

will depend on the nature of the dependent variable (e.g., a continuous versus dichotomous dependent variable). Two hypotheses contain a dependent variable that is measured continuously: Hypothesis 5 (with natural log of incarceration length as the dependent variable) and Hypothesis 6 (natural log of case processing time). Accordingly, ordinary least squares (OLS) multiple regression procedures (Lewis-Beck, 1980; Neter, Wasserman, & Kutner, 1989; Tacq, 1997) appear appropriate for examining the effect of the independent variable (i.e., transfer to adult court) on these 59 dependent variables, while controlling for other influences. Eight hypotheses contain a dichotomous dependent variable: Hypothesis 1 (release from pre-dispositional custody), Hypothesis 2 (conviction), Hypothesis 3 (conviction on a target offense), Hypothesis 4 (incarceration), Hypothesis 7 (pre-dispositional arrest), Hypothesis 8 (pre-dispositional violent felony arrest), Hypothesis 9 (postdispositional arrest), and Hypothesis 10 (post-dispositional violent felony arrest). Since these dependent variables have only two options, multivariate logistic regression (Aldrich & Nelson, 1984; Menard, 1995; Pindyck & Rubinfeld, 1981) appears to be an appropriate statistical technique. Logistic regression, which utilizes maximum likelihood estimation, is designed for a dichotomous dependent variable and overcomes the problems that a binary dependent variable presents 60 for OLS assumptions. While the coefficients obtained through linear regression indicate the amount of change in the dependent variable that is associated with a one-unit change in the independent variable,

59

Use of OLS is contingent upon the necessary diagnostic procedures. One basic assumption is that the dependent variable is normally distributed. As mentioned previously, because the distributions of incarceration length and case processing time were both positively skewed, the natural logs of these dependent variables were used to analyze the data. The distributions of the natural logs appeared normal (the natural log of case processing time was slightly negatively skewed), and later residual analyses revealed that the residuals approached a normal distribution, further supporting the use of OLS regression procedures. 60 For example, heteroscedasticity, nonnormal error term, nonlinearity, and predicted probabilities beyond 1.0.

104

Excluding Violent Youths From Juvenile Court

logistic regression examines the log odds of an event occurring given a one-unit change in the independent variable. Finally, Hypothesis 11 uses time to arrest as the dependent variable. This variable not only measures how quickly a youth reoffends following disposition and release from secure confinement, but it can also be thought of as a “proxy” for the frequency of future offending, as high rate offenders would most likely have a shorter time until failure 61 than that of low rate offenders. While many recidivism studies rely totally upon either a binary measure of future offending or a count of future offenses, the consideration of time to failure adds another dimension of recidivism and also requires an additional statistical technique. To this end, the use of event history or survival analysis has recently become quite popular (see, e.g., Baumer, 1997; Benedict & Huff-Corzine, 1997; DeJong, 1997; Fagan, 1995; Lattimore, Linster, & MacDonald, 1997; Lattimore, Visher, & Linster, 1995; Schmidt & Witte, 1989; Smith & Akers, 1993; Visher, Lattimore, & Linster, 1991). In general, survival analysis can provide a more complete examination of recidivism than can be achieved through the more traditional regression techniques alone (e.g., ordinary least squares, logistic regression). Hazard models, which are used in survival analysis, consider both the occurrence and the timing of recidivism, and they account for the bias that results from estimating regression equations based on censored data (a case is censored if the offender does not experience a rearrest during the follow-up period; see Allison, 1984; Schmidt & Witte, 1988). Therefore, this study will include a survival analysis of the recidivism data, in order to see if transfer to adult court has any effect on the time until rearrest, while controlling for other influences. Specifically, a Cox proportional hazards model (Allison, 1984; Cox, 1972) will be employed, which permits covariates to be included with the key independent variable and subsequently be tested for significance. 61

Due to a reliance on official arrest records, the first arrest following disposition will be defined as a failure, rather than the first new crime. However, the factors affecting timing of first arrest and those affecting timing of first new crime are usually similar, if the timing of these events is independent (Holden, 1985).

Research Agenda

105

Models The current research will examine the effect of transferring violent youths to adult court through the use of several models. To test each hypothesis, a “transfer only” model will be estimated first, containing transfer to adult court as the only independent variable. These models, which will correspond to the estimated bivariate correlations, will reveal the strength of the relationship between transfer and each dependent variable and the explanatory utility of the central independent variable. Along with the transfer only models, full models will be estimated containing transfer to adult court and the control variables discussed above. These models will include the various offense and offender characteristics available for this study, with the purpose of investigating whether the control variables account for any effect of transfer on each 62 dependent variable. In addition to examining whether transfer to adult court has a significant effect while controlling for other factors, the results of the full models will be further employed to assess the impact of transfer. In general, the results will be used to make separate predictions about the dependent variables when transfer is equal to 0 in the full models and also when transfer is equal to 1, with all other explanatory variables set 63 at their mean. For each dependent variable, the two separate 62

For analyses of recidivism during the post-dispositional period, release from pre-dispositional custody, conviction, incarceration length, and case-processing time will also be employed as explanatory variables, to investigated whether any of these four variables mediate a transfer effect. 63 To calculate these predictions, the means of the explanatory variables were computed based on only the offenders who were included at each stage of the data analysis. The analyses of target conviction and of incarceration included only those youths who were convicted (n = 378). The means of the explanatory variables for those youths who were incarcerated (n = 280) were used in the analysis of incarceration length. The means of the explanatory variables for those youths who were released from secure custody prior to final disposition (n = 224) were employed in the analyses of pre-dispositional recidivism. Finally, the means of the explanatory variables for those youths who were back on the street following final disposition (n = 494) were used in the analyses of postdispositional recidivism.

106

Excluding Violent Youths From Juvenile Court

predictions will be inspected with an eye toward judging the strength of the influence of transfer to adult court. The formulas used to make these predictions will depend on the nature of the dependent variable in each model. For the two OLS regression models, which contain dependent variables that are measured continuously, the following equation will be employed:

yˆ = a 0 + b1 x1 + b2 x 2 + b3 x3 +  bk x k

(1)

where a0 is the constant and the subscripts identify each independent variable and the corresponding slope estimate (Bachman & Paternoster, 1997; Lewis-Beck, 1980). Concerning the eight logistic regression models, which contain dichotomous dependent variables, the results of the logit analyses will be used to estimate the probability of the dependent variable occurring. The following equation will be utilized:

pˆ =

e a0 +b1 x1 +b2 x2 +b3 x3 +bk xk 1 + e a0 +b1x1 +b2 x2 +b3 x3 bk xk

(2)

where a0 again represents the constant and the subscripts identify each independent variable and the corresponding slope estimate (Bachman & Paternoster, 1997; Menard, 1995). Finally, with regard to the survival analysis, the impact of transfer to adult court will be assessed based on the “relative risk” associated with the key independent variable in this study. In a Cox proportional hazards model, the natural log of the hazard function (or failure rate at a given point in time) is used as the dependent variable (Allison, 1984; 64 Cox, 1972). Concerning the explanatory variables, those with positive coefficients are associated with an increased risk of failure (or 64

The hazard function is derived from the cumulative survival function, or the proportion of cases surviving at a given point in time. The hazard function shows how likely a case is to experience failure, given that it has survived to that point in time. In other words, it is not a probability, but a failure rate per unit of time. In contrast to the usual regression models, which may be used to predict a single value for an individual case, the Cox regression model produces an entire hazard curve for each value of an explanatory variable.

107

Research Agenda

decreased survival times), while those with negative coefficients are associated with a decreased risk of failure (or increased survival times). For a dichotomous independent variable, such as transfer, the presence of the characteristic measured by the variable may increase, decrease, or not influence the risk of failure at a particular point in time. In order to examine the impact of transfer to adult court, the coefficients generated by the hazards model can be exponentiated (i.e., take their antilogs). For a dichotomous variable, the exponentiated coefficient provides the relative risk of failure for the cases corresponding to the separate values of the variable, while controlling for other factors (Allison, 1984; Cox, 1972). For our purposes, if the exponentiated coefficient of transfer is greater than 1.0, the risk of rearrest at a given point in time for those offenders who were transferred was greater than for those who were retained in juvenile court (i.e., transferred youths experienced shorter times to rearrest). If the exponentiated coefficient of transfer is less than 1.0, the opposite would be concluded. The impact of transfer will be assessed, then, based on the size of this relative risk. To complete our analyses, “transfer specific” models will also be estimated, which amount to separate full models for the transferred and nontransferred youths. After estimating these equations, statistical tests of the differences between the coefficients will be conducted, using the following equation:

z=

b1 − b2

(se1 )2 + (se2 )2

(3)

where b1 is the slope coefficient of group 1, b2 is the slope coefficient of group 2, se1 is the standard error of b1, and se2 is the standard error of b2 (Clogg, Petkova, & Haritou, 1995; Brame, Paternoster, Mazerolle, & Piquero, 1998; Paternoster, Brame, Mazerolle, & Piquero, 1998). By allowing each explanatory variable to have a group specific effect (i.e., in the transfer and nontransfer groups), interaction effects can be examined. In other words, this process will assess whether any variables exert a significantly different effect in one group as compared to the other group. A different strategy could be employed to address this issue, using an interaction model with multiplicative terms involving transfer to adult court and all other explanatory variables.

108

Excluding Violent Youths From Juvenile Court

However, multicollinearity would become a major concern with this latter method, as transfer would be a common component in all the interaction terms.

CHAPTER 5

Quantitative Findings

Bivariate Results In the following pages, Table 3 presents the estimated zero-order correlations between all variables employed in this study. Bivariate coefficients do not control for the effects of other variables. Therefore, they should not be used to make conclusions about cause and effect relationships. However, several points are worth noting. To begin, the bivariate correlations among the independent and control variables suggest that none of the explanatory variables were highly correlated with any of the other explanatory variables. Only two of these zero-order correlations were greater than |.50|, and both were 65 under |.65|. However, as suggested by Lewis-Beck (1980), Menard (1995), Neter et al. (1989), and Tacq (1997), further tests were conducted to examine the possibility of multicollinearity. Tolerance statistics and variance inflation factors confirmed that multicollinearity was not problematic, as all tolerances were greater than .40 and all 66 variance inflation factors were less than 2.5. 65

The estimated coefficient for transfer and the natural log of case processing time (case processing time was used as an explanatory variable in postdispositional analyses of recidivism) was .5413, while the estimated coefficient for age at first offense and prior record was -.6026. 66 As stated by Menard (1995), "A tolerance of less than .20 is cause for concern; a tolerance of less than .10 almost certainly indicates a serious collinearity problem" (p. 66). Similarly, in a discussion of variance inflation factors, Neter et al. (1989) reported, "A maximum VIF value in excess of 10 is often taken as an indication that multicollinearity may be unduly influencing the least squares estimates" (p. 409).

109

Table 3: Bivariate Correlations Among All Variables Variable 1 TRANSFER

1 1.00 (557) 2 AGEREF .34* (557) 3 RACE -.05 (557) 4 COUNTY -.12* (557) 5 SCHOOL -.14* (557) 6 FAMILY .03 (557) 7 WEAPON -.18* (557) 8 AGEFIRST -.21* (557) 9 PRIORREC .45* (557) 10 PRIORVIO .25* (557) 11 RELEASE .17* (557)

2

3

4

5

6

1.00 (557) -.03 (557) -.05 (557) -.17* (557) .01 (557) -.08* (557) .18* (557) .25* (557) .07* (557) .14* (557)

1.00 (557) .23* (557) -.09* (557) .23* (557) .38* (557) -.12* (557) .12* (557) .07* (557) -.11* (557)

1.00 (557) .08* (557) .03 (557) .28* (557) -.01 (557) -.03 (557) .02 (557) .08* (557)

1.00 (557) .03 (557) -.01 (557) .00 (557) -.12* (557) -.03 (557) -.07* (557)

1.00 (557) .08* (557) -.12* (557) .12* (557) .07 (557) -.06 (557)

Note. N’s in parentheses * p < .10 (two-tailed significance test) (Continued . . .)

110

Table 3 (continued) Variable 12 CONVICT

1 .23* (557) 13 TARCON .11* (378) 14 INCARCE .34* (378) 15 INLENGTH .47* (280) 16 PROCESS .54* (557) 17 PREARR .36* (224) 18 PREVIOL .33* (224) 19 RISKTIME -.04 (494) 20 POSTARR .07 (494) 21 POSTVIOL .08* (494) 22 TIMEARR -.11 (152)

2 -.04 (557) -.05 (378) .14* (378) .11* (280) .21* (557) .03 (224) .06 (224) .04 (494) -.09* (494) -.02 (494) -.09 (152)

3 -.13* (557) .10* (378) .16* (378) .14* (280) -.06 (557) .03 (224) .04 (224) .02 (494) .07 (494) .10* (494) .14* (152)

4 -.21* (557) .05 (378) .04 (378) .11* (280) -.02 (557) -.05 (224) .00 (224) .11* (494) .06 (494) .08* (494) .00 (152)

5 -.09* (557) .01 (378) -.05 (378) -.03 (280) -.12* (557) -.02 (224) -.06 (224) -.04 (494) .12* (494) .03 (494) .10 (152)

6 -.09* (557) .07 (378) .13* (378) .03 (280) -.03 (557) -.07 (224) -.02 (224) -.07 (494) .01 (494) .10* (494) .04 (152)

Note. N’s in parentheses. * p < .10 (two-tailed significance test) (Continued . . .)

111

Table 3 (continued) Variable 7 WEAPON

7 1.00 (557) 8 AGEFIRST .00 (557) 9 PRIORREC -.05 (557) 10 PRIORVIO .00 (557) 11 RELEASE -.14* (557) 12 CONVICT -.12* (557) 13 TARCON .17* (378) 14 INCARCE .13* (378) 15 INLENGTH .20* (280)

8

9

10

1.00 (557) -.60* (557) -.31* (557) .03 (557) -.06 (557) -.05 (378) -.17* (378) -.16* (280)

1.00 (557) .45* (557) .02 (557) .09* (557) .03 (378) .25* (378) .20* (280)

1.00 (557) .04 (557) .07 (557) .13* (378) .16* (378) .22* (280)

11

12

1.00 (557) -.13* (557) -.06 (378) -.08* (378) .05 (280)

1.00 (557) . (378) . (378) . (280)

Note. N’s in parentheses. “ . ” is printed if coefficient could not be computed. * p < .10 (two-tailed significance test) (Continued . . .)

112

Table 3 (continued) Variable 16 PROCESS 17 PREARR 18 PREVIOL 19 RISKTIME 20 POSTARR 21 POSTVIOL 22 TIMEARR

7 -.13* (557) .05 (224) .06 (224) -.07 (494) .03 (494) .04 (494) .09 (152)

8 .03 (557) -.07 (224) -.12* (224) .05 (494) -.20* (494) -.20* (494) .05 (152)

9 .16* (557) .17* (224) .22* (224) -.04 (494) .15* (494) .20* (494) -.11 (152)

10 .07* (557) .15* (224) .18* (224) -.03 (494) .04 (494) .06 (494) .02 (152)

11 .40* (557) . (224) . (224) -.03 (494) -.01 (494) -.02 (494) -.15* (152)

12 .15* (557) .20* (224) .13* (224) -.02 (494) -.01 (494) -.01 (494) -.28* (152)

Note. N’s in parentheses. “ . ” is printed if coefficient could not be computed. * p < .10 (two-tailed significance test) (Continued . . .)

113

Table 3 (continued) Variable 13 TARCON

13 1.00 (378) 14 INCARCE .32* (378) 15 INLENGTH .28* (280) 16 PROCESS .05 (378) 17 PREARR .03 (136) 18 PREVIOL .03 (136) 19 RISKTIME -.14* (315) 20 POSTARR -.10* (315) 21 POSTVIOL -.05 (315) 22 TIMEARR .08 (96)

14

15

16

17

18

1.00 (378) . (280) .18* (378) .25* (136) .17* (136) -.03 (315) -.10* (315) .05 (315) -.01 (96)

1.00 (280) .26* (280) .19* (94) .15 (94) -.20* (217) -.07 (217) -.05 (217) .16 (59)

1.00 (557) .31* (224) .21* (224) -.03 (494) -.01 (494) -.05 (494) -.08 (152)

1.00 (224) .69* (224) -.17* (197) .11 (197) .13* (197) -.05 (60)

1.00 (224) -.18* (197) .11 (197) .13* (197) .05 (60)

Note. N’s in parentheses. “ . ” is printed if coefficient could not be computed. * p < .10 (two-tailed significance test) (Continued . . .)

114

Quantitative Findings

115

Table 3 (continued) Variable 19 RISKTIME 20 POSTARR 21 POSTVIOL 22 TIMEARR

19 1.00 (494) .08* (494) .01 (494) -.07 (152)

20

21

22

1.00 (494) .68* (494) . (152)

1.00 (494) -.11 (152)

1.00 (152)

Note. N’s in parentheses. “ . ” is printed if coefficient could not be computed. * p < .10 (two-tailed significance test)

Through an examination of the bivariate correlations between the key independent variable in this study (i.e., transfer to adult court) and the 11 dependent variables, it can be seen that almost all of the 67 coefficients were significant at the .10 level. The direction of each significant coefficient suggests that, as compared to youths retained in juvenile court, those transferred to adult court were more likely to be released from custody prior to disposition and more likely to be convicted. Of those youths who were convicted, transferred juveniles were more likely to be convicted on a target offense (i.e., robbery or aggravated assault) and more likely to be incarcerated. Of those youths who were incarcerated, transferred juveniles experienced longer periods of confinement. Youths in adult criminal court also experienced longer case processing times.

67

Because this study examines a total population of violent offenders (rather than a random sample), an argument could be made that significance tests were not required. However, for purposes of identifying "significant" effects of explanatory variables on dependent variables, we employed a .10 level of significance. The reader may choose to go above or below this level in interpreting the results.

116

Excluding Violent Youths From Juvenile Court

Concerning recidivism, the direction of each significant coefficient suggests that, of those youths released from custody prior to disposition, transferred juveniles were more likely to be rearrested and more likely to be rearrested on a violent felony offense during the predispositional time period. In addition, of those youths “on the street” during the post-dispositional period, transferred juveniles were more likely to be rearrested on a violent felony offense. Finally, with regard to the insignificant correlations, it should be noted that transfer had a positive and nearly significant bivariate association with postdispositional arrest (r = .0681; p = .131) and, among those youths rearrested, a negative association with time to post-dispositional arrest (r = -.1112; p = .173). In terms of the bivariate correlations between the control variables and the dependent variables, numerous significant associations exist. Age at referral was positively associated with release from custody prior to disposition, incarceration (among those youths convicted), incarceration length (among those youths incarcerated), and case processing time. Age at referral was negatively and significantly associated with post-dispositional arrest. Concerning race, as compared to whites, nonwhites were significantly less likely to be released from custody prior to disposition and less likely to be convicted, but among those youths convicted, nonwhites were more likely to be convicted of a target offense and more likely to be incarcerated. Among those youths incarcerated, nonwhites experienced longer periods of incarceration. Finally, nonwhites were significantly more likely to be rearrested on a violent felony offense following disposition, but of those rearrested, nonwhites experienced a longer time to failure. In terms of county of jurisdiction, urban youths were significantly more likely to be released from custody prior to disposition than were suburban and rural youths, but less likely to be convicted. Of those juveniles who were incarcerated, urban youths experienced significantly longer periods of confinement. Lastly, urban juveniles were more likely to be arrested on a violent felony offense during the post-dispositional time period. School status, somewhat surprisingly, was found to have a significant and negative association with release from custody prior to disposition. Those youths who were enrolled, graduated, or had obtained a GED at the time of referral were less likely to be released,

Quantitative Findings

117

but they were also less likely to be convicted. Additionally, school status was significantly and negatively correlated with case processing time. A final interesting finding was the positive association between school status and arrest following disposition, which might be explained based on a process of weakened social bonds (i.e., removal from school) following initial arrest and court processing. With regard to family status, only three bivariate coefficients were significant. Youths with arrangements consisting of something other than living with two parents were less likely to be convicted, but if convicted, they were more likely to be incarcerated. Also, youths from “nontraditional” homes were significantly more likely to be arrested for a violent felony offense during the post-dispositional time period. Concerning weapon type, an examination of the coefficients suggests that firearm use during the offense was significantly associated with all six of the dependent variables that measured case outcomes, but none of the five dependent variables that measured recidivism. Offenders who employed a firearm, as compared to those who used another type of deadly weapon, were less likely to be released from custody prior to disposition and less likely to be convicted. However, of those who were convicted, firearm users were more likely to be convicted of a target offense and more likely to be incarcerated. Of those who were incarcerated, firearm users received longer sentences. Finally, youths who employed a firearm experienced shorter case processing times than did those who used some other deadly weapon. With regard to the variables that account for a youth’s prior offending history, age at first referral was found to have a significant and negative correlation with incarceration (among those youths who were convicted) and incarceration length (among those youths who were incarcerated). Additionally, among those offenders who were released from custody prior to disposition, age at first referral was found to have a significant and negative association with predispositional violent felony arrest. Finally, among those youths back on the street following disposition, age at first referral had a significant, negative correlation with both post-dispositional arrest and violent felony arrest. The coefficients pertaining to the prior record variable and each of the dependent variables indicate numerous significant, bivariate relationships. Prior record had a positive association with conviction, and, among those convicted, a positive association with incarceration.

118

Excluding Violent Youths From Juvenile Court

Of the youths who were incarcerated, those with more extensive prior records received longer sentences. Prior record was also positively correlated with case processing time. Finally, there was a positive and significant association between prior record and all of the dependent variables that measured recidivism, except for post-dispositional time to rearrest. Similar bivariate correlations were revealed between the prior violent felony variable and the dependent variables. Of the youths who were convicted, those who were previously adjudicated delinquent on a violent felony offense were significantly more likely to be convicted of a target offense and also were more likely to be incarcerated. Among those who were incarcerated, having a prior violent felony adjudication had a significant and positive association with incarceration length. Those youths with a prior adjudication on a violent felony offense also experienced longer case processing times. Lastly, among those youths released from secure custody prior to disposition, the prior substantiated violent felony variable was significantly and positively correlated with both pre-dispositional arrest and violent felony arrest. However, having a prior violent felony adjudication was not found to be significantly associated with any of the post-dispositional recidivism variables. In the pre- and post-dispositional analyses of recidivism, two separate explanatory variables were employed to account for differences in time on the street among the released offenders. Concerning the bivariate correlations, of those youths released from custody prior to disposition, case-processing time was found to have a significant and positive association with both pre-dispositional arrest and violent felony arrest. In the post-dispositional time period, time at risk was significantly and positively correlated with arrest, but not with violent felony arrest. In post-dispositional analyses of recidivism, four case outcome variables were also used as explanatory variables: release, conviction, incarceration length, and case processing time. With regard to the bivariate correlations, among those youths who were rearrested following disposition, release from custody prior to disposition had a significant, negative association with time to arrest. Similarly, of the rearrested offenders, conviction also had a significant and negative correlation with time to arrest. Case processing time was not found to

Quantitative Findings

119

be significantly correlated with any of the post-dispositional recidivism variables. For post-dispositional analyses of recidivism, the incarceration length variable was recoded, in order to consider the entire 494 offenders who were on the street following disposition. This variable was measured continuously (in months), and for offenders who were either not convicted or received sentences that did not involve incarceration, incarceration length was set at zero. Separate bivariate correlations were then computed between this variable and each of the post-dispositional recidivism variables. The coefficients revealed a significant, negative association between incarceration length and arrest following disposition (r = -.105; p = .019). However, significant correlations were not found between incarceration length and violent felony arrest following disposition (r = -.023; p = .613), or between incarceration length and time to post-dispositional rearrest (r = -.112; p = .170).

Multivariate Results While the zero-order correlations discussed above are interesting and suggestive, it is again stressed that they do not control for the effects of other explanatory variables. Therefore, they should not be used to make judgments about causality. Instead, multivariate techniques are required, in order to examine the effects of the explanatory variables when other measured determinants are taken into account. To test the 11 hypotheses presented in Chapter 5, a combination of logistic regression, ordinary least squares (OLS) regression, and survival analysis was employed. For each dependent variable, four models were estimated: a bivariate transfer model, a full model, and two separate “transfer specific” models for the transferred and nontransferred youths. The full model was used to estimate the main, additive effects of the explanatory variables, while the transfer specific models were utilized to determine whether, and to what extent, transfer to adult court interacted with the other explanatory variables to influence outcomes.

120

Excluding Violent Youths From Juvenile Court

Hypothesis 1: Release From Pre-Dispositional Custody In testing Hypothesis 1, we sought to determine whether there was a significant difference in the likelihood of release from predispositional secure custody for violent youths in juvenile versus adult court. The logistic regression estimates for the determinants of release 68 are presented in Table 4. Here, we again note the significant, positive bivariate effect of transfer to adult court (b = .8085; p < .01), which suggests that transferred youths were more likely to be released back into the community prior to disposition. More importantly, the transfer effect remained significant in the full model (b = .8095; p < .01), when the effects of the control variables were taken into account. As discussed previously, we may also use the results of the full model to predict the probability of release when transfer is equal to 0 (i.e., not transferred to adult court) and also when transfer is equal to 1 (i.e., transferred to adult court), with all other variables set at their mean. Using Equation 2 (presented in Chapter 5), the estimated probability of release for youths retained in juvenile court was .3484, while the estimated probability of release for transferred offenders was .5457. In other words, the probability of release for a transferred offender was .1973 (almost 20 percentage points) higher than the probability for a retained youth. In addition to the positive transfer effect, the coefficients for age at referral (b = .2356; p < .10), county of jurisdiction (b = 1.1002; p < .01), and weapon type (b = -.8089; p < .01) were also significant. It appears that, not only were transferred offenders more likely to be released from secure custody prior to disposition, but so were older youths and those from urban counties. On the other hand, juveniles who used a firearm during the commission of their offense were less

68 The probability values (p-values) employed in Table 4 and subsequent tables were based on two-tailed significance tests. It should be remembered from Chapter 5 that Ha (4), Ha (5), and Ha (6) indicated a one-tailed test for the effect of transfer to adult court on incarceration, incarceration length, and case processing time, respectively. However, based on the results obtained through testing these hypotheses, using a one- or two-tailed test was immaterial. The two-tailed p-values associated with the "transfer effect" in the full models were all less than .00005.

Table 4: Logistic Regression Estimates for the Determinants of Release (N=557) Variable Bivariate B (SE) |Wald| Multiple B (SE) TRANSFER .808 (.199) 16.450*** .809 (.246) AGEREF .236 (.127) RACE -.424 (.260) COUNTY 1.100 (.307) SCHOOL -.315 (.204) FAMILY -.154 (.215) WEAPON -.809 (.299) AGEFIRST -.002 (.074) PRIORREC -.052 (.037) PRIORVIO .202 (.290) Constant -.605 (.102) 35.021*** -3.833 (1.94) Log734.098 701.412 likelihood

|Wald| 10.801*** 3.435* 2.661 12.805*** 2.393 .511 7.296*** .001 2.028 .484 3.894**

Variable Nontransfer B (SE)|Wald| Transfer B (SE) |Wald| |Z-test| AGEREF .188 (.148) 1.629 .350 (.305) 1.314 .48 RACE -.716 (.304) 5.549** .172 (.527) .106 1.46 COUNTY .947 (.390) 5.904** 1.493 (.509) 8.586*** .85 SCHOOL -.499 (.243) 4.219** .084 (.384) .048 1.28 FAMILY -.264 (.249) 1.121 .375 (.447) .703 1.25 WEAPON -.785 (.384) 4.176** -.918 (.514) 3.196* .21 AGEFIRST .072 (.094) .590 -.109 (.133) .672 1.11 PRIORREC .023 (.055) .172 -.120 (.058) 4.359** 1.79* PRIORVIO -.163 (.410) .158 .384 (.452) .721 .90 Constant -3.759 (2.16) 3.019* -4.310 (5.09) .747 Log517.094 172.839 likelihood

Note. In this table and subsequent tables, the absolute values for all Wald statistics, z-scores, and t-scores are presented. * p < .10 ** p < .05 *** p < .01 121

122

Excluding Violent Youths From Juvenile Court

likely to be released. It should also be noted that, although the coefficient for race does not quite reach the level of significance employed in this study (b = -.4241, p = .1029), there is at least suggestive evidence that nonwhite youths were less likely to be released from custody prior to disposition. Finally, it is interesting that none of the prior offending variables (i.e., age at first offense, prior record, and prior violent felony) were found to have a direct effect, which suggests that previous delinquent behavior did not have a major impact on decisions related to pre-dispositional secure custody. A possible explanation is that prior offending information was not always available to, or not utilized by, the judges who made release decisions. Or, it could be that judges just focused more on the seriousness of the offense (e.g., whether a firearm was involved) in making their decisions (e.g., 69 whether to release a juvenile to his parents, how high to set bail, etc.). We now turn to the question of whether the effects of the explanatory variables were the same for transferred and nontransferred youths. The answer lies in an examination of the transfer specific models. The model pertaining to the nontransferred offenders indicates that, among youths retained in juvenile court, nonwhites (b = -.7157; p < .05), those in school, graduated, or having a GED (b = -.4986; p < .05), and those employing a firearm (b = -.7851; p < .05) were less likely to be released, while youths in urban counties (b = .9475; p < .05) were more likely to be released. Of the juveniles transferred to adult court, those from urban counties (b = 1.4928; p < .01) were more likely to be released, but those with a more extensive prior record (b = -.1205; p < .05) and those employing a firearm (b = -.9183; p < .10) were less likely to be released. In order to judge if any of the explanatory variables exerted a greater effect in one group over the other, the z-tests of the slope coefficients should be examined. These z-scores are a product of Equation 3 (presented in Chapter 5). Here, it can be seen that only one interaction effect was significant at the .10 level, as the effect of prior record was different for transferred and nontransferred youths. The z-test for the slope comparison (z = 1.79) suggests that prior record was 69

Both of these explanations were revealed through the process of interviewing justice system officials. Other results from these interviews are presented in Chapter 7.

Quantitative Findings

123

given more attention in adult court. Youths in adult court with more extensive prior records were less likely to be released than were offenders in juvenile court with similarly extensive prior records. In fact, among the youths retained in juvenile court, the coefficient for prior record was positive. Hypothesis 2: Conviction Hypothesis 2 pertained to the likelihood of conviction for violent youths in juvenile versus adult court. The logistic regression estimates for the determinants of conviction may be found in Table 5. The bivariate model suggests that transferred juveniles were more likely to be convicted than were youths retained in juvenile court (b = 1.4245; p < .01). Furthermore, the effect of transfer on conviction remained significant in the full model (b = 1.4777; p < .01). Again, we can use the results of the full model to predict the probability of conviction for transferred and nontransferred youths, with all other explanatory variables set at their mean. Here, the estimated probability of conviction for transferred offenders was .8856, while the estimated probability for retained youths was .6384. The probability of conviction for a transferred offender was .2472 (nearly 25 percentage points) higher than the probability for a youth retained in juvenile court. Along with the positive effect of transfer, age at referral (b = -.3976; p < .01), county of jurisdiction (b = -1.5996; p < .01), and family status (b = -.4406; p < .10) were found to have significant, negative effects on conviction. While transferred offenders were more likely to be convicted, older youths, those from urban counties, and those not living with two parents were less likely to be convicted. It is again noteworthy that none of the prior offending variables was found to be significant, and weapon type appeared to have little or no direct effect on conviction. Although these factors did not appear to directly influence the likelihood of conviction, they may have had a greater impact on the likelihood of conviction on a target offense, which will be discussed shortly. Regarding the transfer specific models, older youths and those from an urban area were significantly less likely to be convicted in

Table 5: Logistic Regression Estimates for the Determinants of Conviction (N=557) Variable Bivariate B (SE) |Wald| Multiple B (SE) TRANSFER 1.424 (.272) 27.450*** 1.478 (.319) AGEREF -.398 (.135) RACE -.468 (.304) COUNTY -1.600 (.451) SCHOOL -.354 (.228) FAMILY -.441 (.245) WEAPON -.091 (.362) AGEFIRST .026 (.079) PRIORREC .018 (.042) PRIORVIO .105 (.326) Constant .472 (.100) 22.044*** 8.966 (2.13) Log665.067 622.038 likelihood Variable Nontransfer B (SE) |Wald| Transfer B (SE) AGEREF -.328 (.143) 5.208** -1.101 (.533) RACE -.294 (.328) .804 -1.456 (.950) COUNTY -1.508 (.501) 9.077*** -2.053 (1.11) SCHOOL -.162 (.247) .429 -1.476 (.712) FAMILY -.335 (.259) 1.670 -1.800 (1.08) WEAPON -.756 (.468) 2.603 .924 (.697) AGEFIRST .010 (.089) .012 .057 (.192) PRIORREC .033 (.053) .397 -.006 (.071) PRIORVIO -.147 (.386) .146 .716 (.690) Constant 8.235 (2.22) 13.67* 24.350 (9.85) Log524.000 81.622 likelihood

* p < .10 ** p < .05 *** p < .01

124

|Wald| 21.490*** 8.709*** 2.375 12.559*** 2.415 3.266* .063 .112 .180 .104 17.784***

|Wald| |Z-test| 4.273** 1.40 2.347 1.16 3.435* .45 4.294** 1.74* 2.767* 1.32 1.757 2.00** .088 .22 .007 .45 1.078 1.09 6.111**

Quantitative Findings

125

either juvenile or adult court. Furthermore, among the offenders in adult court, those who were enrolled in school, graduated, or had obtained a GED were also less likely to be convicted (b = -1.4761; p < .05), as were those who were not living with two parents (b = -1.8003; p < .10). When the z-tests for the slope coefficients were conducted, two interaction effects were revealed. Here, transferred offenders who were enrolled in school, graduated, or had obtained a GED were less likely to be convicted than were retained youths who had the same school status (z = 1.74). Also, although weapon type was not previously found to have a direct effect on conviction, use of a firearm appeared to have a different influence in the two separate court systems (z = 2.00). Offenders in adult court who had employed a firearm were more likely to be convicted than were youths in juvenile court who had used a gun. Interestingly, among the offenders retained in juvenile court, use of a firearm had a negative and nearly significant (b = -.7557; p = .1067) effect on conviction. Hypothesis 3: Conviction on a Target Offense Hypothesis 3 considered the likelihood of conviction on a target offense (i.e., robbery or aggravated assault) among those youths who were convicted in either juvenile or adult court. The logistic regression estimates are listed in Table 6. The bivariate model again reveals the significant and positive relationship between transfer and conviction on a target offense (b = .6252; p < .05). Moreover, the positive transfer effect became even more significant in the full model (b = 1.0816; p < .01). Of the convicted youths, those in adult court were more likely to be convicted of robbery or aggravated assault, which suggests a lesser amount of charge reduction in adult court. By using our equation to predict the probability of target conviction among the convicted youths, we can again gain a clearer picture of the impact of transfer to adult court. With all other explanatory variables set at their mean, the estimated probability of target conviction for transferred offenders was .8965. In contrast, the estimated probability of target conviction for retained youths was .7459. In other words, among the convicted offenders, the probability of conviction on a robbery or aggravated assault charge was .1506

Table 6: Logistic Regression Estimates for the Determinants of Target Conviction (N=378) Variable Bivariate B (SE) |Wald| Multiple B (SE) TRANSFER .625 (.293) 4.539** 1.082 (.379) AGEREF -.318 (.195) RACE .213 (.343) COUNTY -.010 (.339) SCHOOL .144 (.228) FAMILY .194 (.293) WEAPON .970 (.359) AGEFIRST -.001 (.104) PRIORREC -.055 (.046) PRIORVIO 1.017 (.500) Constant 1.109 (.144) 59.184*** 5.044 (3.06) Log390.511 368.418 likelihood Variable Nontransfer B (SE) |Wald| Transfer B (SE) AGEREF -.101 (.224) .205 -1.673 (.905) RACE .060 (.405) .022 .805 (.785) COUNTY .696 (.390) 3.172* -2.128 (.968) SCHOOL .046 (.342) .018 .051 (.640) FAMILY .045 (.343) .018 .512 (.670) WEAPON .266 (.453) .345 2.137 (.716) AGEFIRST -.087 (.141) .382 .014 (.209) PRIORREC -.077 (.076) 1.019 -.063 (.082) PRIORVIO 1.083 (.694) 2.439 1.475 (.912) Constant 3.223 (3.28) .964 29.264 (16.1) 279.786 67.612 Loglikelihood

* p < .10 ** p < .05 *** p < .01

126

|Wald| 8.142*** 2.652 .385 .001 .250 .437 7.307*** .000 1.455 4.133** 2.713*

|Wald| |Z-test| 3.416* 1.69* 1.052 .84 4.832** 2.70*** .006 .01 .584 .62 8.921*** 2.21** .004 .40 .600 .12 2.612 .34 3.319*

Quantitative Findings

127

(15 percentage points) higher for transferred offenders than for those retained in juvenile court. Two other significant effects were also revealed in the full model. The positive coefficient for weapon type (b = .9700; p < .01) shows that, among convicted youths, those who had employed a firearm during the commission of their offense were more likely to be convicted of robbery or aggravated assault. Similarly, the positive coefficient for prior violent felony (b = 1.0167; p < .05) indicates that, among convicted youths, those with a prior substantiated violent felony were more likely to be convicted of robbery or aggravated assault. Finally, although the negative effect of age at referral did not quite reach the .10 level of significance (b = -.3177; p = .1034), there is again suggestive evidence that older convicted youths were less likely to be convicted of robbery or aggravated assault than were their younger counterparts. Concerning the transfer specific models, three significant differences were found. The first pertained to county of jurisdiction. Of the youths retained and convicted in juvenile court, those from urban counties were more likely to be convicted of robbery or aggravated assault (b = .6956; p < .10). In contrast, of the youths transferred and convicted in adult court, those from urban counties were less likely to be convicted of robbery or aggravated assault (b = -2.1277; p <.05). The z-test for the slope comparison (z = 2.70) confirmed that transfer interacted with county of jurisdiction. Among the convicted youths, those in adult, urban courts were less likely to be convicted of a target offense than were those in juvenile, urban courts. In addition to county of jurisdiction, significant interaction effects were also found between transfer and age at referral and between transfer and weapon type. For youths retained and convicted in juvenile court, age at referral did not have an impact on the likelihood of conviction on a target offense. However, among the offenders transferred and convicted in adult court, the effect of age at referral was negative and significant (b = -1.6731; p < .10). The z-test for the slope comparison (z = 1.69) revealed that older youths convicted in adult court were less likely to be convicted of a target offense than were older youths convicted in juvenile court. Similarly, for the offenders retained and convicted in juvenile court, weapon type did not have an effect on conviction on a target offense. However, among the juveniles transferred and convicted in adult court, the effect of weapon type was positive and significant

128

Excluding Violent Youths From Juvenile Court

(b = 2.1374; p < .01). The slope comparison z-test (z = 2.21) verified that convicted offenders in adult court who had employed a firearm were more likely to be convicted of a target offense than were convicted youths in juvenile court who had used a gun. Hypothesis 4: Incarceration In testing Hypothesis 4, we sought to determine if the offenders who were convicted in adult court had a greater likelihood of incarceration than did those who were convicted in juvenile court. The logistic regression estimates for the determinants of incarceration are presented in Table 7. Here again, a strong, positive, bivariate transfer effect may be noted (b = 2.5621; p < .01). Furthermore, the coefficient associated with transfer to adult court remained highly significant in the full model (b = 2.5104; p < .01). Not surprisingly, among the convicted offenders, the predicted probability of incarceration was much greater for transferred youths than for those retained in juvenile court. With all other explanatory variables set at their mean, the estimated probability of incarceration for youths in adult court was .9621. For the nontransferred offenders, the estimated probability of incarceration was .6736. The probability of incarceration was .2885 (nearly 29 percentage points) higher for transferred youths than for those retained in juvenile court. The only other significant effect found in the full model was the positive influence of weapon type (b = .8376; p < .05). Of the convicted offenders, youths in adult court and youths employing a firearm experienced a greater likelihood of incarceration than did those in juvenile court and those utilizing some other deadly weapon. With regard to the transfer specific models, among the youths retained and convicted in juvenile court, those who employed a firearm were significantly more likely to be incarcerated (b = .7900; p < .10). For the offenders transferred and convicted in adult court, the effect of using a firearm was positive, but insignificant. However, the z-test for the slope comparison (z = .16) did not indicate any differential impact of weapon type on incarceration in the two separate court systems. It appears that firearm use had a similar, positive effect on incarceration for youths who were convicted in either juvenile or adult court.

Table 7: Logistic Regression Estimates for the Determinants of Incarceration (N=378) Variable Bivariate B (SE) |Wald| Multiple B (SE) TRANSFER 2.562 (.475) 29.110*** 2.510 (.533) AGEREF -.012 (.198) RACE .486 (.342) COUNTY .076 (.363) SCHOOL .072 (.300) FAMILY .338 (.294) WEAPON .838 (.396) AGEFIRST -.074 (.125) PRIORREC .078 (.070) PRIORVIO .230 (.521) Constant .573 (.130) 19.551*** .249 (2.94) Log378.871 353.981 likelihood Variable Nontransfer B (SE) |Wald| Transfer B (SE) AGEREF .010 (.209) .002 -.962 (1.06) RACE .467 (.362) 1.661 .463 (1.21) COUNTY .205 (.392) .273 -.240 (1.26) SCHOOL -.030 (.321) .009 .610 (1.00) FAMILY .292 (.312) .874 .665 (.982) WEAPON .790 (.428) 3.405* .592 (1.16) AGEFIRST -.105 (.137) .589 -.014 (.352) PRIORREC .111 (.082) 1.825 .023 (.145) PRIORVIO .482 (.608) .628 -.809 (1.19) Constant .354 (3.07) .013 18.387 (19.4) Log310.419 37.867 likelihood

* p < .10 ** p < .05 *** p < .01

129

|Wald| 22.209*** .004 2.023 .044 .058 1.328 4.468** .349 1.218 .195 .007

|Wald| |Z-test| .821 .90 .147 .00 .036 .34 .371 .61 .459 .36 .260 .16 .001 .24 .023 .53 .461 .96 .901

130

Excluding Violent Youths From Juvenile Court

The other slope comparison z-tests did not suggest any significant interactions between transfer to adult court and the control variables. Hypothesis 5: Incarceration Length Hypothesis 5 pertained to the length of incarceration among those offenders who were both convicted and incarcerated. In testing this hypothesis, we sought to determine whether youths incarcerated by an adult court experienced longer periods of confinement than did those who were placed in a juvenile correctional facility by a juvenile court. The OLS regression estimates for the determinants of the natural log of incarceration length may be found in Table 8. Once again, the initial bivariate model revealed a significant, positive effect of transfer (b = .8651; p < .01). Additionally, transfer alone explained over 22% of the variation in the natural log of 2 incarceration length (R = .2241). In the full model, the transfer effect remained highly significant (b = 1.0378; p < .01), and the coefficient of determination was increased to .3487, meaning the model explained nearly 35% of the variation in the natural log of incarceration length. Based on the fact that we are now considering a continuous dependent variable, we can further assess the impact of transfer by predicting the natural log of incarceration length for transferred and nontransferred youths, with all other explanatory variables set at their mean. Through using Equation 1 (presented in Chapter 5), it was determined that the estimated natural log of incarceration length for youths confined by adult courts was 3.4127 (30.35 months). For the offenders retained and placed by juvenile courts, the estimated natural log of incarceration length was 2.3749 (10.75 months). The predicted natural log of incarceration length for a transferred youth was nearly 1½ times of that for a nontransferred offender. When transformed to months, the predicted period of confinement for a transferred, incarcerated offender was nearly 3 times as long as that of a retained and incarcerated youth. Among the other explanatory variables, two direct effects were found. Weapon type (b = .6756; p < .01) and prior violent felony (b = .2166; p < .10) both exerted a positive influence on the natural log of incarceration length. Not only did the incarcerated, transferred

Table 8: OLS Regression Estimates for the Determinants of the Natural Log of Incarceration Length (N=280) Variable Bivariate B (SE) |T| Multiple B (SE) TRANSFER .865 (.097) 8.960*** 1.038 (.109) AGEREF -.030 (.066) RACE .176 (.131) COUNTY .191 (.124) SCHOOL .110 (.098) FAMILY -.009 (.108) WEAPON .676 (.144) AGEFIRST -.030 (.034) PRIORREC -.018 (.015) PRIORVIO .217 (.123) Constant 2.449 (.062) 39.580*** 2.389 (1.04) R-square .2241 .3487

|T| 9.544*** .460 1.344 1.540 1.117 .084 4.675*** .901 1.202 1.756* 2.296**

Variable Nontransfer B (SE) |T| Transfer B (SE) |T| |Z-test| AGEREF -.051 (.072) .713 -.047 (.128) .368 .03 RACE .509 (.162) 3.148*** -.244 (.212) 1.151 2.82*** COUNTY .335 (.159) 2.107** .081 (.193) .417 1.02 SCHOOL .078 (.119) .656 .092 (.159) .580 .07 FAMILY -.102 (.124) .819 .048 (.187) .256 .67 WEAPON -.138 (.203) .681 1.244 (.206) 6.033*** 4.78*** AGEFIRST -.019 (.045) .428 -.051 (.051) .997 .47 PRIORREC -.010 (.024) .402 -.019 (.019) .987 .29 PRIORVIO .075 (.165) .454 .356 (.182) 1.951* 1.14 Constant 2.995 (1.09) 2.743*** 3.896 (2.22) 1.759* R-square .1216 .3242

* p < .10 ** p < .05 *** p < .01

131

132

Excluding Violent Youths From Juvenile Court

youths experience longer sentences than did their juvenile counterparts, youths who had employed a firearm and those with a prior substantiated violent felony were confined longer than offenders who had used another deadly weapon and those who did not have a previous violent felony adjudication. Several interesting findings were revealed in the transfer specific models. To begin, among the offenders retained and incarcerated by a juvenile court, nonwhites experienced longer periods of confinement than did white youths (b = .5086; p < .01). In contrast, the coefficient for race in the transfer model was negative and insignificant. The z-test for the slope comparison (z = 2.82) confirmed that being nonwhite had a significantly greater, positive impact on the natural log of incarceration length for youths placed by juvenile courts as compared to those incarcerated by adult courts. Concerning county of jurisdiction, urban youths incarcerated by a juvenile court also experienced significantly longer periods of confinement than did those placed by rural and suburban juvenile courts (b = .3351; p < .05). For the offenders transferred and incarcerated by an adult court, the effect of county of jurisdiction was positive, but insignificant. However, the slope comparison z-test (z = 1.02) did not verify a differential impact of county of jurisdiction between the two systems. Still, it seems likely that county of jurisdiction may have had some positive influence on incarceration length, particularly in juvenile courts. This is supported by the fact that in the original full model, the coefficient for county of jurisdiction was positive and nearly reached the .10 level (b = .1913; p = .1248), suggesting incarcerated urban youths received longer sentences than did their rural and suburban counterparts. Among the incarcerated, transferred offenders, weapon type was found to have a positive and highly significant effect on the natural log of incarceration length (b = 1.2437; p < .01), indicating that youths in adult court who had employed a firearm received longer sentences than did those who used another deadly weapon. In contrast, the coefficient for weapon type among youths incarcerated by a juvenile court was negative and insignificant. The slope comparison z-test (z = 4.78) confirmed that firearm use had a significantly greater, positive effect on the natural log of incarceration length for offenders incarcerated by adult courts as compared to those incarcerated by juvenile courts.

Quantitative Findings

133

Finally, among the transferred and incarcerated offenders, those who had a prior substantiated violent felony received significantly longer sentences than did those without a prior violent felony adjudication (b = .3558; p < .10). Among the youths retained and placed in a correctional facility by a juvenile court, the coefficient for prior violent felony was positive, but insignificant. The z-test for the slope comparison (1.14) did not reveal a significant interaction effect. This suggests that having a prior substantiated violent felony offense had a similar, positive effect on incarceration length for youths incarcerated by either court system. Hypothesis 6: Case Processing Time Hypothesis 6 concerned case processing time, and whether violent youths in adult court experienced longer processing times than did those in juvenile court. The OLS regression estimates are presented in Table 9. Here, we again see the strong, positive, bivariate effect of transfer (b = 1.4854; p < .01). Furthermore, transfer alone explained over 29% of the variation in the natural log of case processing time 2 (R = .2930). Not surprisingly, the effect of transfer remained highly significant in the full model (b = 1.5909; p < .01), showing that offenders processed in adult court experienced longer periods of case processing than did those who were handled in juvenile court. Lastly, in the full model, the coefficient of determination increased slightly to .3231, meaning that the model explained over 32% of the variation in the natural log of case processing time. As with the natural log of incarceration length, we can also predict the natural log of case processing time for transferred and nontransferred offenders, with all other explanatory variables set at their mean. For youths in adult court, the estimated natural log of case processing time was 5.4019 (221.83 days). For the offenders retained in juvenile court, the estimated natural log of case processing time was 3.811 (45.20 days). The predicted natural log of case processing time for a transferred youth was again nearly 1½ times of that for a nontransferred offender. When translated to days, the predicted period of case processing for a transferred offender was nearly 5 times as long as that of a youth retained in juvenile court. Two other significant effects were revealed in the full model. With regard to county of jurisdiction, youths from urban counties

Table 9: OLS Regression Estimates for the Determinants of the Natural Log of Case Processing Time (N=577) Variable Bivariate B (SE) |T| Multiple B (SE) TRANSFER 1.485 (.098) 15.167*** 1.591 (.115) AGEREF -.014 (.058) RACE -.062 (.121) COUNTY .260 (.129) SCHOOL -.148 (.095) FAMILY -.079 (.101) WEAPON -.115 (.136) AGEFIRST .088 (.034) PRIORREC -.005 (.016) PRIORVIO -.082 (.133) Constant 3.838 (.049) 78.732*** 2.862 (.897) R-square .2930 .3231

|T| 13.866*** .241 .517 2.012** 1.563 .790 .841 2.606*** .284 .616 3.189***

Variable Nontransfer B (SE) |T| Transfer B (SE) |T| |Z-test| AGEREF -.092 (.067) 1.378 .280 (.125) 2.232** 2.62*** RACE -.103 (.143) .722 .006 (.217) .029 .42 COUNTY .130 (.164) .792 .696 (.202) 3.446*** 2.18** SCHOOL -.169 (.114) 1.487 -.022 (.162) .135 .74 FAMILY -.118 (.116) 1.016 .206 (.192) 1.071 1.44 WEAPON -.329 (.178) 1.847* .262 (.202) 1.298 2.19** AGEFIRST .159 (.042) 3.756*** -.032 (.053) .593 2.80*** PRIORREC .038 (.025) 1.527 -.047 (.020) 2.316** 2.64*** PRIORVIO .041 (.181) .225 -.410 (.184) 2.226** 1.74* Constant 3.335 (.987) 3.377*** .617 (2.16) .285 R-square .0603 .1978

* p < .10 ** p < .05 *** p < .01

134

Quantitative Findings

135

experienced longer periods of case processing than did those from rural and suburban counties (b = .2596; p < .05). Additionally, age at first referral was found to have a positive and significant effect (b = .0878; p < .01), indicating that youths with an older age at first referral to juvenile court experienced longer case processing times. A possible explanation for this finding is that youths with an older age at first referral may be seen as less serious offenders, causing their cases to be “put on the backburner.” This is supported by the fact that while the coefficients for prior record and prior violent felony were insignificant, they were both negative, suggesting that youths with more serious and extensive histories of offending experienced shorter periods of case processing. In comparing the transfer specific models, numerous significant differences were found for youths in juvenile and adult court. To begin, among the offenders retained in juvenile court, those who had employed a firearm experienced shorter periods of case processing (b = -.3290; p < .10). For juveniles transferred to adult court, the effect of gun use was positive and insignificant. The slope comparison z-test (z = 2.19) confirmed that use of a gun had a significantly greater, negative impact on the natural log of case processing time for offenders retained in juvenile court as compared to those transferred to adult court. Next, among the youths retained in juvenile court, those with an older age at first referral experienced longer periods of case processing (b = .1589; p < .01). In contrast, among the transferred juveniles, age at first referral had a negative and insignificant effect. Again, the slope comparison z-test (z = 2.80) indicated a significant interaction, meaning age at first referral had a greater, positive effect on the natural log of case processing time for youths retained in juvenile court as compared to those waived to adult court. The previous finding suggests that age at first offense was given more attention in juvenile court over adult court. The opposite appears to be true for the other two prior offending variables. Among the offenders transferred to adult court, those with more extensive prior records (b = -.0469; p < .05) and those with a prior substantiated violent felony (b = -.4102; p < .05) experienced significantly shorter periods of case processing. For the youths retained in juvenile court, the coefficients for prior record and prior violent felony were positive and insignificant. The slope comparison z-tests for prior record (z = 2.64)

136

Excluding Violent Youths From Juvenile Court

and for prior violent felony (z = 1.74) both verified that these two variables had a greater, negative impact on the natural log of case processing time for juveniles in adult court as compared to those in juvenile court. Two other significant interaction effects were also revealed. Among the transferred offenders, those with an older age at referral experienced lengthier case processing times (b = .2800; p < .05), suggesting that younger juveniles were given more immediate attention. For youths retained in juvenile court, the effect of age at referral was negative and insignificant. The z-test for the slope comparison (z = 2.62) confirmed that age at referral had a stronger, positive effect on the natural log of case processing time for offenders in adult court as compared to those in juvenile court. It seems likely that in juvenile court, older youths may be viewed as “almost adult,” which contributes to a more rapid case disposition. In adult court, however, older juveniles may be viewed in essentially the same way as adults, leading to a more drawn out process. Finally, among the transferred offenders, urban youths experienced longer case processing times than did those from rural and suburban counties (b = .6956; p < .01). For the retained juveniles, the effect of county of jurisdiction was also positive, but insignificant. While the previously discussed full model showed county of jurisdiction to have a direct, positive effect on case processing time, the slope comparison z-test (z = 2.18) revealed that this effect was significantly greater in adult court. Hypothesis 7: Recidivism Prior To Final Disposition While Hypotheses 1 through 6 focused on the nature of punishment imposed on violent offenders in the juvenile and adult court systems, Hypotheses 7 through 11 pertained to the future criminal behavior of these youths. Hypothesis 7 specifically addressed the likelihood of rearrest prior to final disposition, among the 224 offenders who were released from secure custody during this time period. The logistic regression estimates are presented in Table 10. The bivariate coefficient shows a significant, positive effect of transfer (b = 1.6981; p < .01), or that, among the offenders released from secure custody prior to disposition, those who were transferred to adult court were more likely to be rearrested before their original

Table 10: Logistic Regression Estimates for the Determinants of Pre-Dispositional Arrest (N=224) Variable Bivariate B (SE) |Wald| Multiple B (SE) |Wald| TRANSFER 1.698 (.333) 26.030*** 1.471 (.454) 10.524*** AGEREF -.508 (.268) 3.590* RACE .115 (.465) .062 COUNTY -.517 (.546) .897 SCHOOL .137 (.386) .127 FAMILY -.696 (.406) 2.932* WEAPON .769 (.509) 2.288 AGEFIRST .125 (.155) .655 PRIORREC .082 (.070) 1.365 PRIORVIO .298 (.514) .336 PROCESS .575 (.230) 6.229** Constant -1.856 (.240) 59.604*** 1.631 (3.99) .167 Log222.110 203.270 likelihood Variable Nontransfer B (SE) |Wald| Transfer B (SE) AGEREF -.662 (.391) 2.870* -.375 (.483) RACE -.207 (.645) .103 .290 (.818) COUNTY .253 (.895) .080 -2.076 (1.01) SCHOOL .047 (.568) .007 .774 (.635) FAMILY .174 (.597) .084 -2.090 (.807) WEAPON -.252 (.709) .126 2.013 (.866) AGEFIRST .250 (.265) .889 .098 (.236) PRIORREC .114 (.140) .665 .159 (.104) PRIORVIO .027 (1.07) .001 .453 (.701) PROCESS .338 (.301) 1.258 1.490 (.568) Constant 3.178 (5.05) .396 -3.289 (8.15) Log111.503 76.732 likelihood

* p < .10 ** p < .05 *** p < .01 137

|Wald| |Z-test| .602 .46 .126 .48 4.235** 1.73* 1.486 .85 6.712*** 2.26** 5.395** 2.02** .171 .43 2.326 .25 .418 .33 6.868*** 1.79* .163

138

Excluding Violent Youths From Juvenile Court

cases were disposed than were youths retained in juvenile court. Furthermore, the positive effect of transfer remained highly significant in the full model (b = 1.4715; p < .01). Because we are again examining a dichotomous dependent variable, we can use the results of the full model to predict the probability of rearrest prior to final disposition, for the released offenders. With all other explanatory variables set at their mean, the estimated probability of rearrest for transferred youths was .3870. For the retained juveniles, the estimated probability of rearrest was .1266. In other words, the probability of rearrest prior to disposition for an offender released by an adult court was .2604 (26 percentage points) higher than for a youth retained and released by a juvenile court. Three other significant effects were also revealed in the full model. Not surprisingly, the natural log of case processing time was found to have a positive effect (b = .5753; p < .05), suggesting that, among the offenders released from secure custody prior to disposition, those with greater time spent awaiting disposition were more likely to be rearrested 70 than were those who had their cases handled more rapidly. Also, age at referral was found to have a negative effect (b = -.5085; p < .10), meaning that older, released offenders were less likely to be rearrested than were their younger counterparts. Finally, somewhat surprisingly, released offenders with a family status of something other than living with two parents were less likely to be rearrested prior to final disposition than were released youths who were living with two parents (b = -.6961; p < .10). Although hard to explain at first glance, this finding becomes clearer when the results of the transfer specific models are considered. Among the youths retained and released by a juvenile court, the family status variable had a positive and insignificant effect on predispositional rearrest. However, among the transferred and released offenders, the family status coefficient was negative and significant (b = -2.0899; p < .01). This indicates that, while family status apparently had little impact on the pre-dispositional delinquent behavior of youths retained in juvenile court, juveniles in adult court who were 70

As stated by Smith and Paternoster (1990), "All else being equal, we expect the probability of any subsequent criminal behavior and the frequency of such behavior to increase with the length of the follow-up period" (p. 1115).

Quantitative Findings

139

living in “nontraditional” homes were less likely to be rearrested prior to disposition than were transferred youths who were living with two parents. Furthermore, the slope comparison z-test (z = 2.26) revealed that, among the released offenders, transferred offenders with nontraditional living arrangements were less likely to be rearrested prior to disposition than were retained youths with similar living arrangements. The above paragraph suggests that, although family status was previously found to have a direct, negative effect on pre-dispositional arrest, the impact was greater for youths in adult court. A possible explanation is that juveniles who are treated as adults and live with two parents experience family conflict as a consequence of their “criminal” behavior, which, in turn, increases the likelihood of further criminal behavior. On the other hand, juveniles in adult court with other living arrangements may experience less family conflict and, as a result, a reduced likelihood of further offending prior to disposition. Finally, retained juveniles who live with two parents may also experience less family conflict than youths in adult court with similar living arrangements, possibly due to their behavior still be viewed as “juvenile 71 delinquency” rather than criminal. Three other significant interaction effects were revealed in the transfer specific models. Among the retained and released juveniles, the effect of weapon type on pre-dispositional arrest was negative and insignificant. However, among the transferred and released offenders, the effect was positive and significant (b = 2.0126; p < .05), meaning that youths who had employed a firearm were more likely to be rearrested prior to disposition than were offenders who used some other deadly weapon. The z-test for the slope comparison (z = 2.02) confirmed that transferred and released juveniles who had employed a firearm were more likely to be rearrested prior to disposition than were retained and released youths who had used a gun. Somewhat similar findings were revealed in terms of case processing time and county of jurisdiction. Among the retained and 71

An alternative explanation may be that this variable is simply a poor proxy for actual family status. It could be, for instance, that some juveniles who were reported as living with two parents actually had little or no contact with one or both parents.

140

Excluding Violent Youths From Juvenile Court

released youths, the natural log of case processing time had a positive, but insignificant, effect on pre-dispositional arrest. Among the transferred and released offenders, the effect was positive and significant (b = 1.4898; p < .01). The slope comparison z-test (z = 1.79) verified a differential impact of case processing time between the two systems. Released offenders in adult court who experienced longer case processing times were more likely to be rearrested prior to disposition than were those in juvenile court whose cases took longer to be handled. Concerning county of jurisdiction, among the retained and released juveniles, the coefficient was positive and insignificant. In contrast, among the transferred and released youths, the effect was negative and significant (b = -2.0762; p < .05), suggesting that urban offenders were less likely to be rearrested prior to disposition than were their rural and suburban counterparts. The slope comparison z-test (z = 1.73) confirmed an interaction effect, meaning that urban youths who were transferred and released were significantly less likely to be rearrested prior to disposition than were urban youths retained and released by juvenile courts. Finally, in the previously discussed full model, age at referral was found to have a significant, negative effect on pre-dispositional arrest. With regard to the transfer specific models, among the retained and released juveniles, age at referral was again found to have a negative and significant effect (b = -.6625; p < .10). Among the transferred and released offenders, the coefficient for age at referral was also negative, but insignificant. However, the slope comparison z-test (z = .46) failed to support an interaction effect, meaning age at referral had a similar, negative impact on pre-dispositional arrest for released youths in both juvenile and adult court. Hypothesis 8: Violent Recidivism Prior To Final Disposition While Hypothesis 7 concerned the likelihood of rearrest on any charge prior to final disposition, Hypothesis 8 specifically focused on the likelihood of rearrest on a violent felony offense prior to final disposition. The logistic regression results may be found in Table 11. Again, we can start off by noting the positive bivariate effect of transfer (b = 1.9643; p < .01). In addition, the coefficient for transfer to adult court remained highly significant in the full model

Table 11: Logistic Regression Estimates for the Determinants of Pre-Dispositional Violent Felony Arrest (N=224) Variable Bivariate B (SE) |Wald| Multiple B (SE) TRANSFER 1.964 (.443) 19.674*** 1.717 (.563) AGEREF -.384 (.337) RACE -.137 (.586) COUNTY -.284 (.724) SCHOOL -.087 (.467) FAMILY -.467 (.497) WEAPON 1.064 (.688) AGEFIRST .108 (.181) PRIORREC .103 (.077) PRIORVIO .401 (.570) PROCESS .323 (.258) Constant -2.862 (.363) 61.995*** -.304 (5.18) Log153.699 144.301 likelihood

|Wald| 9.298*** 1.297 .054 .154 .035 .883 2.396 .358 1.808 .494 1.567 .003

Variable Nontransfer B (SE) |Wald| Transfer B (SE) |Wald| |Z-test| AGEREF -.853 (.640) 1.774 -.181 (.487) .138 .84 RACE .097 (1.02) .009 -.199 (.818) .059 .23 COUNTY 6.460 (26.8) .058 -1.329 (.947) 1.971 .29 SCHOOL -.631 (.843) .560 .232 (.614) .143 .83 FAMILY -.655 (.866) .571 -.465 (.692) .452 .17 WEAPON .011 (1.20) .000 1.773 (.893) 3.945** 1.19 AGEFIRST .304 (.418) .531 .045 (.232) .037 .54 PRIORREC .035 (.211) .027 .169 (.098) 2.978* .58 PRIORVIO 1.931 (1.27) 2.314 -.048 (.684) .005 1.37 PROCESS -.193 (.469) .169 .917 (.512) 3.208* 1.60 Constant 1.087 (28.0) .001 -4.418 (8.60) .264 Log55.413 79.114 likelihood

* p < .10 ** p < .05 *** p < .01 141

142

Excluding Violent Youths From Juvenile Court

(b = 1.7168; p < .01). Among the offenders who were released from secure custody, those who were transferred to adult court were more likely to be rearrested on a violent felony offense prior to final disposition than were youths who remained in juvenile court. As with rearrest on any charge prior to disposition, we can also use the results of the full model to predict the probability of rearrest on a violent felony offense, for the released offenders. With all other explanatory variables set at their mean, the estimated probability of rearrest on a violent felony offense for the transferred youths was .2296. For the offenders retained in juvenile court, the estimated probability of rearrest on a violent felony offense was .0507. Prior to disposition, the probability that a transferred and released offender would be rearrested on a violent felony offense was .1789 (nearly 18 percentage points) higher than the probability for a retained and released youth. Interestingly, in the full model, there were no other significant, direct effects found. Notably absent was the positive effect of case processing time, which was present in the prior model that examined rearrest on any charges prior to final disposition. Furthermore, in both of the full models that examined pre-dispositional recidivism, the effects of the prior offending variables were insignificant. Youths with more serious and extensive offending histories were not found to have a greater likelihood of rearrest or rearrest on a violent felony offense prior to disposition. However, these variables might play a larger role in the post-dispositional analyses, as a greater number of youths will be considered, generally over a longer follow-up period. While the full model did not reveal any other direct effects on pre-dispositional violent felony arrest, the transfer specific models also did not indicate the presence of any interaction effects. Among the youths retained and released by juvenile courts, none of the explanatory variables exerted a significant effect. Concerning the transferred and released offenders, those who had employed a firearm were significantly more likely to be rearrested on a violent felony offense prior to disposition (b = 1.7729; p < .05). However, the slope comparison z-test (z = 1.19) did not confirm a significant interaction. It should be pointed out that in the model pertaining to the retained offenders, weapon type also exerted a positive, albeit insignificant, effect. Moreover, in the original full model, the coefficient for weapon type was positive and near to being significant (b = 1.0644; p = .1216). These findings seem to suggest that use of a firearm did have some

Quantitative Findings

143

positive effect on pre-dispositional violent felony rearrest, particularly for those juveniles who were transferred to adult court. Two other findings from the transfer specific models are also worth mentioning. Among the transferred and released offenders, prior record (b = .1691; p < .10) and the natural log of case processing time (b = .9166; p < .10) were found to have a positive and significant effect on pre-dispositional violent felony arrest. Both coefficients for these variables were insignificant among the retained and released offenders. However, neither slope comparison z-test supported an interaction effect, meaning that these variables had no significant differential impact in the two separate court systems. Hypothesis 9: Recidivism Following Final Disposition The final three hypotheses in this study concerned the postdispositional recidivism of the 494 youths who were back on the street sometime prior to December 31, 1997. Hypothesis 9 focused on the likelihood of rearrest on any charges following final disposition. The logistic regression estimates are listed in Table 12. To begin, in contrast to the previous bivariate models, transfer to adult court was not initially found to have a significant influence on post-dispositional recidivism (b = .3855; p = .1315). However, in the full model, the transfer effect did become significant (b = .8525; p < .05), indicating that the youths transferred to adult court were more likely to be rearrested following final disposition than were their counterparts in juvenile court. To further assess the impact of transfer, we can again use the results of the full model to predict the probability of rearrest for transferred and nontransferred youths, with all other explanatory variables set at their mean. Among the youths who were back on the street following disposition, the estimated probability of rearrest for a transferred offender was .4433. The estimated probability of rearrest for a youth retained in juvenile court was .2534, meaning that the probability of rearrest for a transferred offender was .1899 (19 percentage points) higher than the probability for a retained youth. In addition to the influence of transfer, several other direct effects were revealed in the full model. Age at referral was found to have a significant, negative effect (b = -.2884; p < .05), meaning that

Table 12: Logistic Regression Estimates for the Determinants of Post-Dispositional Arrest (N=494) Variable Bivariate B (SE) |Wald| Multiple B (SE) TRANSFER .385 (.256) 2.274 .852 (.397) AGEREF -.288 (.145) RACE .519 (.322) COUNTY .281 (.357) SCHOOL .792 (.262) FAMILY -.175 (.252) WEAPON .317 (.363) AGEFIRST -.178 (.085) PRIORREC .065 (.043) PRIORVIO -.188 (.344) RISKTIME .237 (.197) RELEASE .047 (.246) CONVICT .606 (.275) INLENGTH -.066 (.018) PROCESS -.051 (.111) Constant -.876 (.108) 66.121*** .626 (4.01) Log607.611 552.190 likelihood

|Wald| 4.609** 3.957** 2.593 .622 9.119*** .482 .763 4.358** 2.294 .300 1.452 .037 4.842** 12.680*** .210 .024

* p < .10 ** p < .05 *** p < .01 (Continued . . .)

144

Quantitative Findings

145

Table 12 (continued) Variable Nontransfer B (SE) |Wald| Transfer B (SE) |Wald| |Z-test| AGEREF -.424 (.161) 6.887*** .735 (.524) 1.971 2.11** RACE .871 (.376) 5.349** -1.431 (.964) 2.201 2.22** COUNTY .341 (.421) .657 1.376 (.975) 1.990 .97 SCHOOL .826 (.300) 7.606*** 1.009 (.715) 1.989 .24 FAMILY -.297 (.278) 1.143 .583 (.775) .566 1.07 WEAPON .092 (.453) .041 1.820 (.852) 4.563** 1.79* AGEFIRST -.136 (.099) 1.880 -.443 (.221) 4.019** 1.27 PRIORREC .128 (.056) 5.157** -.103 (.097) 1.127 2.06** PRIORVIO -.318 (.416) .586 .160 (.805) .039 .53 RISKTIME .311 (1.63) .036 .124 (.114) 1.170 .11 RELEASE .104 (.276) .143 -.829 (.778) 1.136 1.13 CONVICT .690 (.300) 5.278** .767 (.996) .593 .07 INLENGT -.063 (.201) 9.228*** -.120 (.055) 4.722** .97 PROCESS .007 (.134) .003 -.663 (.348) 3.638* 1.80* Constant .315 (29.4) .001 -5.825 (8.86) .432 Log454.982 75.892 likelihood * p < .10 ** p < .05 *** p < .01

older youths were less likely to be rearrested following final disposition than were younger juveniles. Also, the coefficient for school status was positive and highly significant (b = .7919; p < .01), indicating that offenders who were enrolled, graduated, or had obtained a GED at the time of referral were more likely to be rearrested following disposition than were youths who were not enrolled. As mentioned earlier in this chapter, this finding concerning school status might be explained based on a process of weakened social bonds (Hirschi, 1969). Offenders who were enrolled may have been removed from school as a result of their initial arrest and court processing,

146

Excluding Violent Youths From Juvenile Court 72

increasing the likelihood of future offending. Moreover, youths who had graduated or obtained a GED may have lost employment as a consequence of their arrest and court processing, also increasing the likelihood of future offending. In the pre-dispositional analyses of recidivism, none of the prior offending variables was found to have a significant effect. In the current full model, however, there is evidence that at least one of these variables had an effect on post-dispositional rearrest. The coefficient for age at first referral was negative and significant (b = -.1777; p < .05), indicating that those offenders with an older age at first referral were less likely to be arrested following disposition than were youths with a younger age at first referral. Furthermore, although the positive effect of prior record failed to reach the .10 level of significance, it came fairly close (b = .0651; p = .1298). This at least suggests that offenders with extensive prior records were more likely to be arrested following disposition than were youths with brief offending histories. In the analyses of post-dispositional recidivism, four variables previously employed as outcome variables were now also used as explanatory variables: release from custody prior to disposition, conviction, incarceration length, and the natural log of case processing time. In the current full model, two of these variables were found to have a significant effect on post-dispositional rearrest. Conviction was found to have a positive influence (b = .6061; p < .05), meaning that youths who were convicted were more likely to be rearrested following disposition than were offenders who were not convicted. In contrast, incarceration length was found to have a negative effect (b = -.0657; p < .01), indicating that youths who were incarcerated for longer periods of time were less likely to be rearrested following disposition than were offenders who were either not incarcerated or were incarcerated for shorter time periods. The findings concerning conviction and incarceration length suggest that, for these serious and violent youthful offenders, being 72

Again, an alternative explanation may be that this variable is simply a poor proxy for actual school status. It could be that some juveniles who were reported as being enrolled in school may not have been attending school, or at least had a substantial number of absences.

Quantitative Findings

147

convicted was not enough to deter future offending. In fact, it increased the likelihood of post-dispositional recidivism. On the other hand, longer sentences of incarceration appeared to have a beneficial impact, which suggests a greater deterrent, rehabilitative, or combined effect from longer periods of confinement. In other words, with these offenders, severe punishment (i.e., longer-term incarceration) was more advantageous than was certain punishment (i.e., conviction). While not supported uniformly in the literature, these findings do fit in well with those of two other studies discussed in Chapter 3. The research by Murray and Cox (1979) and by Gottfredson and Barton (1993) also focused on more serious and frequent youthful offenders, and both studies produced evidence that supported the use of institutional confinement. It seems, then, that for the worst offenders, being arrested and convicted is not sufficient as a justice system response. Instead, more severe sanctions (i.e., longer periods of incarceration) may be necessary. We now turn to a consideration of the transfer specific models. These models, and the corresponding z-tests, reveal several significant interaction effects. To begin, among the youths retained in juvenile court, age at referral was found to have a significant, negative effect on post-dispositional rearrest (b = -.4237; p < .01). Among the transferred offenders, the coefficient for age at referral was positive and insignificant. The slope comparison z-test (z = 2.11) confirmed that youths in juvenile court with an older age at referral were significantly less likely to be rearrested following disposition than were older juveniles in adult court. With regard to race, for the nontransferred offenders, nonwhites were significantly more likely to be rearrested following disposition than were whites (b = .8708; p < .05). Among the youths waived to adult court, the coefficient for race was negative and insignificant. Again, the z-test for the slope comparison (z = 2.22) indicated a significant interaction. Nonwhite youths in juvenile court were more likely to be rearrested following disposition than were nonwhite offenders in adult court. Concerning weapon type, among the youths retained in juvenile court, the coefficient was positive and insignificant. For the offenders transferred to adult court, use of a firearm was found to have positive and significant effect on rearrest following final disposition (b = 1.8204; p < .05). The slope comparison z-test (z = 1.79) verified

148

Excluding Violent Youths From Juvenile Court

that youths in adult court who had used a gun during the commission of their offense were more likely to be rearrested following disposition than were offenders retained in juvenile court who had employed a firearm. In terms of prior record, for the nontransferred offenders, those with a more extensive prior record were significantly more likely to be rearrested following final disposition than were those with brief offending histories (b = .1281; p < .05). Among the transferred youths, the effect of prior record was negative and insignificant. The slope comparison z-test (z = 2.06) confirmed the differential impact of prior record. Youths retained in juvenile court with more extensive prior records were more likely to be rearrested following final disposition than were transferred juveniles with similar offending histories. With regard to the natural log of case processing time, among the retained offenders, the coefficient was positive and insignificant. In contrast, for the transferred youths, those with longer case processing times were significantly less likely to be rearrested following final disposition (b = -.6630; p < .10). The z-test for the slope comparison (z = 1.80) suggested a significant interaction. Offenders in adult court with longer case processing times were less likely to be rearrested following disposition than were youths retained in juvenile court who experienced longer periods of case processing. Before leaving the transfer specific models, several other findings are worth mentioning. In the previously discussed full model, the coefficient for school status was positive and significant. Among the offenders retained in juvenile court, again there was a positive and significant effect (b = .8261; p < .01), meaning that youths who were enrolled, graduated, or had obtained a GED at the time of referral were more likely to be rearrested following disposition than were those who were not enrolled. For the transferred offenders, the coefficient for school status was also positive, but insignificant. The slope comparison z-test (z = .24) failed to support a significant interaction. In other words, school status had a similar, positive effect on post-dispositional rearrest for youths in both juvenile and adult court. Somewhat similar findings were revealed with regard to conviction and incarceration length. Among the youths retained in juvenile court, conviction was found to have a positive and significant effect on rearrest following final disposition (b = .6900; p < .05). For the transferred offenders, the coefficient for conviction was also positive,

Quantitative Findings

149

but insignificant. The z-test for the slope comparison (z = .07) again failed to establish a significant interaction. This finding, as well as that of the previously discussed full model, indicates that being convicted had a similar, positive effect on post-dispositional rearrest for youths in both juvenile and adult court. In terms of incarceration length, among the retained juveniles, those who were incarcerated for longer periods of time were significantly less likely to be rearrested following final disposition (b = -.0626; p < .01). For the transferred youths, incarceration length also had a significant, negative effect (b = -.1198; p < .05). Not surprisingly, the slope comparison z-test (z = .97) failed to support any differential impact of incarceration length. This finding, as well as that of the full model, shows that incarceration length had a similar, negative effect on post-dispositional rearrest for offenders in both juvenile and adult court. Finally, in the full model, age at first referral had a negative and significant effect on post-dispositional rearrest. Among the youths retained in juvenile court, the coefficient for age at first referral was again negative, but insignificant. For the transferred offenders, the effect was negative and significant (b = -.4435; p < .05). However, the slope comparison z-test (1.27) did not establish a significant interaction. It appears, then, that age at first referral had a similar, negative effect on post-dispositional rearrest for youths in both juvenile and adult court. Hypothesis 10: Violent Recidivism Following Final Disposition While Hypothesis 9 focused on the likelihood of rearrest on any charge following final disposition, Hypothesis 10 specifically concerned the likelihood of rearrest on a violent felony offense, among the 494 youths who were back on the street sometime prior to December 31, 1997. The logistic regression results are presented in Table 13. We can begin by mentioning the significant, positive, bivariate effect of transfer (b = .5334; p < .10). However, in contrast to all of the previous full models, the positive effect of transfer on post-dispositional violent felony arrest became insignificant when the effects of the other explanatory variables were taken into account. However, it should also be noted that, while the effect of transfer failed to reach the .10 significance level, it came fairly close (b = .6916; p = .1355).

Table 13: Logistic Regression Estimates for the Determinants of Post-Dispositional Violent Felony Arrest (N=494) Variable Bivariate B (SE) |Wald| Multiple B (SE) TRANSFER .533 (.296) 3.250* .692 (.463) AGEREF -.115 (.169) RACE .502 (.439) COUNTY .900 (.515) SCHOOL .295 (.306) FAMILY .437 (.345) WEAPON .303 (.475) AGEFIRST -.153 (.096) PRIORREC .091 (.048) PRIORVIO -.258 (.398) RISKTIME .035 (.070) RELEASE .056 (.299) CONVICT .497 (.335) INLENGTH -.038 (.021) PROCESS -.194 (.130) Constant -1.683 (.135) 155.355*** .067 (3.01) Log-likelihood 447.410 410.821

|Wald| 2.228 .457 1.307 3.052* .929 1.612 .407 2.532 3.674* .422 .250 .035 2.209 3.218* 2.230 .001

* p < .10 ** p < .05 *** p < .01 (Continued . . .)

150

Quantitative Findings

151

Table 15 (continued) Variable Nontransfer B (SE) |Wald| Transfer B (SE) |Wald| |Z-test| AGEREF -.345 (.198) 3.046* .705 (.562) 1.577 1.76* RACE 1.028 (.570) 3.251* -.967 (1.03) .875 1.69* COUNTY .687 (.581) 1.396 1.766 (1.30) 1.836 .76 SCHOOL .359 (.363) .978 -.298 (.795) .141 .75 FAMILY .449 (.399) 1.269 .278 (.886) .098 .18 WEAPON -.464 (.598) .602 1.824 (.938) 3.778* 2.06** AGEFIRST -.025 (.119) .044 -.470 (.232) 4.094** 1.71* PRIORREC .193 (.063) 9.421*** -.084 (.102) .676 2.31** PRIORVIO -.178 (.475) .141 -.155 (.865) .032 .02 RISKTIME .614 (3.61) .029 -.012 (.118) .011 .17 RELEASE .142 (.350) .165 -.622 (.819) .577 .86 CONVICT .571 (.381) 2.239 .179 (1.04) .030 .35 INLENGTH -.028 (.024) 1.400 -.120 (.065) 3.412* 1.33 PROCESS -.293 (.162) 3.283* -.328 (.400) .671 .08 Constant -8.225 (65.2) .016 -4.654 (9.94) .219 Log324.469 62.458 likelihood * p < .10 ** p < .05 *** p < .01

Even though the positive effect of transfer failed to reach the level of significance employed in this study, we can still use the results of the full model to predict the probability of post-dispositional violent felony arrest for transferred and nontransferred offenders, with all other explanatory variables set at their mean. Among the youths who were back on the street following disposition, the estimated probability of rearrest for a transferred offender was .2305. The estimated probability of rearrest for a youth retained in juvenile court was .1304. So, although the effect of transfer was not found to be significant in the full model, the probability of rearrest for a transferred offender was still .1001 (10 percentage points) higher than the probability for a retained youth. The full model did reveal a few other significant effects on post-

152

Excluding Violent Youths From Juvenile Court

dispositional violent felony arrest. County of jurisdiction was found to have a positive effect (b = .8998; p < .10), meaning that urban youths were more likely to be rearrested on a violent felony offense following disposition than were their rural and suburban counterparts. In addition, the effect of prior record was also positive and significant (b = .0912; p < .10). Offenders with more extensive prior records were more likely to be rearrested on a violent felony offense following disposition than were youths with brief offending histories. This finding corresponds with the negative and nearly significant effect of age at first referral (b = -.1532; p = .1115). Although failing to reach the .10 level, there is at least suggestive evidence that offenders with an older age at first referral were less likely to be rearrested on a violent felony offense following disposition than were youths with a younger age at first referral. Finally, we again see a significant, negative effect from incarceration length (b = -.0382; p < .10). This suggests that the original bivariate transfer effect was mediated by incarceration length, as offenders who were incarcerated for longer periods of time were less likely to be rearrested on a violent felony offense following disposition than were those who were either not incarcerated or were incarcerated for shorter time periods. As with the previous full model, this finding implies a greater deterrent, rehabilitative, or combined effect from longer periods of confinement. We now consider the possibility of interaction effects. Based on the results of the transfer specific models, several variables had a differential impact on the transferred and nontransferred offenders. To begin, among the youths retained in juvenile court, age at referral had a significant, negative effect on post-dispositional violent felony arrest (b = -.3454; p < .10). For the transferred offenders, the effect of age at referral was positive and insignificant. The z-test for the slope comparison (z = 1.76) indicated that older youths retained in juvenile court were less likely to be rearrested on a violent felony offense following disposition than were older offenders transferred to adult court. Concerning race, among the youths retained in juvenile court, nonwhites were significantly more likely to be rearrested on a violent felony offense following disposition than were whites (b = 1.0285; p < .10). For the transferred offenders, the coefficient for race was negative and insignificant. The slope comparison z-test (z = 1.69)

Quantitative Findings

153

revealed a significant interaction. Nonwhite youths in juvenile court were more likely to be rearrested on a violent felony offense following disposition than were nonwhite offenders in adult court. With regard to weapon type, among the nontransferred offenders, the coefficient was negative and insignificant. In contrast, for the youths in adult court, use of a firearm had a positive and significant effect on post-dispositional violent felony arrest (b = 1.8239; p < .10). The slope comparison z-test (z = 2.06) confirmed the differential impact of weapon type. Transferred juveniles who had employed a firearm were significantly more likely to be rearrested following final disposition than were retained youths who had used a gun. In terms of age at first referral, among the offenders retained in juvenile court, the coefficient was negative and insignificant. For the transferred youths, age at first referral had a negative and significant effect (b = -.4696; p < .05). The z-test for the slope comparison (z = 1.71) verified a significant interaction. Offenders in adult court with an older age at first referral were less likely to be rearrested on a violent felony offense following disposition than were retained youths with an older age at first referral. Finally, concerning prior record, nontransferred offenders with more extensive prior records were significantly more likely to be rearrested on a post-dispositional violent felony offense than were retained youths with brief offending histories (b = .1934; p < .01). However, for the transferred offenders, the effect of prior record was negative and insignificant. The slope comparison z-test (z = 2.31) again confirmed a significant interaction. Retained youths with extensive prior records were more likely to be rearrested on a violent felony offense following disposition than were transferred offenders with broader offending histories. Two other findings are worth noting with regard to the transfer specific models. Among the nontransferred youths, incarceration length was found to have a negative and insignificant effect. For the transferred offenders, the effect of incarceration length was also negative, but significant (b = -.1201; p < .10). However, the slope comparison z-test (z = 1.33) failed to establish any differential impact from incarceration length. This finding, combined with the results of the full model, suggest that incarceration length had a similar, negative effect on post-dispositional violent felony arrest for youths in both juvenile and adult court.

154

Excluding Violent Youths From Juvenile Court

Lastly, in the full model, the coefficient for the natural log of case processing time was negative and failed to reach the .10 level of significance. However, it came fairly close (b = -.1944; p = .1353). Among the youths retained in juvenile court, the effect of this variable was negative, and now significant (b = -.2930; p < .10). For the transferred offenders, the coefficient was again negative, but insignificant. Although the slope comparison z-test (z = .08) failed to support a significant interaction, the combined results seem to suggest that youths (particularly those retained in juvenile court) who experienced longer periods of case processing were less likely to be rearrested on a violent felony offense following disposition. This could possibly be due to a delayed deterrent effect from having a case “hang over one’s head,” or it may be that rehabilitative services were offered and received during longer periods of case processing. Hypothesis 11: Time To Recidivism Following Final Disposition Hypotheses 9 and 10 concerned the likelihood of recidivism following final disposition, and to test these hypotheses, binary measures of rearrest were employed. Hypothesis 11 specifically addressed time to rearrest among the 494 offenders who were back on the street sometime prior to December 31, 1997. To test this hypothesis, 73 a Cox proportional hazards model was estimated. The results not only suggest factors that have a significant effect on how quickly youths reoffend following disposition and release from secure confinement, but they also provide insight concerning the frequency of future offending, as high-rate offenders would be expected to have a shorter time until rearrest than would low-rate offenders.

73

An assumption of this model is that for any two cases at any point in time, the rate of the estimated hazard rates is a constant (Allison, 1984; Cox, 1972). This assumption was assessed for transferred and nontransferred offenders. First, through a visual inspection, the hazard rates for these youths appeared proportional. Second, a formal test was constructed using a time-dependent covariate, as hazard rates will not be proportional if there is a significant interaction between time and an explanatory variable (in this case, transfer to adult court). The results of this test provided confirmatory evidence that the hazard rates for transferred and nontransferred youths were proportional.

Quantitative Findings

155

The results of the survival analysis may be found in Table 14. To begin, in the bivariate model, the coefficient for transfer is positive and significant (b = .3607; p < .10). This suggests that, at any point in time following disposition, transferred offenders experienced an increased risk of rearrest (i.e., decreased survival times) as compared to nontransferred youths. Furthermore, the effect of transfer became even more significant in the full model (b = .6827; p < .05). These findings indicate that youths waived to adult court were rearrested more quickly following final disposition than were their counterparts who were retained in juvenile court. The results also at least suggest, then, that transferred youths were more frequent offenders following final disposition than were retained juveniles. In order to further assess the impact of transfer on time to rearrest, we can examine the “relative risk” associated with transfer to adult court. As shown in Table 14, the antilog (i.e., exponentiated coefficient) of the coefficient for transfer is 1.9792. This means that the hazard rate of arrest is 1.98 times greater for transferred offenders than for youths retained in juvenile court, while adjusting for the other factors in the model. In other words, we can say that the hazard rate for transferred offenders was about 98% larger than the hazard rate for retained youths. In the full model, there were also several other significant effects revealed. First, the coefficient for school status was positive and highly significant (b = .5975; p < .01). This finding indicates that, at any point in time following disposition, youths who were enrolled in school, graduated, or had obtained a GED at the time of referral experienced an increased risk of rearrest (i.e., decreased survival times) as compared to offenders who were not enrolled in school. In fact, the hazard rate for youths either in school or having completed school was 82% larger than that of offenders who were not in school (exp(b) = 1.8175). This finding corresponds with that of a previous model which examined the likelihood of rearrest following final disposition, and again may suggest a process of weakened social bonds (e.g., removal from school, loss of job) which increases both the likelihood and frequency of offending following final disposition. With regard to the prior offending variables, age at first referral was found to have a significant, negative effect (b = -.1476; p < .05). This signifies that, at any point in time following disposition, youths with an older age at first referral experienced a decreased risk of

Table 14: Cox Regression Estimates for the Determinants of Post-Dispositional Survival Time (N=494) Variable Bivariate B (SE) |Wald| Multiple B (SE) |Wald| TRANSFER .361 (.204) 3.131* .683 (.302) 5.109** AGEREF -.177 (.111) 2.544 RACE .412 (.261) 2.484 COUNTY .369 (.292) 1.596 SCHOOL .597 (.213) 7.882*** FAMILY -.180 (.199) .812 WEAPON .245 (.280) .766 AGEFIRST -.148 (.064) 5.331** PRIORREC .041 (.030) 1.881 PRIORVIO -.121 (.271) .198 RELEASE .110 (.196) .318 CONVICT .632 (.214) 8.710*** INLENGTH -.059 (.015) 15.532*** PROCESS -.078 (.084) .874 1833.401 1782.784 Loglikelihood

Exp(B) 1.979 .837 1.509 1.446 1.817 .835 1.278 .863 1.042 .886 1.117 1.881 .943 .924

* p < .10 ** p < .05 *** p < .01 (Continued . . .)

156

Quantitative Findings

157

Table 14 (continued) Variable Nontransfer B (SE)|Wald| Transfer B (SE) |Wald| AGEREF -.307 (.127) 5.825** .567 (.342) 2.745* RACE .705 (.316) 4.983** -.447 (.568) .621 COUNTY .322 (.339) .905 1.277 (.777) 2.707* SCHOOL .676 (.255) 6.998*** .234 (.435) .289 FAMILY -.247 (.225) 1.201 .112 (.511) .048 WEAPON .032 (.377) .007 1.027 (.519) 3.919** AGEFIRST -.125 (.075) 2.761* -.223 (.147) 2.302 PRIORREC .094 (.040) 5.550** -.080 (.070) 1.277 PRIORVIO -.210 (.327) .413 .377 (.539) .489 RELEASE .172 (.221) .604 -.683 (.474) 2.079 CONVICT .659 (.237) 7.731*** .955 (.601) 2.524 INLENGTH -.050 (.016) 9.423*** -.118 (.040) 8.840*** PROCESS -.014 (.105) .017 -.478 (.194) 6.086** Log1387.496 222.032 likelihood

|Z-test| 2.39** 1.77* 1.13 .87 .64 1.55 .60 2.14** .93 1.64 .46 1.58 2.11**

* p < .10 ** p < .05 *** p < .01

rearrest (i.e., increased survival times) as compared to offenders with a younger age at first referral. The antilog of the coefficient (exp(b) = .8628) shows that a 1-year increase in age at first referral decreases the hazard rate by an estimated 14%. Finally, similar to the previous findings concerning the likelihood of rearrest following disposition, conviction and incarceration length were found to have a significant influence on the time to rearrest during this time period. Conviction had a positive effect (b = .6317; p < .01), meaning that, at any point in time following disposition, youths who were convicted experienced a greater risk of rearrest (i.e., decreased survival times) as compared to offenders who were not convicted. Furthermore, the antilog of the coefficient (exp(b) = 1.8807) suggests that the impact of conviction was substantial, as the hazard rate for

158

Excluding Violent Youths From Juvenile Court

convicted youths was 88% larger than that of the offenders who were not convicted. In contrast to these findings, the coefficient for incarceration length was negative and highly significant (b = -.0588; p < .01). In other words, at any point in time following disposition, youths who were incarcerated for longer periods of time experienced a decreased risk of rearrest (i.e., increased survival times) as compared to youths who were either not incarcerated or were incarcerated for shorter periods of time. The antilog of the coefficient (exp(b) = .9429) shows that a 1-month increase in incarceration length decreases the hazard rate by an estimated 6%. Overall, the findings concerning conviction and incarceration length again suggest that for serious and violent youthful offenders, being convicted is not enough to deter future offending. Instead, longer periods of confinement in correctional facilities appear necessary, whether they provide a deterrent, rehabilitative, or some combination of effects. In order to consider the possibility of interaction effects, we again turn to the transfer specific models. To begin, in the previously discussed full model, age at referral was found to have an insignificant effect on post-dispositional survival times. The transfer specific models suggest that age at referral actually had a contradictive effect in the two separate court systems. Among the offenders retained in juvenile court, the effect of age at referral was negative and significant (b = -.3067; p < .05). This means that, at any point in time following disposition, older youths retained in juvenile court experienced a decreased risk of rearrest (i.e., increased survival times) as compared to younger offenders. In contrast, for the transferred youths, the coefficient for age at referral was positive and significant (b = .5670; p < .10). In other words, at any point in time following disposition, older youths transferred to adult court experienced an increased risk of rearrest (i.e., decreased survival times) as compared to younger offenders. Finally, the slope comparison z-test (z = 2.39) confirmed the significantly different impact of age at referral in the two separate court systems. With regard to race, among the youths retained in juvenile court, the coefficient was positive and significant (b = .7049; p < .05). This indicates that, at any point in time following disposition, nonwhites experienced an increased risk of rearrest (i.e., decreased survival times) as compared to whites. For the transferred offenders, the effect of race was negative and insignificant. The z-test for the slope comparison

Quantitative Findings

159

(z = 1.77) verified a significant interaction. Being nonwhite had a stronger, positive effect on the risk of rearrest for youths retained in juvenile court as compared to those transferred to adult court. Concerning prior record, among the youths retained in juvenile court, the coefficient was positive and significant (b = .0940; p < .05). In other words, at any point in time following disposition, offenders with more extensive prior records experienced an increased risk of rearrest (i.e., decreased survival times) as compared to youths with brief offending histories. For the transferred offenders, the effect of prior record was negative and insignificant. The slope comparison z-test (z = 2.14) supported the differential impact of prior record. Prior record had a greater, positive impact on the risk of rearrest for youths retained in juvenile court as compared to those waived to adult court. A final significant interaction pertained to case processing time. Among the youths retained in juvenile court, the effect of the natural log of case processing time was negative and insignificant. For the transferred offenders, the effect was also negative, but significant (b = -.4777; p < .05). This suggests that, at any point in time following disposition, youths in adult court who had received longer periods of case processing experienced a decreased risk of rearrest (i.e., increased survival times) as compared to those who had received shorter processing times. The z-test for the slope comparison (z = 2.11) confirmed an interaction effect. Longer periods of case processing had a stronger, negative effect on the risk of rearrest for offenders in adult court than for those retained in juvenile court. Before leaving the transfer specific models, several other findings deserve mention. In the previously discussed full model, school status was found to have a significant, positive effect on the risk of rearrest at any point in time following final disposition. Among the youths retained in juvenile court, the effect of school status was also positive and significant (b = .6759; p < .01). For the transferred offenders, the coefficient was positive, but insignificant. The slope comparison z-test (z = .87) did not reveal a significant interaction. These findings suggest that the school status variable had a similar, positive effect for youths in both juvenile and adult court. Similarly, in the full model, age at first referral had a significant, negative effect on the risk of rearrest at any point in time following disposition. Among the youths retained in juvenile court, the effect of age at first referral was also negative and significant (b = -.1247;

160

Excluding Violent Youths From Juvenile Court

p < .10). For the transferred offenders, the coefficient was negative, but insignificant. The z-test for the slope comparison (z = .60) did not reveal a differential impact from age at first referral. This indicates that age at first referral had a similar, negative effect for youths in both juvenile and adult court. In the full model, conviction and incarceration length both had significant effects on the risk of rearrest at any point in time following disposition. Youths who were convicted experienced shorter postdispositional survival times, while those who were incarcerated for longer time periods experienced lengthier post-dispositional survival times. The results of the transfer specific models suggest that the effects of conviction and incarceration length were similar for offenders in both juvenile and adult court. Among the youths retained in juvenile court, the effect of conviction was positive and significant (b = .6588; p < .01), while the impact of incarceration length was negative and significant (b = -.0501; p < .01). For the transferred offenders, conviction had a positive, but insignificant, effect, while incarceration length again had a significant, negative impact (b = -.1176; p < .01). The slope comparison z-test for conviction (z = .46) and for incarceration length (z = 1.58) did not support significant interactions, meaning the effects of these variables were similar for youths in both court systems. Two final findings are worth noting. In the full model, county of jurisdiction and weapon type were both found to have insignificant, positive effects on the risk of rearrest at any point in time following disposition. Among the youths retained in juvenile court, the coefficients for these two variables were also positive and insignificant. However, for the transferred offenders, county of jurisdiction (b = 1.2776; p < .10) and weapon type (b = 1.0267; p < .05) both had a significant, positive impact. Although the slope comparison z-tests for county of jurisdiction (z = 1.13) and weapon type (z = 1.55) did not reveal significant interaction effects, these findings at least suggest that for transferred youths, those from urban counties and those who had used a firearm experienced shorter survival times than did offenders from suburban and rural counties and youths who had employed some other type of deadly weapon.

Quantitative Findings

161

Summary In 11 bivariate and multivariate models, we tested for the direct effect of transfer to adult court on various dependent variables. In general, the dependent variables represented the case outcomes and future criminal behavior of the violent youthful offenders examined in this study. At the .10 level, 10 significant transfer effects were observed, while controlling for the other factors. In terms of case outcomes, as compared to youths retained in juvenile court, transferred offenders were more likely to be released from secure custody prior to disposition and more likely to be convicted. Furthermore, of the convicted youths, those that were transferred were more likely to be convicted on a target offense of robbery or aggravated assault and more likely to be incarcerated. Of the incarcerated offenders, those in adult court experienced longer periods of confinement. Lastly, transferred youths also experienced longer periods of case processing. Concerning future criminal behavior, of those youths released from secure custody prior to disposition, transferred offenders were more likely to be rearrested during the pre-dispositional time period and also more likely to be rearrested on a violent felony offense. Of those youths who were back on the street during the post-dispositional follow-up period, transferred offenders were more likely to be rearrested and experienced shorter survival times as compared to retained youths. The only insignificant transfer effect found in this study pertained to postdispositional rearrest on a violent felony offense. However, in the full model corresponding to that dependent variable, the coefficient for transfer was positive and came fairly close to the .10 level. Some general observations can also be made regarding the other explanatory variables employed in the full models. As compared to younger offenders, older youths tended to be treated less harshly at the beginning of the justice system process, as they were more likely to be released from custody prior to disposition and less likely to be convicted. Concerning recidivism, of the offenders released from custody prior to disposition, older youths were less likely to be rearrested during the pre-dispositional time period. Finally, as compared to their younger counterparts, older offenders were also less likely to be rearrested following final disposition.

162

Excluding Violent Youths From Juvenile Court

County of jurisdiction was also found to have an impact on several dependent variables. With regard to case outcomes, urban youths were more likely to be released prior to final disposition than were suburban and rural offenders. Urban juveniles were less likely to be convicted, and they experienced longer periods of case processing. Lastly, concerning future criminal behavior, urban offenders were more likely to be rearrested on a violent felony offense following final disposition. Type of weapon employed during the offense appeared to be a major determinant of case outcomes. Offenders who used a firearm were less likely to be released from custody prior to disposition than were those who employed some other type of deadly weapon. Of the convicted youths, those who used a gun were more likely to be convicted of robbery or aggravated assault. Convicted offenders who employed a firearm were also sentenced more severely, as they were more likely to be incarcerated, and when incarcerated, they experienced longer periods of confinement. Although weapon type had significant effects on these case outcome variables, it was not found to impact on any of the measures of recidivism. The prior offending variables were found to have numerous direct effects on the dependent variables, particularly in terms of future criminal behavior following final disposition. First, though, concerning case outcomes, of the convicted offenders, those with a prior substantiated violent felony offense were more likely to be convicted of robbery or aggravated assault. Of the incarcerated youths, those with a prior substantiated violent felony offense experienced longer periods of confinement. Lastly, age at first referral was found to impact on case processing time, as juveniles with an older age at first referral experienced longer periods of case processing. While the prior offending variables were not found to be significant determinants of recidivism prior to final disposition, they did have an impact on criminal behavior in the post-dispositional follow-up period. Offenders with an older age at first referral were less likely to be rearrested following final disposition, and they experienced longer survival times. There was also suggestive evidence that these youths were less likely to be rearrested on a violent felony offense following disposition. Conversely, juveniles with a more extensive prior record were more likely to be rearrested on a violent felony offense following disposition than were youths with brief offending histories, and the

Quantitative Findings

163

effect of prior record on rearrest on any charges following disposition was also positive and nearly significant. With regard to the four case outcome variables that were used as explanatory variables in the models that examined pre- and postdispositional recidivism, several significant effects were found. As expected, in the pre-dispositional time period, released offenders who experienced longer periods of case processing were more likely to be rearrested. In the post-dispositional time period, youths who were convicted were more likely to be rearrested, and they experienced shorter survival times. In contrast, offenders who were incarcerated for longer periods of time were less likely to be rearrested, less likely to be rearrested on a violent felony offense, and experienced longer survival times as compared to youths who were either not incarcerated or were confined for shorter periods of time. It seems that merely convicting these violent youthful offenders was counter-productive, while longerterm incarceration had a beneficial impact. While the explanatory variables discussed above appeared to have substantial effects on case outcomes and future criminal behavior, the impact of race, school status, and family status was limited. At the .10 level, no direct race effects were found, although there was suggestive evidence that nonwhites were less likely to be released from custody prior to disposition and more likely to be rearrested following disposition than were whites. Offenders who were enrolled in school, graduated, or had obtained a GED at the time of referral were more likely to be rearrested following final disposition and experienced shorter survival times, possibly explainable based on a process of weakened social bonds resulting from justice system processing. Finally, youths from nontraditional homes (i.e., something other than living with two parents) were less likely to be convicted and less likely to be rearrested during the pre-dispositional time period. In addition to the full models, which tested for direct effects, transfer specific models were also estimated to consider the possibility that the effect of an explanatory variable was dependent on whether the offender was transferred to adult court or retained in juvenile court. Based on the results of the 11 pairs of transfer specific models constructed, we found evidence of 32 statistically significant interaction effects (z > 1.65). More specifically, 28% of the variables in our models had a differential impact for transferred and nontransferred youths.

164

Excluding Violent Youths From Juvenile Court

By considering the significant interaction effects across the 11 dependent variables, several interesting patterns may be observed. To begin, age at referral had a differential impact on five dependent variables. An older age at referral had a differential effect on the natural log of case processing time, significantly increasing processing times for youths in adult court, but not for those in juvenile court.. Also, of the convicted youths, older offenders in adult court were less likely to be convicted of robbery or aggravated assault than were similarly aged youths in juvenile court. These findings may indicate that in adult court, older youths were viewed as less serious offenders, or less in need of attention. If true, the findings concerning post-dispositional recidivism provide reason to question this belief, as older transferred offenders were more likely to be rearrested, more likely to be rearrested on a violent felony offense, and experienced shorter survival times as compared to similarly aged youths retained in juvenile court. Concerning race, among the incarcerated offenders, being nonwhite had a differential impact on the natural log of incarceration length, significantly increasing periods of confinement for offenders retained in juvenile court, but not for those waived to adult court. However, the findings concerning post-dispositional recidivism suggest that this lengthier incarceration did not have a beneficial effect, as nonwhites retained in juvenile court were more likely to be rearrested, more likely to be rearrested on a violent felony offense, and experienced shorter survival times as compared to nonwhites transferred to adult court. In the full models, it was revealed that incarceration length had a negative, direct effect on all three measures of post-dispositional recidivism. The findings from the transfer specific models imply that for nonwhites incarcerated by a juvenile court, the beneficial impact of greater incarceration length was not present. It may be that because incarcerated nonwhites in juvenile court experienced longer periods of confinement than did incarcerated whites, they perceived that they were being treated unfairly (see Paternoster et al., 1997; Paternoster & Piquero, 1995; Sherman, 1993; Tyler, 1990). In turn, this perception may have lead to greater post-dispositional recidivism in comparison to nonwhites who were transferred to adult court. Weapon type was found to have a differential impact on seven dependent variables. In general, gun users were treated more harshly in adult court than in juvenile court. Offenders who employed a firearm were more likely to be convicted in adult court, and of those convicted,

Quantitative Findings

165

transferred gun users were more likely to be convicted of robbery or aggravated assault. Among the incarcerated youths, firearm use had a differential impact on the natural log of incarceration length, increasing periods of confinement for offenders in adult court, but not for those in juvenile court. Gun use also had a different effect on the natural log of case processing time, significantly decreasing processing times for youths in juvenile court, but not for those transferred to adult court. Concerning future criminal behavior, of the offenders released from custody prior to disposition, transferred youths who utilized a firearm were more likely to be rearrested during the pre-dispositional time period than were gun users retained in juvenile court. Similarly, in the post-dispositional time period, transferred youths who employed a firearm were more likely to be rearrested, and also more likely to be rearrested on a violent felony offense, than were gun users retained in juvenile court. The above findings suggest that firearm users who were transferred to adult court were the “worst of the worst.” Initially, they were judged to be beyond the treatment capacity of the juvenile court, resulting in their transfer. In adult criminal court, they were treated more harshly than were youths retained in juvenile court that had employed a gun. Unfortunately, this harsher treatment did not appear to pay off, as the transferred firearm users exhibited greater recidivism both prior to and following disposition in comparison to gun users retained in juvenile court. Finally, a few comments should be made with regard to the importance of prior record. It appears that the extent of a youth’s prior record was given somewhat more attention in adult court. Transferred offenders with more extensive prior records were less likely to be released from custody prior to disposition than were retained youths with similarly extensive records. Prior record also had a differential impact on the natural log of case processing time, significantly reducing processing times for offenders in adult court, but not for those in juvenile court. Lastly, in the post-dispositional follow-up period, youths retained in juvenile court with lengthier prior records were more likely to be rearrested, more likely to be rearrested on a violent felony offense, and experienced shorter survival times as compared to transferred offenders with more extensive prior records.

This page intentionally left blank

CHAPTER 6

Qualitative Findings

The quantitative portion of this study examined over 550 violent youthful offenders from Pennsylvania who would have been excluded from juvenile court jurisdiction in 1994, had a more recently enacted legislative waiver law been in effect at that time. In an additional qualitative component, face-to-face interviews were conducted with 18 justice system officials (i.e., prosecutors, public defenders, juvenile court judges, and juvenile probation officials) in three diverse Pennsylvania counties (i.e., a large metropolitan county, a smaller urban county, and a rural/suburban county). This part of the study was designed to gain a better understanding of how the new legislative waiver law was working in practice and to see how the views of justice system officials corresponded with our quantitative findings.

Overall Effectiveness of Legislative Waiver In general, opinions on the overall effectiveness of Pennsylvania’s recent legislation varied according to the employment position of the justice system official who was interviewed. Prosecutors in the three counties were in uniform agreement that the new law provides for an effective response to violent juvenile offending: [Rural/Suburban Prosecutor:] Any type of legislation giving more controls over younger criminals is a good thing. Really, the whole system should be revamped. Ages should be lowered for adult court jurisdiction, and all violent offenders should be treated as adults. . . . The juvenile court approach is just not working. These kids are dangerous, and they are not going to be rehabilitated. To protect the public, there is a need 167

168

Excluding Violent Youths From Juvenile Court for long-term incarceration, which juvenile courts can’t provide. [Small Urban Prosecutor:] It’s a good law in practice. We can identify the serious cases early, without going through the certification process. [Large Metropolitan Prosecutor:] Prior to the new law, the problem in Pennsylvania was one of two extremes. We either put people away forever, or treated them too leniently. The new law allows for special consideration of older teens with serious cases, which the juvenile court can’t really deal with.

In contrast to prosecutors, public defenders were strongly opposed to the legislative waiver law: [Rural/Suburban Public Defender:] It’s not an effective approach, because it doesn’t deal with people, it deals with actions. Resources and services for teenagers just don’t exist in adult facilities, and we need to remember that the person will almost always get out. . . . The juvenile court has more chance to change an offender. [Small Urban Public Defender:] It’s a dangerous law. We have a juvenile system because we decided that those under 18 should be treated differently, and with the new law, we are quitting on these offenders. Waived juveniles will still get out of prison at a young age, so society still has to put up with them. . . . These kids are poor, come from broken homes, and often have family members in jail. If we give up on them and give them a criminal record, it’s only going to make it tougher for them to stay straight once their sentence is over. [Large Metropolitan Public Defender:] I don’t believe in legislative waiver laws, because they treat the offense rather than the offender and create rigid statutory schemes. We need individual case consideration of life circumstances, which the new law tries to eliminate. Among the juvenile probation officials, general opinions about the new law were mixed. However, these individuals almost uniformly

Qualitative Findings

169

lamented the lack of treatment and rehabilitation possibilities once a youth is sent to adult court: [Rural/Suburban Probation Officer:] It’s not a bad law for the type of offenses listed. More serious cases are immediately removed from juvenile court, and decertification allows for some to be sent back as needed. . . . But it does lead to teenagers being put in adult jails and prisons, where there’s limited services for them. First-time offenders probably are the worst off, since they don’t get the services and have to spend time with adult criminals. [Rural/Suburban Probation Officer:] It may not always be a good law for the offender, who might get put into jail with adults, but it is good for public perception and provides a uniform starting point in every jurisdiction. [Small Urban Probation Officer:] I don’t think the law has had much effect on juvenile crime, especially for hard-core offenders. Many of them want to be transferred anyway. They see it as a sign of status and don’t worry about the long-term consequences. [Large Metropolitan Probation Supervisor:] It’s a diapers-toadulthood approach. Most of these kids have a very poor home-life, with no parenting, and start committing crime at an early age. These things are not dealt with in the adult system, and the problems will not go away overnight. . . . We should have separate facilities for these teenagers. They should not be mingled with adults. They also should receive special services, which they won’t get in prison. [Large Metropolitan Probation Supervisor:] If the goal is to incapacitate for the long term without change, the adult system is a better option. If rehabilitation and change is the goal, then the juvenile court has an advantage. The juvenile court has a concrete system of programs with a demonstrated level of effectiveness, and these programs are not available in the adult system.

170

Excluding Violent Youths From Juvenile Court Juvenile court judges presented somewhat similar views: [Rural/Suburban Judge:] There’s some need to coordinate processing across the state, and the law does establish uniformity. But I don’t think it has much effect on violent behavior, because a hardened criminal or delinquent is still going to take chances. [Small Urban Judge:] If the right facilities and staff are available, the new system will work. . . . These kids come from a background of no father, school failure, and street morals and attitudes. They can be turned around with time and effort, but we need strong techniques and special sentencing options. Without special options, and facility and staff commitment, there’s not much chance for success. [Large Metropolitan Judge:] It’s a shell game. Only people who are not familiar with the system are fooled by it. The law took serious offenders off the street initially, but adult facilities are even more overcrowded than juvenile facilities, and after a short time they are back on the street. . . . Politicians focused on high profile violent crimes by juveniles to create a fear in voters’ minds, but they don’t understand why people commit crime. The inner city is a goddamned jungle, and this law has nothing to do with what teenagers experience on a day-to-day basis.

Overall, in these three counties, prosecutors appear to support the legislative waiver approach based on a perceived increase in public protection obtained through targeting violent juveniles early in the justice system process. On the other hand, public defenders oppose the law, on the basis that it does not consider factors associated with the individual offender. Furthermore, these public defenders believe the juvenile system, through more effective services and facilities, has a better chance of changing the behavior of violent offenders. While juvenile probation officials and juvenile court judges generally presented mixed views, they also stressed the need for special attention and resources devoted to violent youths, which they did not believe to be as prevalent in the adult system.

Qualitative Findings

171

Legislative Waiver Versus Judicial Waiver: Strengths and Weaknesses As discussed in Chapter 2, Pennsylvania essentially moved from a transfer process based on judicial waiver to one that is centered on legislative waiver, at least for violent offenders. With this in mind, the justice system officials were specifically asked to evaluate the strengths and weaknesses associated with the change in waiver provisions. In terms of strengths, while generally stressing the need for public protection, prosecutors more specifically believed that legislative waiver allows for a greater number of “deserving” cases to be transferred to adult court. Additionally, they felt that once in adult court, these offenders might then be controlled for longer time periods. [Rural/Suburban Prosecutor:] There are too many serious cases involving guns and drugs for juvenile courts to handle. In the past, too many violent offenders were kept in juvenile court. Now, they start off in adult court, where they can be incarcerated for a much longer time. [Small Urban Prosecutor:] The law gets them in adult court from the beginning. . . . In adult court, the tail [time on parole] is the important part of the sentence. I’ll push for a longer tail more so than a longer minimum sentence. We can’t really do that in juvenile court, since supervision ends at 21. [Large Metropolitan Prosecutor:] The problem with the juvenile court is its inability to handle the violent felony. . . . Prior to the new law, there were very few transfers unless the case involved a gun and an injury. Now, many more cases go directly to adult court. The big thing is direct file allows supervision to continue past the age of 21, which can’t happen in juvenile court. In juvenile court, it’s rare to even keep kids locked-up or under supervision until 21. In all likelihood, they’ll commit another crime when they get out, and the biggest benefit to the new law is keeping track of them on parole. Especially state parole, because the ramification of violations under state parole is being sent back to state prison.

172

Excluding Violent Youths From Juvenile Court

A few officials believed that legislative waiver is beneficial in reducing the discretion and disparity that may exist with judicial waiver: [Rural/Suburban Probation Officer:] I think it brought order and understanding to a secretive process. Discretion was removed from the hands of juvenile justice professionals, and placed in the hands of district attorneys, who are elected officials that try to serve public wishes. [Rural/Suburban Judge:] Mandatory policies are tough to deal with as a judge. Under the law, the DA [district attorney] has control of the charges, and judges only get involved based on what the DA charges and will prosecute. But judicial discretion and disparity can be a problem across the state, and the law does address that. While the above comments summarize the perceived strengths of legislative waiver over judicial waiver, a few weaknesses were also presented. Several of those interviewed believed the reduction in juvenile court discretion was not appropriate or advantageous: [Rural/Suburban Public Defender:] Judicial waiver is better, because it allows both the DA and judge to filter cases. Both of them can look at the total circumstances of the case. The DA also had the burden to show cause for transfer. Now they don’t have to make a strong decision. [Small Urban Public Defender:] It should be left to juvenile court judges to make the decision. They have the best information and know the case. [Large Metropolitan Public Defender:] I’m slightly in favor of judicial waiver, because I don’t believe in mandatory sentencing and reducing the discretion of the judiciary. [Small Urban Probation Officer:] With judicial waiver, the judge has much more information and can make a more informed decision. It’s a more controlled atmosphere in juvenile court.

Qualitative Findings

173

[Large Metropolitan Probation Officer:] The old system of letting the juvenile judge determine amenability to treatment has the advantage of knowing what the juvenile court has to offer. . . . The adult system just focuses on incapacitation. In addition to the loss of juvenile court discretion, probably the most often cited weakness associated with the legislative waiver law pertained to youths being released on bail following arrest and subsequently being rearrested for another offense. While not seen as a problem by those interviewed in the rural/suburban county, numerous officials in the other two counties mentioned this issue. [Small Urban Prosecutor:] We had an initial problem with release on bail. . . . Violent cases are likely to be detained and handled faster in juvenile court. We’ve been trying to get the police to push for higher bail after the arrest. [Small Urban Public Defender:] The DJ’s [district justices] tend to set bail too low. They see a young kid with no adult record and let them out. Then, they get out and commit more crime. They wouldn’t be out there if they still were in juvenile court. [Large Metropolitan Public Defender:] The major problem is this: Kids who get arrested and sent to juvenile court have many pre-trial services available to them, which are not available in adult court. In our county, the DA opposes releasing direct files to juvenile pre-trial services. So, he either stays in jail or is released to the street. . . . It’s not rare for direct files out on bail to be rearrested, which leads to multiple charges and cases in both systems. [Small Urban Probation Officer:] The law has had the opposite effect that was intended, because it’s not protecting the community. Bail is set too low, and they’re back on the street without supervision. The case isn’t processed for months or a year later. If they were treated as a juvenile from the beginning, they wouldn’t be out on the street. They would be detained on the spot and then placed.

174

Excluding Violent Youths From Juvenile Court [Small Urban Probation Supervisor:] Bail has been a problem, getting out on bail and committing new crime. Then they get treated as juveniles for the new crime, and fingers get pointed at the juvenile system. [Large Metropolitan Probation Supervisor:] They get released on bail and have no supervision. The adult system also has long delays, so it’s pretty likely they’ll commit more crime. [Large Metropolitan Probation Supervisor:] Juveniles often prefer adult court, because they can make bail and think they can beat the case. It’s not long before they’re in trouble again. [Small Urban Judge:] There was a problem with low bail in direct file cases, with some very serious cases being released to the street. New crime was being committed while they were out on bail. . . . The district justices are reluctant to treat them the same as adults, but we’ve made efforts to change this. [Large Metropolitan Judge:] Cases are processed much faster in juvenile court. The new law was poorly thought out in terms of providing services to high-risk youth. In juvenile court, they’re not left out in the street like in the adult system.

It is noteworthy that in the quantitative portion of this study, we found that violent youths transferred to adult court were significantly more likely to be released from secure custody prior to final disposition than were violent offenders retained in juvenile court. Furthermore, of those who were released, transferred juveniles were significantly more likely to be rearrested during the pre-dispositional time period, and significantly more likely to be rearrested on a violent felony offense, than were nontransferred youths. The above comments from justice system officials correspond well with these findings and suggest that youths excluded from juvenile court under the recent Pennsylvania legislation may pose a serious, immediate threat to public safety.

Reverse Waiver to Juvenile Court As discussed in Chapter 2, Pennsylvania’s recent legislation does allow for cases that are initially excluded from juvenile court to be transferred back to that jurisdiction, if the offender establishes by a

Qualitative Findings

175

preponderance of the evidence that a “reverse waiver” or “decertification” will serve the public interest. In other words, it is the responsibility of the defense to convince a judge to send the case back to juvenile court. With this in mind, the justice system officials interviewed were questioned about the type of cases being returned to juvenile court under the new law. While several factors were mentioned as being important in the decertification process, the use of a firearm during the commission of the offense was seen by most as the key ingredient. [Rural/Suburban Prosecutor:] We only decertify about 10% of our direct files. They usually don’t exhibit a history of violence. . . . I’ve never had a gun case decertified. I think it indicates a higher sophistication and justifies adult court. [Small Urban Prosecutor:] The decertification decision depends a lot on the judge hearing the case, but a gun definitely plays a large role with most judges. [Large Metropolitan Prosecutor:] We’ll maybe recommend the case be sent back to juvenile court if the kid shows dramatic improvement or if there’s a problem in the case. . . If a gun’s involved, it’s not likely to go back to juvenile court. . . . I don’t like sending gunpoint robberies back to mix with the other juvenile cases. [Rural/Suburban Public Defender:] We have very few cases decertified. A gun is absolutely a significant factor. It’s seen as indicating a dangerous offender and may be the number one over-riding factor in prohibiting decertification. [Small Urban Public Defender:] A gun is the key issue in keeping the case in adult court. I don’t even see many cases that don’t involve a gun. [Large Metropolitan Public Defender:] About 20% of our direct file cases are dismissed at the preliminary hearing, and about 38% of the cases that reach a decertification hearing are sent back to juvenile court. Most of the cases that stay in adult court are gunpoint robberies. . . . It’s much harder to get a gun case sent back to juvenile court. . . . Most of the nongun cases

176

Excluding Violent Youths From Juvenile Court would never have been certified under the old law. I really question the effectiveness of nongun cases and direct file. [Small Urban Probation Officer:] A gun involved is a big thing now in keeping the case in adult court. But it also depends on the judge that’s hearing the case. There’s not a lot of uniform decision-making. [Large Metropolitan Probation Supervisor:] The direct files should be restricted to gun cases. They’re the ones that stay in adult court. [Small Urban Judge:] In gun cases, I think juveniles are even more dangerous than adults. Some of them are just coldblooded killers. [Large Metropolitan Judge:] The cases sent back to juvenile court are those with special circumstances: good student, supportive family, effective lawyer, less violent offense, no shooting. . . . The big question is: Does the judge identify with the kid? Or is the kid a bad guy?

Evidence of Deterrence In the quantitative portion of this study, we examined the future criminal behavior of violent youths in juvenile and adult court with an eye toward assessing the specific deterrent effect of transfer. To expand on the issue of deterrence, the justice system officials interviewed were asked to estimate both the general and specific deterrent effects produced through Pennsylvania’s recent legislation. [Rural/Suburban Prosecutor:] There are absolutely no deterrent effects. These kids could care less. They don’t consider any law, and when they get out, they’re likely to commit more crime. That’s why we need to protect society for as long as possible. [Large Metropolitan Prosecutor:] I think the word is getting out. Kids are getting rid of the gun before they get arrested, so they know about the law. But I think the incapacitation effect is stronger in preventing crime.

Qualitative Findings [Rural/Suburban Public Defender:] There’s no general deterrent effect from the law. These crimes often depend on emotions or drugs and alcohol, and the law isn’t going to be thought about when the crime occurs. I’d say 90% of all juveniles don’t even know about the law or don’t appreciate what can happen to them. They don’t think they’re going to get caught, so they don’t care about the penalties. . . . Specific deterrence isn’t likely either. It’s more likely that being put in prison and having a criminal record will increase future crimes. [Small Urban Public Defender:] When they’re in jail, kids say they learned a lesson, because they’re motivated to get out of jail. Once they get out, the lesson and motivation goes away. There’s no thought of the law when kids are on the street. Going to jail is actually a badge of honor, or a rite of passage. They think they can do it standing on their head. [Large Metropolitan Public Defender:] I’ve never believed stricter laws have any deterrent effect on crime. . . . The access to guns these kids have is frightening, and there’s no deterrent effect whatsoever. It’s just not reality that they will consider the law on the street. . . . If we label them criminals and put them in jail, they have no future job possibilities, no chance for rehabilitation, and there’s no way we can expect an improvement once they get out. [Rural/Suburban Probation Officer:] There’s no deterrence. They don’t think that far ahead, and most of them don’t even know about the law. [Rural/Suburban Probation Officer:] I do see some deterrent value. Kids see what happens to their friends. At least in smaller communities, kids see cases in the media of juveniles handcuffed and treated as an adult. Under the old system, it was kept quiet. The deterrence might not work so much for the serious and violent kids, but more so for the less serious offenders. . . . The ones it happens to [those excluded from juvenile court], though, are put in a poorer system. They have to get out sometime, and they’ll be worse than when they went in.

177

178

Excluding Violent Youths From Juvenile Court [Small Urban Probation Officer:] I don’t think kids are aware of the law. If they are, they don’t think they will get caught. [Small Urban Probation Supervisor:] I don’t see any deterrent effect. There’s not much evidence kids think about the law. They have a very short-term perspective and don’t think they will get caught. [Large Metropolitan Probation Supervisor:] The case outcomes in adult court do not lead to deterrence. There’s a low rate of state time, long case processing time, and the conviction rates aren’t much different. [Rural/Suburban Judge:] Juveniles don’t know about the law. They don’t know the consequences. . . . It doesn’t have any effect on behavior. [Large Metropolitan Judge:] Kids don’t think about what they do. They don’t question the penalty. Some don’t even know the penalty, and most don’t care. . . . Deterrence only works for people with something to lose. These kids will never be wearing a paisley tie and talking to someone about their Ph.D. The only thing that’s important is having respect among their friends, and the law isn’t going to make any difference to them. . . . Status is gained in the inner city by being a dominant male. They need to use a gun because respect in the neighborhood is the most important thing. You can’t be a pussy. . . . We need to tell the truth about this: You can’t deter someone from something they have to do.

The above comments suggest that justice system officials see little or no deterrent effects (either general or specific) produced by the recent legislation. The findings from the quantitative portion of this study correspond well with these beliefs, at least in terms of specific deterrence, as the transferred youths exhibited greater recidivism in both the pre- and post-dispositional time periods.

Qualitative Findings

179

Alternatives to Treating Violent Youths as Adults Finally, the justice system officials were questioned about the possibility of other alternatives to treating violent juveniles as adults. Several mentioned the possibility of either altering the jurisdiction of the current juvenile court or creating a separate system for dealing with these youths. [Large Metropolitan Prosecutor:] The main thing we need is special consideration of older teens with serious cases. It might be possible to extend the jurisdiction of the juvenile court past 21, if more resources were made available. [Small Urban Public Defender:] Juvenile placements have more to offer than adult facilities. . . . One option would be to expand the juvenile court foundation and raise the maximum age past 21. A series of facilities and programs for serious offenders could be set up for them to progress through. . . . The problem with juvenile facilities now is that they’re usually short-term, a year or less, and kids are cleaned up but then sent back to the same atmosphere and situation. [Small Urban Probation Supervisor:] There are a few alternatives, like blended sentencing or creating a separate third system. The main thing needed is special facilities for high-risk youth, because the treatment and education services in state and county prisons are lacking. [Large Metropolitan Probation Supervisor:] We should separate serious juveniles from minor ones, but they should also be separated from adult criminals, or they’ll just get a good lesson on how to be criminally active. We need an inbetween system. In addition to these recommendations, a few of those interviewed also stressed the need for early intervention and preventative efforts directed at high-risk children and their parents. [Large Metropolitan Prosecutor:] We need to focus more on younger offenders. We can’t wait until they’re almost adults to

180

Excluding Violent Youths From Juvenile Court do something. In the juvenile system, first offenders usually go home on probation regardless of the seriousness of the offense or family background. There’s just not enough assistance given. [Large Metropolitan Probation Supervisor:] We should pay more attention to young first offenders and their parents. If the parents aren’t in order, the kids will be out of order, and they’ll hurt you as soon as look at you. [Small Urban Judge:] We have nowhere near the resources needed to deal with the increase in juveniles coming over the next 10 years. We need to do something now with younger kids. We can’t wait until they are older to respond to serious crime. It’ll be too late. Early intervention is an absolute must. Parents should also be held accountable for court-ordered interventions.

Summary Several general remarks may be made with regard to the views discussed in this chapter. First, prosecutors seem to most strongly support Pennsylvania’s legislative waiver law. All three that were interviewed favored this approach over judicial waiver. Their support was based on a perceived increase in the number of violent offenders sent to adult court, and the adult court’s ability to provide both longer incarceration and longer incapacitation through parole. In contrast, public defenders and juvenile justice officials were less supportive of the recent legislation. They spoke of a reduction in the consideration of individual circumstances and needs, as well as a lack of resources and services in the adult system. Several also mentioned the possible negative effects of placing youths in jails and prisons with adults. The biggest problem cited with the new law was the ability of excluded youths to make bail, get out on the street, and commit new crime. In addition, very few of those interviewed believed that legislative waiver provides much of a deterrent effect. These observations correspond well with the findings from the quantitative portion of this study.

Qualitative Findings

181

These officials also strongly felt that the use of a firearm distinguished the cases that were not likely to be reverse waived to juvenile court. The presence of a gun appears to suggest a more sophisticated and violent offender. Lastly, several of those interviewed recommended that more specific attention be paid to high-risk youths, perhaps through an extension of juvenile court jurisdiction or creation of a separate system, but also through early intervention and prevention efforts.

This page intentionally left blank

CHAPTER 7

Discussion and Conclusions

In recent years, most states have moved to strengthen the sanctions available for responding to serious and violent youthful offending. Although a variety of “get tough” mechanisms have been adopted, treating juvenile offenders as adults currently appears most popular. Specifically, while many states have expanded the use of judicial and prosecutorial waiver, statutorily excluding certain youths from juvenile court (i.e., legislative waiver) has been embraced as a method for sending increased numbers of adolescents to adult criminal court. This study sought to determine the overall effectiveness of one such legislative waiver policy that targets violent youths. Proponents of transferring juveniles to adult court generally emphasize two perceived advantages with this approach: stronger punishment and greater public safety. Previous studies have raised questions about these assertions, and the current research also provided cause to reconsider the large-scale “criminalization” of delinquent behavior. Overall, while the transferred violent offenders in this study received harsher punishment than did their counterparts retained in juvenile court, the waived youths also exhibited greater recidivism once they were returned to the community. This suggests that the price of any short-term public safety achieved through juvenile transfer might be an increased risk of long-term victimization. Within these general findings, though, are several points that deserve further consideration.

Evidence of a “Custody Gap” A key finding was that, as compared to retained youths, transferred offenders were more likely to be released from secure custody prior to final disposition of their case. In fact, the estimated probability of release for a waived youth was nearly 20 percentage points higher than 183

184

Excluding Violent Youths From Juvenile Court

that of a retained juvenile. This evidence of a “custody gap” corresponded well with the findings obtained through interviews with various justice system officials. In discussing Pennsylvania’s recently enacted legislative waiver law, various officials reported a problem with waived youths posting bail and returning to the street within a short period of time following their initial arrest. These same officials also voiced their concerns about the released offenders committing further crime upon being returned to the community. The quantitative results of this study suggest that these worries too are valid. Of those youths released from secure custody prior to final disposition, the transferred offenders were both more likely to be rearrested and more likely to be rearrested on a violent felony offense during the pre-dispositional time period than were youths retained in juvenile court. It appears, then, that by statutorily excluding violent youths from juvenile court, legislative waiver laws can have the unintended effect of sending offenders back into the community. These adolescents are often released with little or no supervision, and they generally experience longer periods of case processing than do similar youths in juvenile court. In a significant number of cases, the end result is a rearrest on new charges, which many times may include violent felony offenses.

No Evidence of a “Leniency Gap” Although the transferred offenders in this study were more likely to be released from pre-dispositional secure custody than were the retained youths, the subsequent treatment received by the waived youths was consistently of a harsher nature. The transferred offenders were more likely to be convicted, and of those convicted, youths in adult court were more likely to be convicted of a target offense of robbery or aggravated assault. Of the convicted offenders, those who were waived were also more likely to be incarcerated. Of those who were incarcerated, the transferred juveniles experienced longer periods of confinement. These findings contrast with those of earlier studies that found evidence of a “leniency gap” for youths waived to adult court (see Bortner, 1986; Champion, 1989; Emerson, 1981; Gillespie & Norman, 1984; Hamparian et al., 1982; Royscher & Edelman, 1981; Sagatun et al., 1985).

Discussion and Conclusions

185

However, our findings are fairly consistent with those of more recent research that has focused on violent youthful offenders and found that those in adult court experience greater conviction rates (Fagan, 1990; Podkopacz & Feld, 1996; but see also Fagan, 1995; Rudman et al., 1986), greater incarceration rates (Barnes & Franz, 1989; Bishop et al., 1996; Fagan, 1990, 1995; Podkopacz & Feld, 1996; Rudman et al., 1986), and lengthier periods of confinement (Bishop et al., 1996; Fagan, 1990; Podkopacz & Feld, 1996; Rudman et al., 1986; but see also Fagan, 1995; Fritsch et al., 1996). Again, these results also coincide with the opinions of justice system officials who support the use of legislative waiver for violent offenders. District attorneys, in particular, emphasized their increased ability to prosecute serious cases in adult court and thereby secure long-term incarceration. They also noted the added benefit of lengthier periods of parole. The fact that violent youths can be, and seemingly are, punished more severely in adult criminal court may be seen by some as enough reason to justify the expanded use of treating juveniles as adults. With the current “get tough” philosophy, there is strong support for harsher sanctions, particularly if they appear to increase community safety. Because violent youths in adult court seem to be held more accountable and are subjected to greater incapacitation (through both longer periods of incarceration and additional time on parole), politicians and the public alike may continue to join prosecutors in their backing of transfer provisions. However, the limits of this approach should also be noted. Although the transferred offenders in this study experienced longer periods of incarceration than did their counterparts in juvenile court, 57% of the waived youths had their cases disposed and were returned to the community within 4 years of their initial arrest. In other words, a majority of the transferred juveniles were released from incarceration while they were still in their late teens or early twenties. Undoubtedly, many more will be released while they are still young. Although these offenders may undergo strict parole supervision, a central question is whether the short-term incapacitative benefit achieved through juvenile transfer is offset by further criminal activity over the long-term once

186

Excluding Violent Youths From Juvenile Court 74

waived youths are released from confinement. The results of the final area of this study suggest an answer to this question.

Little Evidence of Deterrence Many supporters of treating juvenile offenders as adults suggest that transfer sends notice to violent youths that their behavior will no longer be tolerated, particularly when harsh sanctions are imposed in adult court. Consequently, it is expected that the transfer process will reduce or eliminate future motivations toward offending, thereby further increasing public safety. The results of the current study fail to support this contention. Of the offenders who were returned to the community following final disposition of their case, transferred youths were more likely to be rearrested than were those retained in juvenile court. The waived juveniles also experienced shorter survival times. These findings correspond well with those of three other recent studies (Bishop et al., 1996; Fagan, 1995; Podkopacz & Feld, 1996; see also Winner et al., 1997) that compared the recidivism of transferred juveniles with that of similar youths retained in juvenile court. Furthermore, our interviews with justice system officials revealed little support for the idea that transferring violent offenders to adult court provides either a specific or general deterrent effect. It seems, then, that legislative waiver laws (such as the one recently enacted in Pennsylvania) can realistically be expected to have little or no deterrent utility. In fact, the evidence actually suggests a “brutalization effect,” or that these laws may serve to increase the frequency and seriousness of future offending by those youths who are excluded from juvenile court. However, as noted by Podkopacz and Feld (1996, p. 491), there is more than one possible explanation for the findings of greater recidivism among transferred youths. First, it is possible that more effective treatment services are offered in the juvenile correctional system. In our interviews, public defenders, juvenile court judges, and juvenile probation officials all stressed the need for special resources and programs for violent youthful offenders. The general consensus was that the adult system is 74

Bishop and her colleagues (1996) also alluded to this issue.

Discussion and Conclusions

187

ill prepared to deal with these adolescents. Furthermore, research to date indicates that juveniles in adult prisons and jails receive inferior treatment services and are more often exposed to violent victimization as compared to youths in juvenile facilities (Forst et al., 1989; Howell, 1997; Reddington & Sapp, 1997). At this point, it should be noted that in the current study, while controlling for court of jurisdiction, offenders who were incarcerated for longer periods of time exhibited lesser recidivism during the postdispositional follow-up period than did youths who were either not incarcerated or were confined for shorter time periods. It is not possible for us to determine whether longer incarceration produced a deterrent, rehabilitative, or some combined effect. However, the results clearly suggest that for these violent offenders, lengthier confinement was beneficial. Although there was no significant differential impact of incarceration length for offenders in juvenile versus adult court, when considered with the findings of greater recidivism among transferred youths, it would seem that longer-term treatment in juvenile correctional facilities provides the best chance for future law-abiding behavior among violent adolescents. A second possible reason for the findings of greater recidivism by transferred juveniles is that these youths learned from other adult criminals, or that the adult court and incarceration experience served as a training ground for future criminality. Aside from prosecutors, most of the other justice system officials who were interviewed believed that having adolescents spend time with adult criminals produces adverse consequences. Some might argue that by placing violent youths in the company of older and more seasoned offenders, a forum is provided that serves to enhance the very behavior that is hoped to be prevented. Thirdly, the labeling perspective would seem to provide a sound explanation for why transferred juveniles exhibit greater recidivism than do similar offenders retained in the juvenile system. Early labeling models were criticized for being formulated too simplistically (Gove, 1980; Tittle, 1975), in that a deviant label was viewed as producing a nearly unconditional “self-fulfilling prophecy” (Becker, 1963) of greater deviance. More recent efforts have focused on identifying the intervening factors and contingencies that exist in the labeling process (see, e.g., Braithwaite, 1989; Link, Cullen, Struening, Shrout, & Dohrenwend, 1989; Paternoster & Iovanni, 1989; Sherman, 1993; Tyler, 1990). As discussed below, being sent to adult criminal court can

188

Excluding Violent Youths From Juvenile Court

have a variety of consequences for youthful offenders, all of which may occur prior to any further criminal behavior. None are inevitable, but the degree to which these consequences occur may explain our findings of greater recidivism among transferred youths. To begin, as argued by Paternoster and Iovanni (1989, p. 376), the imposition of a label seems unlikely to amplify future offending if the label is applied and kept in private. While the juvenile court process still tends to be somewhat secretive, when youths are transferred to adult court, they are very likely to be publicly labeled as a criminal. Subsequently, several additional outcomes may take place, either alone or in conjunction with each other. One obvious potential consequence of public labeling is exclusion from conventional activities, such as jobs, school, and other social functions (Braithwaite, 1989; Link, 1982; Link et al., 1989; Paternoster & Iovanni, 1989). In other words, a possible result of public labeling is weakened social bonds (Hirschi, 1969). This is exemplified by Freeman’s (1992) finding that the processing of juveniles in the criminal system (with its public records and probable stigmatization) leads to lower employment rates in the future, especially when a prison sentence is imposed. It is noteworthy that in the current study, youths who were enrolled in school, graduated, or had obtained a GED at the time of referral were more likely to be rearrested, and rearrested in a shorter time period, following disposition than were those who were not enrolled. Although there was no significant differential impact of school status for offenders in juvenile versus adult court, it is feasible that these youths experienced greater recidivism due to such factors as being removed from school, losing a job, or being excluded from future employment following their arrest. Another possible consequence arising from transfer and public labeling is that waived juveniles may perceive that they are being treated unjustly. Recent research has found a sense of injustice over formal sanctioning to be associated with greater future offending, including such diverse behaviors as adolescent drug use, governmental regulation violations, domestic violence, shoplifting, and drunk driving (LanzaKaduce & Radosevich, 1987; Makkai & Braithwaite, 1994; Paternoster et al., 1997; Tyler, 1990). It is conceivable that many youths processed in the adult system perceive unfair treatment and become alienated, which predisposes them to recidivate (see Sherman, 1993). This seems particularly likely if juveniles are physically or mentally mistreated in

Discussion and Conclusions

189

adult facilities, or if they are excluded from conventional activities upon their return to the community. A final potential consequence of transferring and publicly labeling juveniles as criminals is that their personal identities may be altered (Paternoster & Iovanni, 1989). Rooted within labeling theory is the symbolic interactionist tradition, which holds that the application of a deviant label and efforts to formally control deviant acts serve to produce a deviant self-image (Becker, 1963; Lemert, 1972). This, in turn, leads to worsened future behavior. Although labeling juveniles as delinquents and processing them in the juvenile system may create a certain amount of stigmatization, the stigma associated with criminal court processing would appear to have a stronger and more lasting effect on a youth’s personal identity. Moreover, this change in personal identity may be exacerbated by the presence and influence of those who both support and exemplify the status of being a criminal (e.g., adult 75 prison inmates). Overall, then, based on the various intervening factors that can occur following waiver to adult court, modern labeling theorists would undoubtedly concur with the findings of greater recidivism among transferred youths. Although none of the conditions discussed above are certain, they seem much more likely for transferred juveniles than for offenders retained in juvenile court. If all or most do occur, we should not be surprised if youths who are treated as adults continue, and even escalate, their offending in the future. While all of the above explanations for the findings of greater recidivism among transferred youths are plausible, we cannot rule out the possibility that Pennsylvania juvenile courts were simply successful in identifying and transferring the most chronic violent offenders who 75

As discussed by Paternoster and Iovanni (1989), "[Our presentation] reveals clearly the reciprocal nature of the labeling process. Exclusive social reactions from others are likely to lead to an alteration of personal identity and to a closer affinity with deviant others. Such closer association with supportive deviants may lead in turn to additional exclusion (likely also as a result of intentional choice on actor's part) and to a further recasting of self in line with the deviant label" (p. 378).

190

Excluding Violent Youths From Juvenile Court

had the greatest likelihood of subsequent criminal behavior. Although all efforts were made to control for variables that may influence and predict the future behavior of violent youths (e.g., prior offending, age of onset, firearm use), it is possible that other relevant factors were not considered. However, while a certain degree of selection bias may contribute to the results of this study, it is important to point out that this explanation still does not favor the increased use of transferring juveniles to adult court. Legislative waiver laws, such as the one recently enacted in Pennsylvania, seek to greatly reduce or eliminate the discretionary powers of the juvenile court and treat mass quantities of juveniles as adults. None of the explanations discussed above, including the possibility of selection bias, support this practice. If the juvenile court can effectively distinguish the “worst” offenders and subsequently send them to adult court, what is gained by removing this ability and handling broad categories of youthful offenders as adults? At best, the answer appears to be very little. At worst, it would seem that there is much to lose through this practice. While transferred youths may, in general, experience harsher punishment, our findings suggest that only the most chronic and violent (e.g., those with a prior violent felony adjudication and those that use a firearm) are immediately removed from the community and undergo periods of confinement lasting more than a few years. Therefore, it is unlikely that broad legislative waiver policies will provide an increased, long-term, incapacitative effect. Furthermore, our results (and those of a few previous studies) suggest that if legislatures are seeking to deter future youthful offending, there is little reason to expect success. On the other hand, by statutorily excluding violent youths from juvenile court, adolescents may receive inadequate treatment services, be trained by more veteran offenders, experience the negative consequences of a criminal label, or all three.

Policy Implications Contemporary juvenile courts have come under intense criticism, resulting in efforts to “criminalize” delinquent behavior (Feld, 1993; Singer, 1996). In response to these attacks, some have called for the abolishment of a separate juvenile court system (Feld, 1993, 1998), or at least reconceptualizing it as a court for “bankrupt families” (Moore & Wakeling, 1997). Many supporters of the juvenile court are likely to

Discussion and Conclusions

191

resist these efforts, and the results of this study also provide reason for caution in this movement. The findings suggest that excluding broad categories of violent offenders from juvenile court may not produce the expected outcomes and may actually be counterproductive. Therefore, rather than waiving as many youths as possible to adult court (or ending juvenile court jurisdiction altogether), it seems much more practical and beneficial to ensure that most youthful offenders are treated as juveniles and direct our attention toward improving the services provided for them. Alternatives and Alterations To Transfer One justification mentioned for transferring older youths is that they are about to reach the age of criminal court jurisdiction, so juvenile courts should allow adult courts to handle their cases. Research shows that age is an important factor in the waiver process (see Eigen, 1981a, 1981b; Fagan & Deschenes, 1990; Feld, 1989; Podkopacz & Feld, 1996; Poulos & Orchowsky, 1994), and the practice of transferring older juveniles is unquestionably influenced by a perceived need for longer sentences than are available in the juvenile system (Feld, 1989). While arguments have been made that waiver should occur only when an appropriate term of confinement greatly exceeds the period available to a juvenile court judge (Feld, 1989; Zimring, 1991), many may feel that even very young violent offenders should be transferred, due to the finite ability of juvenile courts to maintain custody and supervision. However, other options exist. Although there are some exceptions, most states use 18 as the age at which criminal courts receive jurisdiction over young offenders (DeFrances & Strom, 1997; Sickmund et al., 1997). However, many states also define a maximum age at which the juvenile court can retain custody and supervision beyond the original age of jurisdiction. Juvenile courts in some states may retain control up to age 21 or even longer, such as in California, where the juvenile court may retain jurisdiction for certain offenses up to age 25 (Krisberg & Austin, 1993). These youths may not be tried again for the same offense after reaching age 18, but the juvenile court can maintain supervision or institutionalization. This continuing jurisdiction could be an important factor in dealing with youth violence.

192

Excluding Violent Youths From Juvenile Court

Rather than waiving increased numbers of juveniles (both young and old) to adult court, an alternative in most states would be to raise the maximum age at which the juvenile court can retain jurisdiction (e.g., at least to age 24). This option, which was mentioned by several justice system officials who were interviewed for this study, would allow for lengthier confinement and supervision when needed and would avoid the adverse consequences associated with sending youths to adult court. Furthermore, raising the maximum age at which juvenile courts can maintain jurisdiction would be in accordance with the fact that crime, including violent offending, peaks by late adolescence and declines thereafter (Blumstein, 1988, 1995a, 1995b; Blumstein et al., 1986; Farrington, 1986, 1998). It only seems logical, then, that juvenile courts be able to keep control of known offenders into young adulthood, rather than “cutting them loose” at a time when they are most likely to break the law. A somewhat similar alternative is to create a “blended jurisdiction” between juvenile and adult courts, or an intermediate system for handling serious and violent young offenders. A few justice system officials also alluded to this option, and it is currently being employed in a variety of states (DeFrances & Strom, 1997; Feld, 1998; Sickmund et al., 1997). In general, blended sentencing seeks to achieve proportionality and heighten formal control by combining juvenile and adult sanctions. In sentencing youthful offenders, several states allow adult criminal courts to impose a sanction involving either juvenile or adult correctional services (or both, in some states). Other states have enabled juvenile courts to identify certain offenders for special processing and sanctioning, which may or may not include relocation to adult correctional facilities at some future point. If a goal is to avoid the potential negative impact of transfer to adult court, blended sentencing options that retain most youths in juvenile court appear best. New Mexico, Minnesota, and Texas are examples of states that currently employ versions of this process (Feld, 1998). Although slightly different in various ways, all three approaches focus on serious and violent youths who have not been waived to adult court. While the right to move certain cases to the adult system is preserved, the overriding purpose behind these schemes is to maintain access to juvenile correctional services and provide longer periods of supervision and control under the jurisdiction of the juvenile court.

Discussion and Conclusions

193

Of course, as long as there is some separate system for dealing with juvenile offenders, there will always be a need and desire to treat some of them as adults. Few would argue that there are not certain chronic, violent, youths who, for the sake of public safety, should be removed from society for long periods of time. However, the results of this study indicate that extending the transfer of violent youths beyond those who are the “most deserving” is not good public policy. Several guidelines for treating juveniles as adults are also suggested. While juvenile court judges are likely to retain the ability to waive at least some youths to adult court, it also seems likely that the trend toward statutorily excluding certain offenders will continue. Rather than applying legislative waiver laws to broad categories of offenders, the findings of the current research suggest a more selective approach. Youths of a specified age, such as 14 and older, who are charged with murder will undoubtedly continue to be targeted for automatic transfer (although the possibility of reverse waiver should continue to exist). However, the evidence indicates that the transfer of other violent youths should be reduced in scope. First, concerning age, Pennsylvania’s recent legislation targets violent adolescents aged 15 and older. Of the youths identified for this study, 75% were at least age 16 at the time of referral. Furthermore, of the transferred juveniles, 93% were age 16 or older at the time of referral, and the average age was 16.7 years. In contrast, of the youths retained in juvenile court, only 69% were at least age 16 at the time of referral, and the average age was 16.0 years. These findings correspond with those of previously mentioned research that examined the relationship between age and transfer. In combination with the other results of this study, it appears that a minimum age of 16 for automatically waiving violent offenders would be appropriate. This would ensure that younger adolescents could receive juvenile correctional services and also avoid the potential negative consequences of contact with adult criminals and public labeling in the adult system. Second, instead of including violent offenses committed with any type of deadly weapon under legislative waiver laws, a focus on firearms seems more justified. In this study, youths who employed a gun during their offense were less likely to be released from secure custody prior to final disposition than were those who used some other type of deadly weapon (although there was no differential impact of weapon type in juvenile versus adult court). Firearm users were also

194

Excluding Violent Youths From Juvenile Court

subsequently treated more harshly. Of the convicted youths, those who employed a gun were more likely to be convicted of a target offense of robbery or aggravated assault, and were more likely to be incarcerated, than were nonfirearm users. Of the incarcerated offenders, those who employed a gun experienced longer periods of confinement. Furthermore, the positive effect of firearm use on conviction, conviction on a target offense, and the natural log of incarceration length were significantly greater in adult court over juvenile court. Finally, many of the justice system officials who were interviewed believed that Pennsylvania’s legislative waiver law should be limited to cases involving a firearm, as they are the most likely cases to remain in adult court. From a public safety standpoint, then, statutorily excluding violent firearm users from juvenile court jurisdiction would seem to have the best chance of providing both immediate and longer-term protection. However, it should also be remembered that transferred youths who employed a gun exhibited greater recidivism, both in the pre- and postdispositional time periods, than did firearm users retained in juvenile court. These findings imply that the transferred gun users were the most chronic violent offenders. They also suggest that juvenile courts can effectively deal with at least some youths who resort to firearm violence, while others may require even longer periods of incarceration than they are typically given in adult court. Therefore, identifying some other factor or factors to consider in combination with firearm use may be appropriate in determining which offenders require longer-term periods of incapacitation in the adult system. Not surprisingly, a youth’s prior offending history may supply the needed information. Our results indicated that offenders with more serious and extensive offending backgrounds were given more immediate attention and punished more severely, particularly in adult court. These same youths, though, also exhibited greater recidivism in the post-dispositional follow-up period. This implies that adolescents with substantial histories of offending continued their chronic behavior into young adulthood, despite the imposition of harsher punishment. Based on the findings concerning firearm use and prior offending history, it seems logical to suggest that legislative waiver laws target both of these factors together. In other words, only youths who employ

Discussion and Conclusions

195

a gun and display a notable delinquent background should be excluded 76 from juvenile court. While use of a firearm during the commission of a violent act would appear fairly easy to define, specifying the needed prior record could prove more difficult. One option would be to require 77 a prior adjudication of delinquency on a violent felony offense. Another would be to develop a prior record scoring system, common in adult sentencing guidelines, which could take into account all prior adjudications and even age at first referral. In any event, if a youth’s offending history is going to take a more prominent role in the transfer process (and adult court sanctioning), juvenile records must be improved and made more accessible, at least for those working in the juvenile and criminal justice systems (see Blumstein et al., 1986; Farrington et al., 1986; Feld, 1998). Intervention and Prevention of Youth Violence The recommendations discussed above focus on extending the supervision and control of violent youths who are retained in juvenile court and also identifying the relatively fewer adolescents who require long-term incapacitation in the adult system. However, enabling longer incarceration of violent young offenders is not a complete approach. In combination with these tactics, increased resources should be devoted to improving the custody and services provided for these juveniles, as well as toward delinquency prevention. In our study, it was revealed that longer periods of incarceration were associated with lesser recidivism during the post-dispositional follow-up period, perhaps due to a deterrent, rehabilitative, or some combined effect. Furthermore, incarceration length had a similar beneficial impact for youths in both the juvenile and adult systems. However, this does not mean that widespread confinement of 76

Pennsylvania's recent legislation statutorily excludes both violent offenders who employ a deadly weapon and repeat violent offenders. The current recommendation is that firearm use be combined with prior record to produce two criteria that must be met in order to exclude youths from juvenile court. 77 Of the 557 youths in this study, only 70 both used a firearm during the commission of their offense and had a prior adjudication of delinquency on a violent felony offense. If 15 year olds are excluded, this number is reduced to 54.

196

Excluding Violent Youths From Juvenile Court

delinquents in large-scale correctional facilities should be employed, as this approach has not been found to be very effective in rehabilitating juveniles (Howell, 1997; Krisberg et al., 1995). Instead, a system of graduated sanctions, with an emphasis on early intervention and treatment, appears more justified. Recently, there has been growing support for a “comprehensive strategy” for dealing with serious and violent juvenile offenders (Howell, 1997; Howell et al., 1995; see also Krisberg & Austin, 1993). This strategy is based on research regarding effective correctional programs and the causes and correlates of delinquency. Rather than responding to youth violence after it has escalated to a high level, the evidence suggests that a combination of earlier intervention and prevention programs will be more effective. In discussing the progression of delinquency toward chronic and violent behavior, Krisberg and his colleagues (1995) assert the following: The lack of consistent intervention with juvenile offenders soon after their initial contact with the police or other referring agency has long been recognized as perhaps the single largest gap in services for troubled youth, including serious, violent, and chronic juvenile offenders. . . . Too often, the juvenile justice system’s response early in a young offender’s career is either too much or too little. . . . It gives rise to an all-toocommon pattern: repeated encounters with authorities, perhaps coupled with several short-term detentions, but with no coherent or intensive help provided, culminating in repeated offenses and, ultimately, incarceration in the juvenile and adult systems. (pp. 153-154) In the current study, younger offenders appeared to be given more immediate and serious consideration. However, these youths were already in their mid-teens, and many exhibited a substantial history of offending. Often times, very young juveniles (especially those who are first-offenders) are not given much attention by the juvenile justice system, despite the strong relationship that exists between early onset and later chronic and violent behavior (Blumstein et al., 1986; Farrington, 1986, 1998; Loeber & LeBlanc, 1990; Thornberry et al., 1995). Our quantitative findings also indicated that age at first referral was negatively associated with post-dispositional recidivism, and

Discussion and Conclusions

197

several justice system officials reported a need for greater focus on younger offenders and their parents. Although secure confinement is probably not needed in most cases, more attention should be given to the needs and risk factors associated with the early onset of problem behavior. These needs and risk factors will be discussed shortly in the context of delinquency prevention. In conjunction with early and immediate intervention, an emphasis on individualized treatment and expanding the scope of dispositional alternatives available to the juvenile court appears appropriate. Although the rehabilitative effectiveness of large-scale correctional facilities is minimal at best, recent meta-analyses and other program assessments dispute the claim that “nothing works” (see, e.g., Andrews et al., 1990; Brewer, Hawkins, Catalano, & Neckerman, 1995; Gendreau, Little, & Goggin, 1996; Gendreau & Ross, 1987; Howell, 1997; Krisberg et al., 1995; Lipsey, 1992). The evidence suggests that the most successful correctional programs are those that provide risk and needs assessments, strong case management, and clear and consistent punishment for misconduct. Furthermore, highly structured and smaller community-based facilities that provide intensive services in a secure environment offer the greatest chance for successful treatment of youths who require a structured setting. Finally, an organized system of intermediate sanctions could provide judges with more dispositional options for those young offenders who have not been identified as the most serious and violent. In terms of this latter recommendation, the most effective intermediate sanction programs have been found to be those with a strong treatment component, as well as aftercare services and follow-up supervision (Howell, 1997; Krisberg et al., 1995). In discussing the treatment and confinement of youthful offenders, reason for caution exists with regard to the potential for discrimination. This study uncovered at least suggestive evidence that nonwhites were initially treated more harshly than were whites, and that nonwhites subsequently exhibited greater recidivism in the post-dispositional follow-up period than did whites. This was particularly true for youths retained in juvenile court. These findings may conceivably be explained based on a perception of unfair treatment among nonwhite offenders. It would seem, then, that to maximize the effectiveness of any strategy for dealing with youth violence, further efforts should be made to reduce

198

Excluding Violent Youths From Juvenile Court

the racial bias that appears to operate in the juvenile justice system (see Leonard et al., 1995; Pope & Feyerherm, 1990). Finally, a complete policy for managing violent juvenile behavior must include a prevention component. In this regard, current evidence indicates that the identification of early risk factors is extremely important (Farrington, 1998; Hawkins, Catalano, & Brewer, 1995; Howell, 1997; Howell & Hawkins, 1998; Tremblay & Craig, 1995). These factors can include individual characteristics (e.g., disruptive behavior at a young age, cognitive deficits, low IQ), family influences (e.g., poor parenting, child abuse, parental criminality), school experiences (e.g., academic failure, lack of commitment to school), peer group influences (e.g., delinquent and drug using friends, gangs), and community characteristics (e.g., poverty, presence of drugs and crime, physical deterioration). Various types of programs have been found effective in preventing problem behavior and delinquency, especially those targeting multiple risk factors within the family and school environments (see, e.g., Bry, 1982; Gottfredson, 1986; Kazdin, Siegel, & Bass, 1992; Olds, Henderson, Chamberlin, & Tatelbaum, 1986; Olds & Kitzman, 1993; Tremblay et al., 1992; Yoshikawa, 1994, 1995). These efforts can be enhanced through an emphasis on protective factors that moderate a juvenile’s exposure to risk (Brewer et al., 1995; Farrington, 1998; Howell, 1997; Howell & Hawkins, 1998). These protective factors can include a resilient temperament, pro-social bonding with the family and community, and strong moral beliefs and standards for behavior. Overall, the research on both delinquency prevention and responding to delinquent behavior demonstrates the importance of focused programs and an individualized approach, as well as early intervention.

Recommendations for Future Research This study attempted to provide a comprehensive estimation of the effectiveness of statutorily excluding violent youths from juvenile court. Although transferring juveniles to adult court has been receiving increased attention in the literature, very few studies have examined legislative waiver policies. The current research also improved and expanded upon prior studies by observing a recent cohort of violent offenders, in terms of both their justice system outcomes and future criminal behavior. A variety of statistical techniques and models were

Discussion and Conclusions

199

employed to control for relevant variables and to consider both additive and interactive effects. Finally, qualitative interviews with justice system officials were utilized to further investigate the effectiveness of automatic waiver to adult court. The best test of the effectiveness of transferring juveniles to adult court would be a randomized experiment. Political opposition may preclude this from happening. In the absence of a randomized experiment, future researchers will no doubt continue to employ a variety of matching, cohort, and time series designs. Several new directions are also suggested. First, past studies on juvenile transfer have generally been quantitative in nature and have utilized official records. Ethnographic research with at-risk youths could add to what is known about the effect of transfer policies on adolescent behavior. In addition, interviews with serious and violent youthful offenders, both in the juvenile and adult systems, may provide insight into their perceptions of justice system legitimacy. Information gained from family members could also prove helpful, as well as further knowledge obtained from those working in the system. Lastly, surveys could be used to supplement official data, in order to gather information on a wider variety of variables (e.g., perceptions of unfair treatment, exclusion from conventional activities) and investigate whether findings based on self-reported offending conform to those produced through official measures of recidivism. Second, future research should explore differences in the treatment and experiences of youths in both the juvenile and adult systems. Limited evidence suggests better services in the juvenile system, and the potential for victimization, learning from more knowledgeable criminals, or both, in the adult system. Studies should seek to establish whether these factors (or some other factors) explain differences in subsequent behavior, or whether the findings of greater recidivism among transferred youths are purely a result of selection bias. Finally, adult criminal courts and correctional systems are beginning to respond to the recent influx of youthful offenders. Special facilities are being built (some have opened) for transferred youths, and new programs are constantly being developed. These efforts will provide researchers with natural laboratories to test hypotheses pertaining to the practice of treating juveniles as adults and also allow for the further specification of the effectiveness of this practice.

This page intentionally left blank

References

Aldrich, J. H., & Nelson, F. D. (1984). Linear probability, logit, and probit models (Sage University Paper Series on Quantitative Applications in the Social Sciences, No. 07-045). Beverly Hills, CA: Sage Publications, Inc. Allison, P. D. (1984). Event history analysis: Regression for longitudinal event data (Sage University Paper Series on Quantitative Applications in the Social Sciences, No. 07-046). Newbury Park, CA: Sage Publications, Inc. Andenaes, J. (1968). Does punishment deter crime? Criminal Law Quarterly, 11, 76-93. Andenaes, J. (1974). Punishment and deterrence. Ann Arbor, MI: University of Michigan Press. Andrews, D. A., Zinger, I., Hodge, R. D., Bonta, J., Gendreau, P., & Cullen, F. T. (1990). Does correctional treatment work? A clinically relevant and psychologically informed meta-analysis. Criminology, 28, 369-404. Apospori, E., Alpert, G., & Paternoster, R. (1992). The effect of involvement with the criminal justice system: A neglected dimension of the relationship between experience and perceptions. Justice Quarterly, 9 (3), 379-392. Bachman, R., & Paternoster, R. (1997). Statistical methods for criminology and criminal justice. New York: McGraw-Hill. Bachman, R., Paternoster, R., & Ward, S. (1992). The rationality of sexual offending: Testing a deterrence/rational choice conception of sexual assault. Law and Society Review, 26 (2), 343-372. Baker, H. H. (1979). Procedure of the Boston juvenile court. In F. L. Faust & P. J. Brantingham (Eds.), Juvenile justice philosophy: nd Readings, cases, and comments (2 ed., pp. 146-155). St. Paul, MN: West Publishing. 201

202

References

Barnes, C. W., & Franz, R. S. (1989). Questionably adult: Determinants and effects of the juvenile waiver decision. Justice Quarterly, 6 (1), 117-135. Baumer, E. (1997). Levels and predictors of recidivism: The Malta experience. Criminology, 35 (4), 601-628. Beccaria, C. (1986). On crimes and punishments (D. Young, Trans.). Indianapolis, IN: Hackett Publishing. (Original work published 1764) Becker, H. S. (1963). Outsiders. New York: Free Press. Benedict, W. R., & Huff-Corzine, L. (1997). Return to the scene of the punishment: Recidivism of adult male property offenders on felony probation, 1986-1989. Journal of Research in Crime and Delinquency, 34 (2), 237-252. Bernard, T. J. (1992). The cycle of juvenile justice. New York: Oxford University Press. Bishop, D. M., & Frazier, C. E. (1991). Transfer of juveniles to criminal court: A case study and analysis of prosecutorial waiver. Notre Dame Journal of Law, Ethics & Public Policy, 5, 281-302. Bishop, D. M., Frazier, C. E., & Henretta, J. C. (1989). Prosecutorial waiver: Case study of a questionable reform. Crime & Delinquency, 35 (2), 179-201. Bishop, D. M., Frazier, C. E., Lanza-Kaduce, L., & Winner, L. (1996). The transfer of juveniles to criminal court: Does it make a difference? Crime & Delinquency, 42 (2), 171-191. Biskup, M. D., & Cozic, C. P. (1992). Youth violence. San Diego, CA: Greenhaven Press, Inc. Blumstein, A. (1988). Prison populations: A system out of control? In M. Tonry (Ed.), Crime and justice: An annual review of research (Vol. 10, pp. 231-265). Chicago: University of Chicago Press. Blumstein, A. (1995a). Prisons. In J. Q. Wilson, & J. Petersilia (Eds.), Crime (pp. 387-419). San Francisco: Institute for Contemporary Studies. Blumstein, A. (1995b). Youth violence, guns, and the illicit-drug industry. The Journal of Criminal Law and Criminology, 86 (1), 10-36. Blumstein, A., Cohen, J., Martin, S. E., & Tonry, M. (1983). Research on sentencing: The search for reform. Washington, DC: National Academy Press.

References

203

Blumstein, A., Cohen, J., & Nagin, D. (1978). Deterrence and incapacitation: Estimating the effects of criminal sanctions on crime rates. Washington, DC: National Academy of Sciences. Blumstein, A., Cohen, J., Roth, J., & Visher, C. (1986). Criminal careers and “career criminals.” Washington, DC: National Academy Press. Bortner, M. A. (1986). Traditional rhetoric, organizational realities: Remand of juveniles to adult court. Crime & Delinquency, 32 (1), 53-73. Braithwaite, J. (1989). Crime, shame, and reintegration. Cambridge: Cambridge University Press. Brame, R., Paternoster, R., Mazerolle, P., & Piquero, A. (1998). Testing for the equality of maximum likelihood regression coefficients between two independent equations. Journal of Quantitative Criminology, 14, 245-261. Breed v. Jones, 4. U. S. 5. (1975). Brewer, D. D., Hawkins, J. D., Catalano, R. F., & Neckerman, H. J. (1995). Preventing serious, violent, and chronic offending: A review of evaluations of selected strategies in childhood, adolescence, and the community. In J. C. Howell, B. Krisberg, J. D. Hawkins, & J. J. Wilson (Eds.), A sourcebook: Serious, violent, and chronic juvenile offenders (pp. 61-141). Thousand Oaks, CA: Sage Publications, Inc. Bridges, G. S., & Stone, J. A. (1986). Effects of criminal punishment on perceived threat of punishment: Toward an understanding of specific deterrence. Journal of Research in Crime and Delinquency, 23 (3), 207-239. Brown, W. K., Miller, T. P., & Jenkins, R. L. (1987). The favorable effect of juvenile court adjudication of delinquent youth on the first contact with the juvenile justice system. Juvenile and Family Court Journal, 38, 21-26. Brown, W. K., Miller, T. P., Jenkins, R. L., & Rhodes, W. A. (1989). The effect of early juvenile court adjudication on adult outcomes. International Journal of Offender Therapy and Comparative Criminology, 33, 177-183. Brown, W. K., Miller, T. P., Jenkins, R. L., & Rhodes, W. A. (1991). The human costs of “Giving the kid another chance.” International Journal of Offender Therapy and Comparative Criminology, 35 (4), 296-302.

204

References

Bry, B. H. (1982). Reducing the incidence of adolescent problems through preventive intervention. American Journal of Community Psychology, 10, 265-276. Butts, J. A. (1996). Speedy trial in the juvenile court. American Journal of Criminal Law, 23, 515-560. Butts, J. A. (1997). Necessarily relative: Is juvenile justice speedy enough? Crime & Delinquency, 43 (1), 3-23. Butts, J. A., & Halemba, G. J. (1994). Delays in juvenile justice: Findings from a national survey. Juvenile and Family Court Journal, 45, 31-46. Campbell, D. T., & Stanley, J. C. (1966). Experimental and quasiexperimental designs for research. In N. L. Gage (Ed.), Handbook of research on teaching (pp. 1-76). Chicago: Rand McNally. Champion, D. J. (1989). Teenage felons and waiver hearings: Some recent trends, 1980-1988. Crime & Delinquency, 35 (4), 577-585. Champion, D. J., & Mays, G. L. (1991). Transferring juveniles to criminal courts: Trends and implications for criminal justice. New York: Praeger Publishers. Chiricos, T. G., & Waldo, G. P. (1970). Punishment and crime: An examination of some empirical evidence. Social Problems, 18, 200-217. Clarke, E. E. (1996). A case for reinventing juvenile transfer. Juvenile and Family Court Journal, 47, 3-22. Clogg, C. C., Petkova, E., & Haritou, A. (1995). Statistical methods for comparing regression coefficients between models. American Journal of Sociology, 100 (5), 1261-1293. Coates, R. B., Miller, A. B., & Ohlin, L. E. (1978). Diversity in a youth correctional system: Handling delinquents in Massachusetts. Cambridge, MA: Ballinger. Cohen, L. (1978). Sanction threats and violation behavior: An inquiry into perceptual variation. In C. Wellford (Ed.), Quantitative studies in criminology (pp. 84-99). Beverly Hills, CA: Sage Publications, Inc. Cook, P. J., and Laub, J. H. (1998). The unprecedented epidemic in youth violence. In M. Tonry & M. H. Moore (Eds.), Crime and justice: A review of research (Vol. 24, pp. 27-64). Chicago: University of Chicago Press. Cook, T. D., & Campbell, D. T. (1979). Quasi experimentation: Design and analysis issues for field settings. Chicago: Rand McNally.

References

205

Cornish, D. B., & Clarke, R. V. (1986). The reasoning criminal: Rational choice perspectives on offending. New York: SpringerVerlag. Cornish, D. B., & Clarke, R. V. (1987). Understanding crime displacement: An application of rational choice theory. Criminology, 25 (4), 933-947. Cox, D. R. (1972). Regression models and life tables. Journal of the Royal Statistics Society, Series B, 34, 187-202. Creswell, J. W. (1994). Research design: Qualitative & quantitative approaches. Thousand Oaks, CA: Sage Publications, Inc. DeFrances, C. J., & Strom, K. J. (1997). Juveniles prosecuted in state criminal courts. (Report No. NCJ-164265). Washington, DC: Bureau of Justice Statistics/Office of Juvenile Justice and Delinquency Prevention. DeJong, C. (1997). Survival analysis and specific deterrence: Integrating theoretical and empirical models of recidivism. Criminology, 35 (4), 561-575. DiIulio, J. J. (1995). The coming of the super-predators. Weekly Standard, 1, 23-28. DiIulio, J. J. (1996). They’re coming: Florida’s youth crime bomb. Impact, 25-27. DiIulio, J. J., Smith, S. K., & Saiger, A. J. (1995). The federal role in crime control. In J. Q. Wilson & J. Petersilia (Eds.), Crime (pp. 445-462). San Francisco: Institute for Contemporary Studies. Eigen, J. P. (1981a). The determinants and impact of jurisdictional transfer in Philadelphia. In J. C. Hall, D. M. Hamparian, J. M. Pettibone, & J. L. White (Eds.), Major issues in juvenile justice information and training: Readings in public policy (pp. 333-350). Columbus, OH: Academy for Contemporary Problems. Eigen, J. P. (1981b). Punishing youth homicide offenders in Philadelphia. The Journal of Criminal Law and Criminology, 72 (3), 1072-1093. Emerson, R. (1981). On last resorts. American Journal of Sociology, 87, 1-22. Empey, L. T., & Erickson, M. L. (1972). The Provo experiment: Evaluating community control of delinquency. Lexington, MA: D. C. Heath.

206

References

Empey, L. T., & Stafford, M. C. (1991). American delinquency: Its meaning & construction. Belmont, CA: Wadsworth Publishing Company. Fagan, J. (1990). Social and legal policy dimensions of violent juvenile crime. Criminal Justice and Behavior, 17 (1), 93-133. Fagan, J. (1995). Separating the men from the boys: The comparative advantage of juvenile versus criminal court sanctions on recidivism among adolescent felony offenders. In J. C. Howell, B. Krisberg, J. D. Hawkins, & J. J. Wilson (Eds.), A sourcebook: Serious, violent, & chronic juvenile offenders (pp. 238-260). Thousand Oaks, CA: Sage Publications, Inc. Fagan, J., & Deschenes, E. P. (1990). Determinants of judicial waiver decisions for violent juvenile offenders. The Journal of Criminal Law and Criminology, 81 (2), 314-347. Fagan, J., Forst, M., & Vivona, T. S. (1987). Racial determinants of the judicial transfer decision: Prosecuting violent youth in criminal court. Crime & Delinquency, 33 (2), 259-286. Fagan, J., & Wilkinson, D. L. (1998). Guns, youth violence, and social identity in inner cities. In M. Tonry & M. H. Moore (Eds.), Crime and justice: A review of research (Vol. 24, pp. 105-188). Chicago: University of Chicago Press. Farrington, D. P. (1977). The effects of public labelling. British Journal of Criminology, 17 (2), 112-125. Farrington, D. P. (1983). Randomized experiments in crime and justice. In M. Tonry & N. Morris (Eds.), Crime and justice: An annual review of research (Vol. 4, pp. 257-308). Chicago: University of Chicago Press. Farrington, D. P. (1986). Age and crime. In M. Tonry & N. Morris (Eds.), Crime and justice: An annual review of research (Vol. 7, pp. 189-250). Chicago: The University of Chicago Press. Farrington, D. P. (1998). Predictors, causes, and correlates of male youth violence. In M. Tonry & M. H. Moore (Eds.), Crime and justice: A review of research (Vol. 24, pp. 421-475). Chicago: University of Chicago Press. Farrington, D. P., Ohlin, L., & Wilson, J. Q. (1986). Understanding and controlling crime: Toward a new research strategy. New York: Springer-Verlag. Farrington, D. P., Osborn, S. G., & West, D. J. (1978). The persistence of labeling effects. British Journal of Criminology, 18, 277-284.

References

207

Faust, F. L., & Brantingham, P. J. (1979). Juvenile justice philosophy. St. Paul, MN: West Publishing. Federal Bureau of Investigation. (1989-1998). Uniform crime reports for the United States. Washington, DC: U.S. Government Printing Office. Feld, B. C. (1987). The juvenile court meets the principle of the offense: Legislative changes in juvenile waiver statutes. The Journal of Criminal Law and Criminology, 78 (3), 471-533. Feld, B. C. (1989). Bad law makes hard cases: Reflections on teen-aged axe-murderers, judicial activism, and legislative default. Journal of Law and Inequality, 8 (1), 1-101. Feld, B. C. (1993). Criminalizing the American juvenile court. In M. Tonry (Ed.), Crime and justice: A review of research (Vol. 17, pp. 197-280). Chicago: University of Chicago Press. Feld, B. C. (1998). Juvenile and criminal justice systems’ responses to youth violence. In M. Tonry & M. H. Moore (Eds.), Crime and justice: A review of research (Vol. 24, pp. 189-261). Chicago: University of Chicago Press. Fields, G., & Overberg, P. (1998, March 26). Juvenile homicide arrest rate on rise in rural USA. USA Today, p. A11. Forst, M. L., & Blomquist, M. E. (1991). Cracking down on juveniles: The changing ideology of youth corrections. Notre Dame Journal of Law, Ethics & Public Policy, 5, 323-375. Forst, M., Fagan, J., & Vivona, T. S. (1989). Youths in prisons and training schools: Perceptions and consequences of the treatmentcustody dichotomy. Juvenile and Family Court Journal, 40, 1-14. Frazier, C. E., & Bishop, D. M. (1995). Reflections on race effects in juvenile justice. In K. K. Leonard, C. E. Pope, & W. H. Feyerherm (Eds.), Minorities in juvenile justice (pp. 16-46). Thousand Oaks, CA: Sage Publications, Inc. Freeman, R. B. (1992). Crime and the employment of disadvantaged youth. In G. Peterson, & W. Vroman (Eds.), Urban labor markets and job opportunity (pp. 201-238). Washington, DC: Urban Institute Press. Fritsch, E. J., Caeti, T. J., & Hemmens, C. (1996). Spare the needle but not the punishment: The incarceration of waived youth in Texas prisons. Crime & Delinquency, 42 (4), 593-609.

208

References

Fritsch, E., & Hemmens, C. (1995). Juvenile waiver in the United States 1979-1995: A comparison and analysis of state waiver statutes. Juvenile and Family Court Journal, 46, 17-35. Geerken, M. R. (1994). Rap sheets in criminological research: Considerations and caveats. Journal of Quantitative Criminology, 10 (1), 3-21. Geerken, M. R., & Gove, W. R. (1975). Deterrence: Some theoretical considerations. Law and Society Review, 9, 498-513. Gendreau, P., Little, T., & Goggin, C. (1996). A meta-analysis of the predictors of adult offender recidivism: What works! Criminology, 34 (4), 575-607. Gendreau, P., & Ross, R. R. (1987). Revivification of rehabilitation: Evidence from the 1980s. Justice Quarterly, 4 (3), 349-407. Gibbs, J. P. (1968). Crime, punishment, and deterrence. Southwestern Social Science Quarterly, 48, 515-530. Gibbs, J. P. (1975). Crime, punishment, and deterrence. New York: Elsevier. Gold, L. K., & Norman, M. D. (1984). Does certification mean prison: Some preliminary findings from Utah. Juvenile and Family Court Journal, 35, 23-34. Gold, M., & Williams, J. R. (1969). National study of the aftermath of apprehension. Prospectus, 3, 3-19. Gottfredson, D. C. (1986). An empirical test of school-based environmental and individual interventions to reduce the risk of delinquent behavior. Criminology, 24, 705-731. Gottfredson, D. C., & Barton, W. H. (1993). Deinstitutionalization of juvenile offenders. Criminology, 31 (4), 591-607. Gove, W. R. (1980). The labelling perspective: An overview. In W. R. nd Gove (Ed.), The labelling of deviance (2 ed., pp. 9-33). Beverly Hills: Sage Publications, Inc. Grasmick, H. G., & Bursik, R. J. (1990). Conscience, significant others, and rational choice: Extending the deterrence model. Law and Society Review, 24 (3), 837-861. Grasmick, H. G., Bursik, R. J., & Arneklev, B. J. (1993). Reduction in drunk driving as a response to increased threats of shame, embarrassment, and legal sanctions. Criminology, 31 (1), 41-67. Greenbaum, S. (1997). Kids and guns: From playgrounds to battlegrounds. Juvenile Justice, III (2), 3-10.

References

209

Greenwood, P. W. (1995). Juvenile crime and juvenile justice. In J. Q. Wilson & J. Petersilia (Eds.), Crime (pp. 91-117). San Francisco: Institute for Contemporary Studies. Hagan, J., & Bumiller, K. (1983). Making sense of sentencing: A review and critique of sentencing research. In A. Blumstein, J. Cohen, S. E. Martin, & M. Tonry (Eds.), Research on sentencing: The search for reform (pp. 1-54). Washington, DC: National Academy Press. Hamparian, D., Estep, L. K., Muntean, S. M., Prestino, R., Swisher, R. G., Wallace, P. L., & White, J. L. (1982). Youth in adult court: Between two worlds. Columbus, OH: Academy for Contemporary Problems. Hawkins, J. D., Catalano, R. F., & Brewer, D. D. (1995). Preventing serious, violent, and chronic offending: Effective strategies from conception to age 6. In J. C. Howell, B. Krisberg, J. D. Hawkins, & J. J. Wilson (Eds.), A sourcebook: Serious, violent, and chronic juvenile offenders (pp. 36-60). Thousand Oaks, CA: Sage Publications, Inc. Hirschi, T. (1969). Causes of delinquency. Berkeley: University of California Press. Holden, R. T. (1985). Failure time models for thinned crime commission data. Sociological Methods and Research, 14, 3-30. Horney, J., & Marshall, I. H. (1992). Risk perceptions among serious offenders: The role of crime and punishment. Criminology, 30 (4), 575-594. Houghtalin, M., & Mays, G. L. (1991). Criminal dispositions of New Mexico juveniles transferred to adult court. Crime & Delinquency, 37 (3), 393-407. Howell, J. C. (1997). Juvenile justice and youth violence. Thousand Oaks, CA: Sage Publications, Inc. Howell, J. C., & Hawkins, J. D. (1998). Prevention of youth violence. In M. Tonry & M. H. Moore (Eds.), Crime and justice: A review of research (Vol. 24, pp. 263-315). Chicago: University of Chicago Press. Howell, J. C., Krisberg, B., Hawkins, J. D., & Wilson, J. J. (1995). A sourcebook: Serious, violent, & chronic juvenile offenders. Thousand Oaks, CA: Sage Publications, Inc.

210

References

Howell, J. C., Krisberg, B., & Jones, M. (1995). Trends in juvenile crime and youth violence. In J. C. Howell, B. Krisberg, J. D. Hawkins, & J. J. Wilson (Eds.), A sourcebook: Serious, violent, & chronic juvenile offenders (pp. 1-35). Thousand Oaks, CA: Sage Publications, Inc. In re Gault, 3. U. S. 1. (1967). In re Winship, 3. U. S. 3. (1970). Jacobs, M. D. (1990). Screwing the system and making it work. Chicago: University of Chicago Press. Jensen, E. J., & Metsger, L. K. (1994). A test of the deterrent effect of legislative waiver on violent juvenile crime. Crime & Delinquency, 40 (1), 96-104. Johnson, B. D., Williams, T., Dei, K. A., & Sanabria, H. (1990). Drug abuse in the inner city: Impact on hard-drug users and the community. In M. Tonry & J. Q. Wilson (Eds.), Crime and justice: A review of research (Vol. 13, pp. 9-67). Chicago: The University of Chicago Press. Joo, H. J., Ekland-Olson, S., & Kelly, W. R. (1995). Recidivism among paroled property offenders released during a period of prison reform. Criminology, 33 (3), 389-410. Kazdin, A. E., Siegel, T. C., & Bass, D. (1992). Cognitive problem solving skills training and parent management training in the treatment of antisocial behavior in children. Journal of Consulting and Clinical Psychology, 60, 733-747. Keiter, R. B. (1973). Criminal or delinquent? A study of juvenile cases transferred to the criminal court. Crime & Delinquency, 19, 528538. Kent v. United States, 3. U. S. 5. (1966). Kinder, K., Veneziano, C., Fichter, M., & Azuma, H. (1995). A comparison of the dispositions of juvenile offenders certified as adults with juvenile offenders not certified. Juvenile and Family Court Journal, 46, 37-42. Klein, M. W. (1986). Labeling theory and delinquency policy: An experimental test. Criminal Justice and Behavior, 13 (1), 47-79. Klemke, L. W. (1978). Does apprehension for shoplifting amplify or terminate shoplifting activity? Law and Society Review, 12, 391403.

References

211

Klepper, S., & Nagin, D. (1989). The deterrent effect of perceived certainty and severity of punishment revisited. Criminology, 27 (4), 21-46. Krisberg, B., & Austin, J. F. (1993). Reinventing juvenile justice. Newbury Park, CA: Sage Publications, Inc. Krisberg, B., Austin, J. F., & Steele, P. A. (1989). Unlocking juvenile corrections. San Francisco: National Council on Crime and Delinquency. Krisberg, B., Currie, E., Onek, D., & Wiebush, R. G. (1995). Graduated sanctions for serious, violent, and chronic juvenile offenders. In J. C. Howell, B. Krisberg, J. D. Hawkins, & J. J. Wilson (Eds.), A sourcebook: Serious, violent, & chronic juvenile offenders (pp. 142170). Thousand Oaks, CA: Sage Publications, Inc. Lanza-Kaduce, L. (1988). Perceptual deterrence and drinking and driving among college students. Criminology, 26 (2), 321-341. Lanza-Kaduce, L., & Radosevich, M. J. (1987). Negative reactions to processing and substance use among young incarcerated males. Deviant Behavior, 8, 137-148. Lattimore, P. K., Linster, R. L., & MacDonald, J. M. (1997). Risk of death among serious young offenders. Journal of Research in Crime and Delinquency, 34 (2), 187-209. Lattimore, P. K., Visher, C. A., & Linster, R. L. (1995). Predicting rearrest for violence among serious youthful offenders. Journal of Research in Crime and Delinquency, 32 (1), 54-83. Lee, L. (1994). Factors determining waiver in a juvenile court. Journal of Criminal Justice, 22 (4), 329-339. Lemert, E. M. (1951). Social pathology: A systematic approach to the theory of sociopathic behavior. New York: McGraw-Hill. Lemert, E. M. (1972). Human deviance, social problems and social nd control (2 ed.). Englewood Cliffs, NJ: Prentice-Hall. Lemmon, J. H., Sontheimer, H., & Saylor, K. A. (1991). A study of Pennsylvania juveniles transferred to criminal court in 1986. Harrisburg, PA: The Pennsylvania Juvenile Court Judges’ Commission. Leonard, K. K., Pope, C. E., & Feyerherm, W. H. (1995). Minorities in juvenile justice. Thousand Oaks, CA: Sage Publications, Inc. Lewis-Beck, M. S. (1980). Applied regression: An introduction (Sage University Paper Series on Quantitative Applications in the Social Sciences, No. 07-022). Newbury Park, CA: Sage Publications, Inc.

212

References

Link, B. (1982). Mental patient status, work and income: An examination of the effects of a psychiatric label. American Sociological Review, 47, 202-215. Link, B., Cullen, F. T., Struening, E., Shrout, P., & Dohrenwend, B. P. (1989). A modified labeling theory approach to mental disorders: An empirical assessment. American Sociological Review, 54, 400423. Lipsey, M. W. (1992). Juvenile delinquency treatment: A meta-analytic inquiry into the variability of effects. In D. T. Cook, H. Cooper, D. S. Cordray, H. Hartman, L. V. Hedges, R. J. Knight, T. A. Louis, & F. Mosteller (Eds.), Meta-analysis for explanation (pp. 83-127). New York: Russell Sage Foundation. Lipton, D., Martinson, R., & Wilks, J. (1975). The effectiveness of correctional treatment: A survey of treatment evaluation studies. New York: Praeger Publishers. Loeber, R., & LeBlanc, M. (1990). Toward a developmental criminology. In M. Tonry & N. Morris (Eds.), Crime and justice: An annual review of research (Vol. 12, pp. 375-473). Chicago: University of Chicago Press. Logan, C. H. (1975). Arrest rates and deterrence. Social Science Quarterly, 56, 376-389. Mahoney, A. R. (1974). The effect of labeling upon youths in the juvenile justice system: A review of the evidence. Law and Society Review, 8, 583-614. Makkai, T., & Braithwaite, J. (1994). The dialectics of corporate deterrence. Journal of Research in Crime and Delinquency, 31, 347-373. Maltz, M. D. (1984). Recidivism. Orlando, FL: Academic Press. Martinson, R. (1974). What works? Questions and answers about prison reform. The Public Interest, 35, 22-54. Masters, B. A., & Shear, M. D. (1998, April 5). As suburbia surges, violence tags along. The Washington Post, pp. A1, A20, A21. Matsueda, R. L. (1992). Reflected appraisals, parental labeling, and delinquency: Specifying a symbolic interactionist theory. American Journal of Sociology, 97 (6), 1577-1611. McCoy, C. R., Lowe, H., & Gibbons, T. J. (1998, March 22). Very public death shines a light on murder in the city. The Philadelphia Inquirer, pp. A1, A26.

References

213

McDowall, D., Loftin, C., & Wiersema, B. (1992). A comparative study of the effects of mandatory sentencing laws for gun crimes. The Journal of Criminal Law and Criminology, 83 (2), 378-394. McKeiver v. Pennsylvania, 4. U. S. 5. (1971). Meddis, S. V. (1993, October 29). Poll: Treat juveniles the same as adult offenders. USA Today, pp. A1, A11. Menard, S. (1995). Applied logistic regression analysis (Sage University Paper Series on Quantitative Applications in the Social Sciences, No. 07-106). Thousand Oaks, CA: Sage Publications, Inc. Miller, S. L., & Iovanni, L. (1994). Determinants of perceived risk of formal sanction for courtship violence. Justice Quarterly, 11 (2), 281-312. Minor, K. I., Hartmann, D. J., & Terry, S. (1997). Predictors of juvenile court actions and recidivism. Crime & Delinquency, 43 (3), 328344. Minor, W. W., & Harry, J. P. (1982). Deterrent and experiential effects in perceptual research: A replication and extension. Journal of Research in Crime and Delinquency, 19, 190-203. Moore, M. H., & Tonry, M. (1998). Youth violence in America. In M. Tonry & M. H. Moore (Eds.), Crime and justice: A review of research (Vol. 24, pp. 1-26). Chicago: University of Chicago Press. Moore, M. H., & Wakeling, S. (1997). Juvenile justice: Shoring up the foundations. In M. Tonry (Ed.), Crime and justice: A review of research (Vol. 22, pp. 253-301). Chicago: University of Chicago Press. More teens report seeing gangs in school: Survey also finds big rise in students citing victimization by violent crime. (1998, April 13). The Washington Post, p. A4. Murray, C. A., & Cox, L. A. (1979). Beyond probation: Juvenile corrections and the chronic delinquent. Beverly Hills, CA: Sage Publications, Inc. Myers, D. (1997). Excluding juvenile offenders from juvenile court: The effects of Pennsylvania’s recent legislation. Paper presented at the annual meeting of the Academy of Criminal Justice Sciences, Louisville, KY.

214

References

Nagin, D. (1978). General deterrence: A review of the empirical evidence. In A. Blumstein, J. Cohen, & D. Nagin (Eds.), Deterrence and incapacitation: Estimating the effects of criminal sanctions on crime rates (pp. 95-139). Washington, DC: National Academy of Sciences. Nagin, D., & Paternoster, R. (1993). Enduring individual differences and rational choice theories of crime. Law and Society Review, 27 (3), 467-496. Neter, J., Wasserman, W., & Kutner, M. H. (1989). Applied linear nd regression models (2 ed.). Burr Ridge, IL: Richard D. Irwin, Inc. Olds, D. L., Henderson, C. R., Chamberlin, R., & Tatelbaum, R. (1986). Preventing child abuse and neglect: A randomized trial of nurse home visitation. Pediatrics, 78, 65-78. Olds, D. L., & Kitzman, H. (1993). Review of research on home visiting for pregnant women and parents of young children. The Future of Children, 3 (3), 53-92. Osbun, L. A., & Rode, P. A. (1984). Prosecuting juveniles as adults: The quest for “objective” decisions. Criminology, 22 (2), 187-202. Paternoster, R. (1987). The deterrent effect of the perceived certainty and severity of punishment: A review of the evidence and issues. Justice Quarterly, 4 (2), 173-217. Paternoster, R. (1989a). Absolute and restrictive deterrence in a panel of youth: Explaining the onset, persistence/desistence, and frequency of delinquent offending. Social Problems, 36 (3), 289309. Paternoster, R. (1989b). Decisions to participate in and desist from four types of common delinquency: Deterrence and the rational choice perspective. Law & Society Review, 23 (1), 7-40. Paternoster, R., Brame, R., Bachman, R., & Sherman, L. W. (1997). Do fair procedures matter? The effect of procedural justice on spouse assault. Law and Society Review, 31, 163-204. Paternoster, R., Brame, R., Mazerolle, P., & Piquero, A. (1998). Using the correct statistical test for the equality of regression coefficients. Criminology, 36 (4), 859-866. Paternoster, R., & Iovanni, L. (1986). The deterrent effect of perceived severity: A reexamination. Social Forces, 64, 751-777. Paternoster, R., & Iovanni, L. (1989). The labeling perspective and delinquency: An elaboration of the theory and an assessment of the evidence. Justice Quarterly, 6 (3), 359-394.

References

215

Paternoster, R., & Piquero, A. (1995). Reconceptualizing deterrence: An empirical test of personal and vicarious experiences. Journal of Research in Crime and Delinquency, 32 (3), 251-286. Paternoster, R., Saltzman, L. E., Chiricos, T. G., & Waldo, G. P. (1982). Perceived risk and deterrence: Methodological artifacts in perceptual deterrence research. Journal of Criminal Law and Criminology, 73, 1238-1258. Paternoster, R., Saltzman, L. E., Waldo, G. P., & Chiricos, T. G. (1983). Estimating perceptual stability and deterrent effects: The role of perceived legal punishment in the inhibition of criminal involvement. Journal of Criminal Law and Criminology, 74, 270297. Paternoster, R., Saltzman, L. E., Waldo, G. P., & Chiricos, T. G. (1985). Assessments of risk and behavioral experience: An exploratory study of change. Criminology, 23 (3), 417-436. Pennsylvania Juvenile Court Judges’ Commission. (1992). The Juvenile Act. 42 PA C.S. Sec. 6301 et seq. Harrisburg, PA: Juvenile Court Judges’ Commission. Pennsylvania Juvenile Court Judges’ Commission. (1996). The Juvenile Act. 42 PA C.S. Sec. 6301 et seq. Harrisburg, PA: Juvenile Court Judges’ Commission. Piliavin, I., Gartner, R., Thorton, C., & Matsueda, R. (1986). Crime, deterrence, and rational choice. American Sociological Review, 51, 101-119. Pindyck, R. S., & Rubinfeld, D. L. (1981). Econometric models and nd economic forecasts (2 ed.). New York: McGraw-Hill. Piquero, A., & Tibbetts, S. (1996). Specifying the direct and indirect effects of low self-control and situational factors in offenders’ decision making: Toward a more complete model of rational offending. Justice Quarterly, 13 (3), 481-510. Platt, A. (1969). The child savers. Chicago: University of Chicago Press. Podkopacz, M. R., & Feld, B. C. (1995). Judicial waiver policy and practice: Persistence, seriousness and race. Journal of Law and Inequality, 14 (1), 73-178. Podkopacz, M. R., & Feld, B. C. (1996). The end of the line: An empirical study of judicial waiver. The Journal of Criminal Law and Criminology, 86 (2), 449-492.

216

References

Pope, C. E., & Feyerherm, W. H. (1990). Minority status and juvenile justice processing: An assessment of the research literature. Criminal Justice Abstracts, 22, 327-336, 527-542. Poulos, T. M., & Orchowsky, S. (1994). Serious juvenile offenders: Predicting the probability of transfer to criminal court. Crime & Delinquency, 40 (1), 3-17. Reddington, F. P., & Sapp, A. D. (1997). Juveniles in adult prisons: Problems and prospects. Journal of Crime and Justice, 20 (2), 139152. Regnery, A. S. (1985). Getting away with murder: Why the juvenile system needs an overhaul. Policy Review, 34, 65-68. Regnery, A. S. (1986). A federal perspective on juvenile justice reform. Crime & Delinquency, 32 (1), 39-51. Rendleman, D. R. (1979). Parens patriae: From chancery to the juvenile court. In F. L. Faust & P. J. Brantingham (Eds.), Juvenile justice philosophy (pp. 58-96). St. Paul, MN: West Publishing. Richards, P., & Tittle, C. R. (1981). Gender and perceived chances of arrest. Social Forces, 59, 1182-1199. Richards, P., & Tittle, C. R. (1982). Socioeconomic status and perceptions of personal arrest probabilities. Criminology, 20, 329346. Rothman, D. J. (1980). Conscience and convenience: The asylum and its alternatives in progressive America. Boston: Little, Brown. Royscher, M., & Edelman, P. (1981). Treating juveniles as adults in New York: What does it mean and how is it working? In J. C. Hall, D. M. Hamparian, J. M. Pettibone, & J. L. White (Eds.), Major issues in juvenile justice information and training: Readings in public policy (pp. 265-293). Columbus, OH: Academy for Contemporary Problems. Rudman, C., Hartstone, E., Fagan, J., & Moore, M. (1986). Violent youth in adult court: Process and punishment. Crime & Delinquency, 32 (1), 75-96. Sagatun, I., McCollum, L. L., & Edwards, L. (1985). The effect of transfers from juvenile to criminal court: A loglinear analysis. Journal of Crime and Justice, 8, 65-92. Saltzman, L. E., Paternoster, R., Waldo, G. P., & Chiricos, T. G. (1982). Deterrent and experiential effects: The problem of causal order in perceptual deterrence research. Journal of Research in Crime and Delinquency, 19, 172-189.

References

217

Sampson, R. J. (1986). Effects of socioeconomic context on official reaction to juvenile delinquency. American Sociological Review, 51, 876-885. Sanborn, J. B. (1994). Certification to criminal court: The important policy questions of how, when, and why. Crime & Delinquency, 40 (2), 262-281. Schall v. Martin, 4. U. S. 2. (1984). Schmidt, P., & Witte, A. D. (1988). Predicting recidivism using survival models. New York: Springer-Verlag. Schmidt, P., & Witte, A. D. (1989). Predicting criminal recidivism using “split population” survival models. Journal of Econometrics, 40, 141-159. Schneider, A. L., & Ervin, L. (1990). Specific deterrence, rational choice, and decision heuristics: Applications in juvenile justice. Social Science Quarterly, 71 (3), 585-601. Schwartz, I. M. (1989). (In)justice for juveniles. Lexington, MA: Lexington Books. Schwartz, I. M., Guo, S., & Kerbs, J. J. (1993). The impact of d emo gr ap hic var iab les o n p ub lic o p inio n r egar d ing juvenile justice: Implications for public policy. Crime & Delinquency, 39 (1), 5-28. Sechrest, L., White, S. O., & Brown, E. D. (1979). The rehabilitation of criminal offenders: Problems and prospects. Washington, DC: National Academy of Sciences. Shannon, L. W. (1980). Assessing the relationship of adult criminal careers to juvenile careers. In C. Abt (Ed.), Problems in American social policy (pp. 232-246). Cambridge, MA: Abt Books. Sherman, L. W. (1992). Policing domestic violence: Experiments and dilemmas. New York: Free Press. Sherman, L. W. (1993). Defiance, deterrence, and irrelevance: A theory of the criminal sanction. Journal of Research in Crime and Delinquency, 30 (4), 445-473. Sherman, L. W., & Smith, D. A. (1992). Crime, punishment, and stake in conformity: Legal and informal control of domestic violence. American Sociological Review, 57, 680-690. Sickmund, M., Snyder, H. N., & Poe-Yamagata, E. (1997). Juvenile offenders and victims: 1997 update on violence. (Report No. NCJ165703). Washington, DC: Office of Juvenile Justice and Delinquency Prevention.

218

References

Singer, S. I. (1993). The automatic waiver of juveniles and substantive justice. Crime & Delinquency, 39 (2), 253-261. Singer, S. I. (1996). Recriminalizing delinquency: Violent juvenile crime and juvenile justice reform. New York: Cambridge University Press. Singer, S. I., & McDowall, D. (1988). Criminalizing delinquency: The deterrent effects of the New York juvenile offender law. Law and Society Review, 22 (3), 521-535. Smith, D. A. (1986). The neighborhood context of police behavior. In A. Reiss, & M. Tonry (Eds.), Crime and justice: An annual review of research (Vol. 8, pp. 313-341). Chicago: University of Chicago Press. Smith, D. A., & Brame, R. (1994). On the initiation and continuation of delinquency. Criminology, 32 (4), 607-629. Smith, D. A., & Gartin, P. R. (1989). Specifying specific deterrence: The influence of arrest on future criminal activity. American Sociological Review, 54, 94-105. Smith, D. A., & Paternoster, R. (1990). Formal processing and future delinquency: Deviance amplification as a selection artifact. Law and Society Review, 24 (5), 1109-1131. Smith, D. A., & Polsenberg, C. (1992). Specifying the relationship between arrestee drug test results and recidivism. Journal of Criminal Law and Criminology, 83, 364-377. Smith, D. A., Wish, E. D., & Jarjoura, G. R. (1988). Drug use and pretrial misconduct in New York City. Journal of Quantitative Criminology, 5, 101-126. Smith, L. G., & Akers, R. L. (1993). A comparison of recidivism of Florida’s community control and prison: A five-year survival analysis. Journal of Research in Crime and Delinquency, 30 (3), 267-292. Snyder, H. N. (1997). Juvenile arrests 1996. (Report No. NCJ167578). Washington, DC: Office of Juvenile Justice and Delinquency Prevention. Snyder, H. N., & Sickmund, M. (1995). Juvenile offenders and victims: A national report. (Report No. NCJ-153569). Washington, DC: Office of Juvenile Justice and Delinquency Prevention. Stafford, M. C., & Warr, M. (1993). A reconceptualization of general and specific deterrence. Journal of Research in Crime and Delinquency, 30 (2), 123-135.

References

219

Tacq, J. (1997). Multivariate analysis techniques in social science research. London: Sage Publications, Ltd. Tannenbaum, F. (1938). Crime and the community. New York: Columbia University Press. Thomas, C. W., & Bilchik, S. (1985). Prosecuting juveniles in criminal courts: A legal and empirical analysis. The Journal of Criminal Law and Criminology, 76 (2), 439-479. Thomas, C. W., & Bishop, D. (1984). The effect of formal and informal sanctions on delinquency: A longitudinal comparison of labeling and deterrence theories. Journal of Criminal Law and Criminology, 75 (4), 1222-1245. Thornberry, T. P., & Christenson, R. L. (1984). Juvenile justice decision-making as a longitudinal process. Social Forces, 63, 164171. Thornberry, T. P., Huizinga, D., & Loeber, R. (1995). The prevention of serious delinquency and violence: Implications from the program of research on the causes and correlates of delinquency. In J. C. Howell, B. Krisberg, J. D. Hawkins, & J. J. Wilson (Eds.), A sourcebook: Serious, violent, and chronic juvenile offenders (pp. 213-237). Thousand Oaks, CA: Sage Publications, Inc. Tittle, C. R. (1969). Crime rates and legal sanctions. Social Problems, 16, 408-423. Tittle, C. R. (1975). Deterrents or labeling. Social Problems, 22, 332345. Tittle, C. R., & Rowe, A. R. (1974). Certainty of arrest and crime rates: A further test of the deterrence hypothesis. Social Forces, 52, 455462. Tonry, M. (1996). Sentencing matters. New York: Oxford University Press. Torbet, P., Gable, R., Hurst, H. I., Montgomery, I., Szymanski, L., & Thomas, D. (1996). State responses to serious and violent juvenile crime. (Report No. NCJ-161565). Washington, DC: Office of Juvenile Justice and Delinquency Prevention. Tremblay, R. E., & Craig, W. M. (1995). Developmental crime prevention. In M. Tonry & D. P. Farrington (Eds.), Crime and justice: A review of research (Vol. 19, pp. 151-236). Chicago: The University of Chicago Press.

220

References

Tremblay, R. E., Vitaro, F., Bertrand, L., LeBlanc, M., Beauchesne, H., Boileau, H., & David, H. (1992). Parent and child training to prevent early onset of delinquency: The Montreal longitudinalexperimental study. In J. McCord & R. E. Tremblay (Eds.), Preventing antisocial behavior: Interventions from birth through adolescence (pp. 117-138). New York: Guilford. Tyler, T. R. (1990). Why people obey the law. New Haven: Yale University Press. Visher, C. A., Lattimore, P. K., & Linster, R. L. (1991). Predicting the recidivism of serious youthful offenders using survival models. Criminology, 29 (3), 329-362. Waldo, G. P., & Chiricos, T. (1972). Perceived penal sanction and selfreported criminality: A neglected approach to deterrence research. Social Problems, 19, 522-540. Ward, D. A., & Tittle, C. (1993). Deterrence or labeling: The effects of informal sanctions. Deviant Behavior, 14, 43-64. Wellford, C. F., & Triplett, R. (1993). The future of labeling theory: Foundations and promises. In F. Adler & W. Laufer (Eds.), New directions in criminological theory (Vol. 4, pp. 1-22). New Brunswick, NJ: Transaction Publishers. Wilson, J. Q. (1983). Thinking about crime (Rev. ed.). New York: Vintage Books. Wilson, J. Q. (1995). Crime and public policy. In J. Q. Wilson & J. Petersilia (Eds.), Crime (pp. 489-507). San Francisco: Institute for Contemporary Studies. Winner, L., Lanza-Kaduce, L., Bishop, D. M., & Frazier, C. E. (1997). The transfer of juveniles to criminal court: Reexamining recidivism over the long term. Crime & Delinquency, 43 (4), 548-563. Wolfgang, M. (1982). Abolish the juvenile court system. California Lawyer, 2, 12-13. Wolfgang, M., Figlio, R., & Sellin, T. (1972). Delinquency in a birth cohort. Chicago: University of Chicago Press. Wolfgang, M., Thornberry, T., & Figlio, R. (1987). From boy to man: From delinquency to crime. Chicago: University of Chicago Press. Wooldredge, J. D. (1988). Differentiating the effects of juvenile court sentences on eliminating recidivism. Journal of Research in Crime and Delinquency, 25 (3), 264-300.

References

221

Wordes, M., & Bynum, T. S. (1995). Policing juveniles: Is there bias against youths of color? In K. K. Leonard, C. E. Pope, & W. H. Feyerherm (Eds.), Minorities in juvenile justice (pp. 47-65). Thousand Oaks, CA: Sage Publications, Inc. Yoshikawa, H. (1994). Prevention as cumulative protection: Effects of early family support and education on chronic delinquency and its risks. Psychological Bulletin, 115, 1-27. Yoshikawa, H. (1995). Long-term effects of early childhood programs on social outcomes and delinquency. The Future of Children, 5, 51-75. Yu, J. (1994). Punishment celerity and severity: Testing a specific deterrence model on drunk driving recidivism. Journal of Criminal Justice, 22 (4), 355-366. Zimring, F. E. (1981). Notes toward a jurisprudence of waiver. In J. C. Hall, D. M. Hamparian, J. M. Pettibone, & J. L. White (Eds.), Major issues in juvenile justice information and training: Readings in public policy (pp. 193-205). Columbus, OH: Academy for Contemporary Problems. Zimring, F. E. (1991). The treatment of hard cases in American juvenile justice: In defense of discretionary waiver. Notre Dame Journal of Law, Ethics & Public Policy, 5, 267-280. Zimring, F. E., & Hawkins, G. J. (1973). Deterrence: The legal threat in crime control. Chicago: The University of Chicago Press.

This page intentionally left blank

Index

68, 70, 74, 77, 80-81, 88, 100-102, 185-186 Biskup, M. D., 1 Blomquist, M. E., 4, 12-13, 15, 17-18, 21-22, 24 Blumstein, A., 1, 19-20, 36, 88, 94-95, 192, 195-196 Bortner, M. A., 5-7, 9, 23, 25, 58, 62, 79, 184 Braithwaite, J., 43, 54, 187-188 Brame, R., 44, 54, 107 Brantingham, P. J., 16 Breed v. Jones, 17, 22, 29 Brewer, D. D., 197-198 Bridges, G. S., 40 Brown, W. K, 17, 51-52 Brutalization effect, 186 Bry, B. H, 198-199 Bursik, R. J., 37, 39 Butts, J. A., 7, 66, 67, 80 Bynum, T. S., 88 Caeti, T. J., 7 Capital punishment, 4, 12, 41 Case processing time, 7, 47, 67, 68, 69, 73, 80, 85, 96, 98-99, 103, 109, 115, 116, 117, 118, 120, 133, 135, 136, 138, 139, 140, 142, 143, 146, 148, 154, 159,

Akers, R. L., 104 Aldrich, J. H, 103 Allison, P. D., 104, 106, 107, 154 Alpert, G., 39 Andenaes, J., 36, 41 Andrews, D. A, 197 Apospori, E, 39, 40 Arneklev, B. J., 37 Austin, J. F., 2, 4, 11, 12, 13, 21, 29, 51, 65, 191, 196, Azuma, H., 5 Baby boomers, 20 Bachman, R., 39, 54, 106 Baker, H. H, 13 Barnes, C. W., 4, 5, 6, 7, 26, 62, 63, 80, 185 Barton, W. H., 49, 50, 52, 81, 147 Bass, D., 198 Baumer, E., 43, 104 Beccaria, C., 34-36 Becker, H. S., 41, 187, 189 Benedict, W.R., 104 Bernard, T., 12-17, 93 Bilchik. S., 4-7, 9, 12, 14, 18, 20, 24-25, 28, 58, 62, 65 Bishop, D. M., 3-10, 24-25, 33, 40-42, 58, 62, 64-65, 223

224 161, 162, 163, 164, 165, 178, 184 Catalano, R. F., 197-198 Center for Juvenile Justice Training and Research, xi, 10, 93 Chamberlin, R., 198 Champion, D. J., 2, 5-7, 9, 21, 23, 25, 62, 184 Chiricos, T. G., 37-38, 40 Christenson, R. L., 94 Clarke, E. E., 5-6, 9, 39, 58, 62, 84 Clogg, C. C, 107 Coates, R. B, 51 Cohen, A., 36, 40, 94 Colonial period, 12 Confidentiality, 21 Constitution, 7, 17, 66 Conviction, 6, 9, 58-69, 72-73, 78-79, 86, 96-97, 103-105, 117-118, 123-127, 146148, 157-160, 178, 185, 194 Cook, P. J., 1, 14, 18-19, 20 Cook, T. D., 89 Cornish, D. B, 39 Cox, D. R., ix, 47, 52, 81, 104107, 147, 154-156 Creswell, J. W., 83 Crime control conservatives, 28 Cullen, F. T., 187 Currie, E., 8 Custody gap, 184 Defiance theory, 53-54 DeFrances, C., 2, 3, 5, 21, 22, 23, 25, 26, 27, 65, 191, 192

Index DeFrances, C. J., 2-5, 21-27, 65, 84, 191-192 Dei, K. A., 74 DeJong, C., 37, 41, 43, 52, 104 Deschenes, E. P., 5, 22-25, 60, 89, 92, 191 Deterrence, 2-3, 8, 18, 28, 3449, 53-62, 68-74, 77, 81, 176-178 Deviance amplification, 42-43, 81 DiIulio, J. J, 20, 33 Dohrenwend, B. P., 187 Double jeopardy, 17 Due process rights, 11-12, 1417, 67, 78 Due process liberals, 28 Edelman, P., 7, 62, 79, 184 Edwards, L., 5 Eigen, J. P., 5-7, 9, 22-23, 5859, 62-65, 79-80, 89 Ekland-Olson, S., 43 Emerson, R., 7, 62, 79, 184 Erickson, M. L., 15, 18, 20, 46, 54, 93 Ervin, L., 40, 49-53 Experiential effect, 38-40 Fagan, J., 1, 4-11, 19, 22-25, 29, 34, 58, 60-68, 70-75, 77-81, 89, 92, 100, 104, 185-186, 191 Farrington, D. P., 3, 20, 42-43, 53, 95, 192, 195-196 Faust, F. L., 16 Federal Bureau of Investigation, 27 Feld, B. C., 3-8, 10-14, 21-26, 28, 30, 32, 58, 61-65, 71,

Index 79-81, 87, 89, 92-94, 100, 103, 185-186, 190-192, 195 Feyerherm, W. H., 6, 198 Fichter, M., 5 Fields, G., 1 Figlio, R., 53 Forst, M., 4, 5, 12, 15, 17, 18, 22, 72, 187 Forst, M. L., 4-5, 12-13, 15, 17-18, 21-24, 72, 187 Franz, R. S., 4-7, 26, 62-63, 80, 185 Frazier, C. E., 3, 5-9, 24-25, 58, 62, 65, 88 Freeman, R. B, 188 Fritsch, E. J., 3, 7, 9, 21, 25, 65-66, 80, 98, 185 Future research, 54, 199 Gartin, P. R., 41, 44-45, 52-53, 100 Gartner, R., 38 Geerken, M. R., 38-39, 99 Gendreau, P., 197 Get tough, 2, 11, 18-19, 33, 53, 55, 101, 183, 185 Gibbons, T. J., 1 Gibbs, J. P., 36-37, 41, 43 Gillespie, L., 5-6, 9, 58, 62, 184 Goggin, C., 197 Gold, M., 43-45 Gottfredson, D. C., 49-50, 52, 81, 147, 198 Gove, W. R., 38-39, 187 Grasmick, H., 37, 39 Greenbaum, S., 1, 19 Greenwood, P. W., 2, 11, 78 Guo, S., 41

225 Hagan, J., 62 Halemba, G. J., 7, 66, 80, 204 Hamparian, D., 5-7, 9, 22-25, 58, 62, 79, 93, 184 Haritou, A., 107 Hartmann, D. J., 51 Hartstone, E., 6 Hawkins, J. D., 1, 36-37, 43, 197, 198 Hemmens, C., 3, 7, 9, 21, 25 Henderson, C. R., 198 Henretta, J. C., 5 Hirschi, T., 145, 188 Holden, R. T., 188 Horney, J., 39, 40 Houghtalin, M., 5-6, 58, 62-63 Houses of refuge, 12, 14 Howell, J. C., 1, 3, 7, 9-10, 18, 20, 24-27, 67, 72-73, 187, 196-198 Huff-Corzine, L., 104 Huizinga, D., 95 In re Gault, 16 In re Winship, 17 Incapacitation, 2-3, 8, 15, 18, 28, 33, 55, 62, 68, 73, 173, 177, 180, 185, 194195

Incarceration, 6-9, 37, 48-49, 52-55, 62-66, 73, 78-80, 87, 96-105, 116-120, 128133, 146-149, 152-153, 157-160, 163-165, 168, 180, 185, 187, 194-196 Interaction effects, 107, 125, 127, 136, 139, 142, 147, 152, 158, 160, 163-164

226 Iovanni, L., 37-39, 41, 43, 92, 187-189 Jacobs, M. D., 51-52 Jarjoura, G. R., 101 Jenkins, R. L., 51 Jensen, E. J., 8, 70 Johnson, B. D., 74 Jones, M., 17, 20, 22, 29 Joo, H. J., 43, 210 Judicial waiver, 9, 21-22, 2532, 72, 85, 171-172, 180 Just deserts, 18 Justice by geography, 22 Kazdin, A. E, 198 Keiter, R. B., 5, 9, 23 Kelly, W. R., 43 Kent v. United States, 16, 22, 29 Kerbs, J. J., 41 Kinder, K., 5-7, 9, 26, 58, 60, 68, 79, 80, 84 Kitzman, H., 198 Klein, M. W., 47-49, 52 Klemke, L. W., 43-45 Klepper, S., 39

Krisberg, B., 1-2, 4, 8, 11-13, 20-21, 29, 51, 65, 67, 191, 196-197 Labeling, 41-44, 81, 92, 187189, 193 Lanza-Kaduce, L., 3, 8, 40, 54, 188 Lattimore, P. K., 104 Laub, J. H., 1, 18-20 LeBlanc, M., 95, 196 Lee, L., 5, 9 Legislative waiver, 9, 21, 24-

Index 29, 57, 69, 72, 82-88, 95, 167-173, 180, 183-186, 190, 193-194, 198 Lemert, E. M., 41, 189, 211 Lemmon, J. H., xi, 5-7, 25, 30, 58, 62, 67 Leniency gap, 6, 62-63, 184 Leonard, K. K., 6, 198 Lewis-Beck, M. S., 103, 106, 109 Link, B., 187, 188 Linster, P. K., 104 Lipsey, M., 197 Lipton, D., 17, 46 Little, T., 186, 197 Loeber, R., 95, 196 Logan, C. H., 37 Logistic regression, 103-106, 119-120, 123, 125, 128, 136, 140, 143, 149 MacDonald, J. M., 104 Mahoney, A. R., 43 Makkai, T., 43, 54, 188 Maltz, M. D., 47, 100 Marshall, I. H., 39-40 Martin, S.E., 78 Martinson, R., 17, 46 Matsueda, R. L., 38, 44 Mays, G. L., 2, 5-6, 21, 23, 26, 58, 62-63 Mazerolle, P., 107 McCollum, L. L., 5 McCoy, C. R., 1 McDowall, D., 8, 69 McKeiver v. Pennsylvania, 17 Meddis, S. V., 3, 57 Menard, S., 103, 106, 109 Metsger, L. K., 8, 70 Miller, S., 37, 39, 51, 62

Index Minor, K. I., 38, 40, 51-52 Minor, W. W., 52 Montrose Training School, 50 Moore, M. H., 1, 6, 11, 190 Multicollinearity, 108-109 Murray, C. A., 47, 52, 81, 147 Myers, D. L., 30 Nagin, D., 36-39 Neckerman, H. J., 197 Nelson, F. D., 103 Neter, J., 103, 109 Norman, M. D., 5-9, 58, 62, 184 Nothing works, 17, 46, 197 Ohlin, L., 3, 51 Olds, D., 198 Orchowsky, S., 5, 22-23, 89, 92-93, 191 Ordinary least squares, 103104, 119 Osborn, S. G., 53 Osbun, L. A.,, 5, 9, 24 Overberg, P., 1 Parens patriae, 12, 15 Paternoster, R., xi, 36-45, 54, 81, 85, 92-95, 100-101, 106-107, 138, 164, 187189 Pauperism, 12 Pennsylvania, 10-11, 29-32, 51, 67, 75, 84-88, 93-100, 167-168, 171, 174, 176, 180, 184, 186, 189-190, 193-195 Pennsylvania Juvenile Court Judges’ Commission, 2932 Petkova, E., 107 Piliavin, I., 38-40

227 Pindyck, R. S., 103 Piquero, A., 37, 39-41, 45, 54, 107, 164 Platt, A., 12

Podkopacz, M. R., 5-10, 2223, 26, 58, 61-65, 71, 7981, 87, 89, 92, 94, 100, 185-186, 191 Polsenberg, C., 101 Pope, C. E., 6, 198 Poulos, T. M., 5, 22, 23, 89, 92-93, 191 Prior record, 5, 9, 22-23, 64, 72, 86, 95, 109, 117, 122, 135, 143, 146, 148, 152153, 159, 162, 165, 195 Progressive movement, 4, 12, 14 Prosecutorial waiver, 9, 21, 2427, 183 Provo experiment, 46-47 Punishment certainty, 9-10, 33-41, 49, 55-61, 66, 69, 74-79, 96 severity, 9-10, 33-38, 41, 49, 55-58, 62, 65-66, 69, 73-80, 97 swiftness, 9-10, 33-36, 41, 55, 58, 66, 69, 74-78, 98 Racial bias, 5, 23, 92, 198 Radosevich, M. J., 54, 188 Recidivism, 8-10, 46-53, 58, 64, 70-77, 81-87, 95-105, 109, 116-119, 142-147, 154, 161-165, 178, 183, 186-189, 194-199 Reddington, F. P., 72, 187

228 Regnery, A. S., 18 Rehabilitation, 4, 11-13, 17-18, 28-31, 169, 177 Rendleman, D. R., 13, 216 Retribution, 2-3, 8, 28, 33, 55, 62, 68, 73 Reverse waiver, 23, 32, 175, 193 Rhodes, W. A., 51 Richards, P., 40 Risk factors, 197-198 Rode, P. A., 5, 9, 24 Ross, R. R., 197 Roth, J., 94 Rothman, D. J., 4, 15 Royscher, M., 7, 62, 79, 184 Rudman, C., 6-7, 58-59, 63, 65, 68, 79-80, 185 Sagatun, I., 5, 6, 7, 58, 62, 184 Saiger, A. J., 33 Saltzman, L. E., 38, 40 Sampson, R. J., 88 Sanabria, H., 74 Sanborn, J. B., 22-25 Sapp, A. D., 72, 187 Saylor, K., 5 Schall v. Martin, 17, 78 Schmidt, P., 104 Schneider, A., 40, 49-53 Schwartz, I. M, 11, 41, 57 Sechrest, L., 17 Selection bias, 42, 46-49, 5965, 71-74, 81-86, 190, 199 Sellin, T., 53 Sentencing, 2, 7, 13, 18, 21, 33, 49, 62, 67-68, 73, 78, 96-100, 170, 172, 179, 192, 195 Shannon, L. W., 53

Index Sherman, L. W., xi, 41, 43, 53, 54, 164, 187-188 Shrout, P., 187 Sickmund, M., 1-5, 9-10, 1827, 191-192 Siegel, T. C., 198 Singer, S. I., 5-8, 11, 24-25, 29, 58, 60, 65, 68-69, 94, 190 Smith, D., 5, 23, 33, 41-45, 5254, 81, 85, 93, 95, 100101, 138 Smith, L., 104 Snyder, H., 1, 5, 10, 18-20, 27 Sontheimer, H., 5 Speedy trial, 7, 67 Stafford, M. C., 15, 18, 20, 37, 40, 45-46, 54, 93 Steele, P. A., 51 Stone, J. A., 40 Strom, K. J., 2-3, 5, 21-27, 65, 84, 191-192 Struening, E., 187 Suppression effect, 47 Supreme Court, 7, 15-16, 22, 67, 78 Survival analysis, 104, 106, 119, 155 Tacq, J., 103, 109 Tannenbaum, F., 41 Tatelbaum, R., 198 Terry, S., 51 Thomas, C. W., 4-9, 12, 14, 18, 20, 24-25, 28, 40-41, 58, 62, 65 Thornberry, T. P., 94, 95, 196 Thorton, C., 38 Tibbetts, S., 39 Time to failure, 70, 102, 104,

Index 116 Tittle, C., 37, 40, 44, 187 Tonry, M., 1, 18 Torbet, P., 2-3, 9, 21-26 Tremblay, R. E., 198 Triplett, R., 44 Tyler, T. R., 54, 164, 187, 188 Veneziano, C., 5 Visher, C. A., 94, 104 Vivona, T. S., 5, 72 Wakeling, S., 11, 190, 213 Waldo, G. P., 37-40 Ward, D., 37, 44, 39 Warr, M., 37, 40, 45 Wellford, C. F., 44 West, D. J., 53 White, S. O., 17, 23 Wiebush, R. G., 8 Wilkinson, D. L., 1, 19 Wilks, J., 17 Williams, J. R., 43-45, 74 Wilson, J. Q., 1, 3, 18, 20, 33, 36, 43 Winner, L., 3, 8, 24, 70, 81, 186 Wish, E. D., 101 Witte, A. D., 104 Wolfgang, M., 18, 53, 89, 99 Wooldredge, J. D., 47-49, 52 Wordes, M., 88 Yoshikawa, H., 198 Youth violence, 1, 5, 8, 11, 19, 20, 26, 32, 56-57, 64, 7374, 191, 196-197 Yu, J., 37, 43 Zimring, F. E., 4, 22, 25, 3637, 43, 191

229

This page intentionally left blank