No title

Archives of Sexual Behavior, Vol. 26, No. 1, 1997

Are Transvestites Necessarily Heterosexual? Bonnie Bullough, R.N., Ph.D.1 and Vern Bullough, R.N., Ph.D.2,3

A survey of 372 male cross-dressers gathered data about present and childhood experiences and attitudes in light of the growing knowledge about transvestism. This article focuses on data related to sexual orientation, particularly in relationship to the definition of transvestism in the Diagnostic and Statistical Manual of Mental Disorders of the American Psychiatric Association. It is argued that transvestism is not necessarily a heterosexual phenomenon. KEY WORDS: transvestism; sexual orientation; heterosexual; gender identity; transgenderist.

INTRODUCTION Most documents written before the middle of the 20th century made no clear distinction between sexual orientation and gender; a man described as effeminate was often assumed to be homosexual (V. Bullough, 1976). Similarly a masculine woman was often thought of as lesbian. An example of this was the prototype lesbian woman pictured by Hall (1929) in Well of Loneliness (Devor, in press). Thus the historical record of crossdressing sometimes included an assumption of homosexuality but other times it did not. This was particularly true of women, the most frequent cross-dressers before the 20th century whose motivation was often described as economic or a desire for freedom rather than a sexual urge (V. Bullough and Bullough, 1993). Hirschfeld, the first serious student of cross-dressing, coined the term "transvestitism" and indicated that the group was primarily heterosexual 1University of Southern California, 2State University of New York.

1540 Alcazar Street, Los Angeles, California 90033. Visiting Professor of Nursing, University of Southern

California, Los Angeles. whom correspondence should be sent at 17434 Mayall Street, Northridge, California 91325.

3To

1 0004-0002/97/0200-0001$12.50/0 C 1997 Plenum Publishing Corporation

2

Bullough and Bullough

and male. A German physician, he was himself an avowed homosexual, a reformer, and a specialist in the study of sexuality. In 1910 he published what may be the key work on cross-dressing, The Transvestites: An Investigation of the Erotic Drive to Cross Dress (1991). He reviewed the cases of 16 men and 1 woman, most of whom were patients. The others were located through newspaper stories, correspondence with people he heard about, and referrals from colleagues. He described cross-dressing as starting in early childhood, increasing during puberty, and remaining almost unchanged after that. Although most of the persons he studied were heterosexual, Hirschfeld noted there were a few homosexuals but argued that the dominant sexual urge among transvestites was focused on themselves dressed in women's finery rather than other persons of either sex. The one woman in the group indicated some attraction to females, but she eventually married and had children. In her history she focused on the freedom and life-style which a man's identity provided her more than she did on erotic pleasure (Hirschfeld, 1991). Ellis, another early 20th century sex researcher essentially agreed with Hirschfeld's findings, although not his terminology since he felt the phenomenon went beyond simple cross-dressing (Ellis, 1913, 1936). Then for the next SO years the research on transvestism was done by treatment-oriented psychiatrists who characterized transvestism as an illness and sought to treat it with psychotherapy. They searched for key elements in the childhood history of their clients that would help them understand and treat them. Castration anxiety and homosexual panic emerged as the major explanations (Gutheil, 1930; Fenichel, 1930; Hora, 1953; Peabody et al., 1953; Lukianowitz, 1959). More recently psychiatrists and psychologists have sought understanding using a broader range of variables but have continued to use an illness model to conceptualize all types of cross dressing (Person and Ovesey, 1978; Brierley, 1979, Beatrice, 1985; Fagan et al., 1988; Docter, 1988, 1993). Stoller (1971) who focused on the family constellation as a causal factor in both transsexualism and transvestism, argued that a strong mother figure was the major causal factor. His position has been supported in a recent study by Schott (1995). Codification of the thinking of the psychiatric and psychological community on transvestism was reflected in the definition adopted in DSM-III-R (American Psychiatric Association [APA], 1987). Though homosexuality was omitted from the DSM list of mental illnesses in 1974 (Bayer, 1981), the continued listing of transvestism can be partly explained by the fact that transvestism had neither been studied as much as homosexuality nor were the transvestites as politically astute as the gay-power movement in demanding that their diagnosis be removed from

Are Transvestites Heterosexual?

3

the DSM. Only recently has such an effort begun to be mounted in the cross-dressing community. The new DSM-IV (APA, 1994) definition of transvestism is similar to the earlier definition. Although the discussion section indicates that heterosexual men may occasionally have homosexual encounters, the definition is specifically focused on heterosexual men. The definition excludes transsexuals, people who cross-dress for relief of tension or gender discomfort, but who have sexual excitement, and men whose sexual orientation is homosexual. The diagnostic criteria for Transvestic Fetishism (disorder #302.30) reads as follows: A. Over a period of at least 6 months, in a heterosexual male, recurrent intense sexually arousing fantasies, sexual urges, or behaviors involving cross-dressing. B. The fantasies, sexual urges, or behaviors cause clinically significant distress or impairment in social, occupational, or other important areas of functioning. Specify if With Gender Dysphoria if the person has persistent discomfort with gender role or identity. (APA, 1994)

At the same time that the psychiatric or illness approach to transvestism was in sway, another influence on the perception emerged—namely, the transvestite club movement. Virginia Prince, the founder of the movement was well educated, holding a doctorate in science, and she could and did read the literature in the field. Although she knew Stoller and talked to him at length she did not accept the illness model of transvestism but agreed with him that it was heterosexual. She found the work of Money more persuasive, partly because he was a dedicated writer of definitions, which was also one of Prince's great interests. She defined transvestism as strictly a heterosexual activity, and argued that it was therefore more respectable than homosexuality. She did not deny that there were homosexual cross-dressers but said they were drag queens who focused on seduction rather than the joy of dressing (V. Bullough and Bullough, 1993, p. 324). She also supported the definitions of Money and Ehrhardt (1972) who pointed out that sexual orientation and gender identity were totally different concepts. Prince started publication of the magazine Transvestia in 1960 and it was for almost a decade the only periodical in the field. She was also the founder of the many cross-dressing clubs, starting in the 1960s in the United States and extending throughout Europe, Canada, and Australia by the 1970s. The club movement provided a social support system and a sexual script for cross-dressing that made cross-dressing at home or at club meetings a pleasant and respectable activity, enabling the members to temporarily forget the hostility of society and family (V. Bullough and Bullough, 1993). The historical record is not clear on whether Prince influenced the

4

Bullough and Bullough

DSM definition on the issue of heterosexuality or the DSM definition influenced Prince. The behavioral science approach to studying and classifying crossdressers developed after the emergence of the transvestite clubs in the 1960s since the club movement provided for the first time a ready population for the study of cross-dressing people who were not necessarily patients. For a time Prince refused to make her lists available unless she was involved in the study. She and Bender did a survey of 502 readers of Transvestia, the club publication which circulated to cross-dressers throughout the country, and also had a small international following. Although data were gathered from 1964-1966 the findings of this survey were not published until 1972 (Prince and Bender, 1972). Other studies followed with the Australian team, Buhrich and McConaghy (1977a, 1977b, 1978,1979) using 83 club members as their sample. Buhrich and Beaumont (1981) did a comparison of 86 Australians and 126 Americans. V. Bullough et al., (1983a, 1983b; B. Bullough et al., 1985) comparing the life histories of 65 transvestites with transsexuals and gay men. Docter (1988) studied 110 cross-dressers who were located through clubs. There were still other studies (Randall, 1959; Buhrich and McConaghy, 1985; Stoller, 1985). In spite of the fact that this first generation of survey researchers used nonpatients as their study samples, they did not move very far away from either the psychiatric model or the ideology of the club members. The Australians, Buhrich, McConaghy, and Beaumont followed the DSM definition on the issue of fetishism and excluded people who had no sexual arousal related to cross-dressing. Similarly Docter excluded people who did not fit the definition in two areas; if they had never experienced excitement in association with cross-dressing or if they were homosexual (Docter, 1988 p. 135). Although V. Bullough et al., avoided the flaws of exclusion of homosexual men and people who did not report excitement with cross-dressing, their sample was undoubtedly biased by the fact that it used Los Angeles clubs as a sample source. Los Angeles was Prince's home ground and those clubs, even those who had broken with her, still felt her powerful influence. The Prince and Bentler (1972) sample was the most representative since it covered a national readership of Transvestia and in spite of the ideology espoused by Prince they reported that 9% of their sample of 504 men indicated that they were homosexual in orientation, with 29% reporting some homosexual experience. Buhrich and McConaghy (1977) reported 3% homosexual in their sample with 17% having some homosexual experience. V. Bullough et al. (1983) found no one who was exclusively homosexual, although 18% indicated they had some homosexual experience.

Are Transvestites Heterosexual?

5

In their cross-cultural study, Buhrich and Beaumont (1981) asked subjects to indicate their sexual orientation when they were dressed as males and when they were dressed as females: 87% of an American sample said they were exclusively heterosexual when they were dressed as men but only 52% were exclusively heterosexual when cross dressed. Similarly, an Australian sample went from a 72% heterosexual orientation in men's clothing to a 56% heterosexual orientation in women's clothing. Homosexual fantasies or actual homoerotic activities are clearly a part of the cross-dressing scene. We therefore decided to survey a larger more geographically dispersed sample of cross-dressers to again look at the definition that specified that transvestites must be heterosexual, as well as to study selected variables related to the childhood and present patterns of cross-dressing with an emphasis on those variables that might help us understand the phenomenon.

METHOD Questionnaires were sent to 1200 members of Tri Ess, a national cross-dressing organization, by enclosing questionnaires with the association's magazine, Femme Mirror. In addition the American Educational Gender Information Service (AEGIS) sent out a notice of the study to its mailing list of 500 which includes persons who identify as transgendered or transsexual, members of the helping professions, and related support groups. News of the study also spread through other newsletters and friendship which prompted people to write to the researchers to secure questionnaires. These sources brought in approximately 40 individual requests for the questionnaires. An ad in The Advocate, a gay newspaper published in Los Angeles, yielded no responses. Data gathering started in the Summer of 1994, and analysis was carried out in 1995. There were 372 respondents, including men from every state in the union, 8 from Canada, and 7 from other countries. Structured questions were answered carefully and answers to open-ended questions were often answered at length. Some subjects indicated they had never told anyone all of their background before. Many people sent their names so we could contact them for additional data. A six-item homosexual attitudes scale was constructed and included in the questionnaire. It could be called a homophobia scale although it is important to remember that attitudes towards homosexuality tap more than fear (Plasek and Allard, 1984) so we call it a homosexual attitudes scale. The scale had an alpha score of .75.

6

Bullough and Bullough Table I. Current Sexual Orientation (N = 368) Orientation Heterosexual Bisexual Homosexual Sex not a part of my life now

N

%

248 39 9 72

67.4 10.6

2.4 19.6

Table II. Present Sexual Orientation and Fantasy While Cross-Dressed Fantasies

Man Woman Self Both men and women

Hetero

9.6% 52.8 27.0 10.5

Bi/homo

Asexual

30.2 23.3

29.0 32.3 24.2 11.3

9.4 37.2

RESULTS

The 372 people in this sample all started life as men and all crossdressed at some time in their lives. Their median age when they filled out the questionnaire was 48; 64% are presently married or living with a woman, 1.4% live with a man, and 35% live alone. They were an affluent group with 66% holding professional, business, or clerical jobs with an additional substantial group of retired professionals. The subjects' father's jobs had a much more normal distribution so the sample represents an upwardly mobile successful group of men. Table I shows the current self-reported sexual orientation of the sample. The homosexual group was small, only 9 persons, but the bisexual group was larger, 39 persons. Some comments suggested that bisexuality was a less stigmatized identity so even some people who seemed to be clearly oriented totally towards a same-sex partner chose the bisexual option to identity themselves. Note the finding in this table of the sizable group who report no sexual activity at present (other than cross-dressing), a finding consistent with comments made by Hirschfeld, and by Docter (1993; Docter and Fleming, 1992). Fantasy life can also be an expression of sexual orientation. Although the influence of orientation can be seen with the heterosexual group most likely to focus on a fantasy woman, and bisexual/homosexual persons focused on a fantasy man, fantasies focused on the subject are also common. Fantasy orientation seems to vary significantly.

Are Transvestites Heterosexual?

7

Table III. What Did You Fear Would Happen if You Were Caught Cross-Dressing? (N = 312) Fears Rejection Sissy label Crazy, mentally ill label Sinful label

N

%

148 78 71 15

47.4 25.0 22.8

4.8

Childhood Experiences The median age at which this group started to cross-dress was 8.5 and 32% of the sample cross-dressed before they were 6. This is slightly younger (about 1 year) than the median age reported in other studies. Most of the subjects in this research were clandestine cross-dressers as children, and 56% (by their accounts) were never caught; although 93% of the sample indicated that they were afraid of being caught. An open-ended question solicited the reasons they feared being caught. Their answers could be coded into four groups shown (Table III). The most common fear was a fear of rejection which was emphasized by 47% of respondents: 25% feared a "sissy" label; 23% were afraid of a "crazy" label, and 5% figured cross-dressing was a sin. The sissy category also included those who feared being called a "faggot" or a "queer" as well as those that used the term "sissy." From the context of the answers it seemed that sissy sometimes meant girlish and weak and sometimes it meant a homosexual orientation. This confusion is a part of the childhood culture because labels are often applied before the participants know what the words mean. Even queer can mean "strange," "girlish," or "silly" rather than denoting a sexual orientation. In addition some of the answers which were primarily focused on parental or peer rejection included a secondary fear that they would be rejected because they were sissies or crazy. Nevertheless, the fact that these labels were used in variable ways did not lessen their stigma. Fear of a label was a powerful deterrent to open expression of feminine traits including an interest in women's clothing. It may have influenced the child to stick with a clandestine activity to express his gender feelings rather than seeking out a same-sex partner which would have been less secret. Childhood and adolescent sexual experiences may also touch on the variable of sexual orientation. Ninety-seven persons or 26% of the group reported some homosexual experiences as a child or adolescent and 135 persons or 41% indicated they had some heterosexual experience as children or adolescents. For most of the respondents this was a positive expe-

8

Bullough and Bullough Table IV. Sexual Experience as a Child or Adolescent

N

%

Homosexual Yes, a positive experience Yes, ambivalent about it Yes, negative No homosexual experience

42 31 20 271

12.5

Heterosexual Yes, a positive experience Yes, ambivalent about it Yes, negative No heterosexual

104 31 20 217

Experience

8.4 5.4 73.6 28.4

8.4 5.4 59.3

rience, but a small number reported negative feelings. Table IV shows these data. Ten percent of the sample reported they were raped or sexually assaulted as children and although there is some overlap between this group and the people who reported that they had a negative sexual experience, there were also those who reported both a rape and pleasant sexual experiences. Crossing Oven The Transgenderists Twenty-five percent of the study group took hormones at some time. Most of the time hormone use was under 2 years, which is long enough to stimulate breast development but avoid some of the other side effects of estrogen. This can be thought of as a step towards a change of gender identity. There is a growing trend in both the male and female cross-dressing community to move into the opposite-sex role without benefit of surgery. Eleven percent of this study sample members are now living full time as women; 2% had sex reassignment surgery (SRS) and there were a few people who indicated that they are preoperative transsexuals. However most of the remaining 9% who are living full time as a woman are what are now called "transgenderists." Most (85%) of these people had taken or were now taking hormones. This population of transgenderists is a growing phenomenon in the cross-gender community. Although in the historical past many people crossed over without surgery, most of them were women who changed their identity to become men, with only a few men cross-dressing and changing their identity to women. (V. Bullough and Bullough, 1993). When SRS

Are Transvestites Heterosexual?

9

Table V. Sexual Orientation of Transgenderists and Persons Living as Men Orientationa

Transgenderists Living as a man aPearson:

Heterosexual

Bi-homosexual

Asexual

10 233

8 36

9 58

26.51: sig .0000.

developed after 1950 it became the rite of passage to a different sex role, and it was primarily a male to female journey. People, mostly men, believed that the surgery could give them permission to change their sexual identity by changing their sex organs (Bolin, 1988). Since SRS was not as well developed for women, many of them who wanted to change sex simply had a mastectomy and left their sex organs intact. In effect they too can be called transgenderists instead of sexually reassigned transsexuals. In this sample there were 40 persons living as women; only 11 of whom had experienced SRS. The remaining 29 were living full time as women. Seven of these indicated they planned to have surgery, but for some it seemed remote because they did not have the money or there were other significant barriers. They were analyzed with the other transgenderists even though they may not remain in this status. Some transgenderists reported other types of surgery, including facial surgery or mammoplasty, but did not plan to have genital surgery. The sexual orientation of this group of men who are living as women without SRS is variable: Table II shows the 27 transgenderists who indicated a sexual orientation, 10 called themselves heterosexual, 8 indicated they were bisexual or homosexual, and 9 asexual: one third of the transgenderists indicated that they were not much interested in sexual activities. This area of their life seemed to be centered around their gender orientation rather than their sex life. Table V summarizes these figures in comparison with remaining component of the sample who are living as men. The attitude of the transgenderists and transsexuals is less prejudiced against homosexuals than other components of the sample. The mean score of the total sample on the homosexual attitudes scale was 12.5, while the mean score of the transgenderists was 13.23. The nontransgenderists had a mean score of 12.00. These differences are statistically significant. DISCUSSION

These data clearly indicate that while a majority of transvestites are heterosexual, a significant portion are bisexual, homosexual, or not sexually

10

Bullough and Bullough

active with another person. Although the DSM-IV (APA, 1994) definition of transvestism is better than the older definition, it seems to be in error on this issue. There is support for this broader definition in the cross-cultural literature. Whitman studied cross-dressing communities in Java, Thailand, Guatemala, Peru, Brazil, and the Philippines and has found that the men who cross-dress are often homosexual. Some of them consider themselves preoperative transsexuals but few actually have sexual reassignment surgery. They have a distinctive name in each of the cultures. In Java for example, they are known as waria, and in Brazil they are known as travesties (Whitam and Mathy, 1986: Whitam, in press). Favorite occupations are hairdressing, prostitution, and entertainment including dancing, theater, and other art forms. Whatever the local culture, Whitam (in press) reports that the crossdressing men outside of United States, Canada, and Western Europe have close ties to the gay community and in many countries are primarily homosexual. The transvestite prostitutes of Costa Rica serve a clientele of heterosexual men who do not consider themselves homosexual because the client is the high-status person and the prostitute is a subordinate. This Latin American definition of the situation has more to do with power in the encounter than with sexuality (Schifter and Madrigal, in press). The transgenderists who live in the opposite sex, both men and women, are a growing phenomenon (Devor, 1989, Boswell, 1991; Bornstein, 1994, Bolin, 1994). In this sample, the transgenderists were quite variable in their sexual orientation, with orientations towards the same sex, the opposite sex, or no sex at all. The DSM model of obligatory heterosexual orientation for transvestites needs reconsideration since there is significant variation in sexual orientation among people who cross-dress. REFERENCES American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders, 4th ed., APA, Washington, DC, pp. 530-531. American Psychiatric Association. (1987). Diagnostic and Statistical Manual of Mental Disorders, 3rd ed., APA, Washington, DC. Bayer, R. (1981). Homosexuality and American Psychiatry: The Politics of Diagnosis, Basic

Books, New York. Beatrice, I. (1985). A psychological comparison of heterosexuals, transvestites, preoperative transsexuals and postoperative transsexuals. J. Nerv. Ment. Dis., 73: 358-365. Bolin, A. (1988). In Search of Eve, Bergin and Garvey, South Hadley, MA. Bolin, A. (1994). Third Sex, Third Gender, Herdt, G. (ed.). Zone Books, New York. Bornstein, K. (1994). Gender Outlaw, on Men, Women and the Best of Us, Rutledge, New York.

Are Transvestites Heterosexual?

11

Boswell, H. (1991). The transgender alternative. Chrysalis Quart., p. 1.

Brierly, H. (1979).A Handbook with Case Studies for Psychologists. Psychiatrist, and Counselors, Pergamon, Oxford. Buhrich, N., and Beaumont, T. (1981). Preadult feminine behaviors of male transvestites./4rc/i. Sex. Behav. 14: 413-419. Buhrich, N., and McConaghy, N. (1977a). The clinical syndromes of femmiphilic transvestism. Arch. Sex. Behav. 6: 397-412. Buhrich, N., and McConaghy, N. (1977b). The discrete syndromes of transvestism and transsexualism. Arch. Sex. Behav. 6: 483-495. Buhrich, N., and McConaghy, N. (1978). Parental relationships during childhood in homosexuality, transvestism and transsexualism. N.Z.J. Psychiat. 12: 103-108. Buhrich, N., and McConaghy, N. (1979). Three clinically discrete categories of fetishistic transvestism. Arch. Sex. Behav. 8: 151-157. Buhrich, N., and McConaghy, N. (1985). Preadult feminine behaviors of male transvestites. Arch. Sex. Behav. 14: 413-419. Bullough, V. L. (1976). Sexual Variance in Society and History, University of Chicago Press, Chicago, II. Bullough, B., Bullough, V., and Smith R. (1985). Masculinity and femininity in transvestite, transsexual and gay males. Western J. Nursing Res. 7: 317-332. Bullough, V. L., and Bullough, B. (1993). Cross Dressing, Sex and Gender, University of Pennsylvania Press, Philadelphia. Bullough, V., Bullough, B., and Smith R. (1983). A comparative study of male transvestites, male to female transsexuals, and male homosexuals. J.. Sex Res. 19: 238-257. Bullough, V., Bullough, B., and Smith R. (1983b). Childhood and family values of male sexual

minority groups. Health Values: Achieving High Level Wellness T. 19-26. Devor, H. (1989). Gender Blending: Confronting the Limits of Duality, Indiana University Press, Bloomington. Devor, H. (in press). More than manly woman: How female transsexuals reject lesbian identities. In Bullough, B., Bullough, V., and Elias, J. (eds.). Gender and Transgender Issues, Prometheus, Amherst, NY. Docter, R. F. (1988). Transvestites and Transsexuals, Plenum Press, New York. Docter, R. F. (1993). Dimensions of transvestism and transsexualism. J. Psychol. Hum. Sex. 5: 15-37. Docter, R. P., and Fleming, J. S. (1992). Dimensions of transvestism and transsexualism: The validation and factorial structure of the cross-gender questionnaire. Interdisciplinary Approaches in Clin. Management, pp. 15-37. Ellis, H. (1913). Sexo-aesthetic inversion, Alienist and Neurologist 34, Part 1 (May 1913) 3-14; Part 2 (August 1913) 1-31. Ellis, H. (1936). Eonism. In Studies in the Psychology of Sex, Vol. 2, Part 2, Random House, New York. pp. 1-120. (Originally published 1906). Pagan, P. J., Wise, T., Derogatis, L. J. and Schmidt, C. W. (1988). Distressed transvestites: Psychometric characteristics. J. Nerv. Ment. Dis. 176: 214-217. Fenichel, O. (1930). The psychology of transvestism. Int. J. Psychoanal. 11: 285-298. Gutheil, E. (1930). An analysis of a case of transvestism. In Stekel, W. (ed.), Sexual Aberrations; the Phenomenon of Fetishism in Relation to Sex, Liveright, New york, pp. 345-351.

Hall, R. (1929). The Well of Loneliness, Covici, Friede, New York. Hirschfeld, M. (1991). The Transvestites: An Investigation of the Erotic Drive to Cross Dress (transl. by Michael Lombardi-Nash), Prometheus, Buffalo New York. (Originally published 1910; Medical Publishing House: Max Spohr, Leipzig. Hora, T. (1953). The structural analysis of transvestism, Psychoanalyt. Rev. 40: 268-274. Krafft-Ebing, R. (1933). Psyckopathia Sexualis: With Especial Reference to the Antipathic Sexual Instinct: A Medico-Forensic Study (Trans, and adapted from 12 German edition by F. J. Rebman), Physicians and Surgeons Book, Brooklyn, NY. Lukianowicz, N. (1959). Survey of various aspects of transvestism in light of our present knowledge, J. Nerv. Mental Dis. 128: 3644.

12

Bullough and Bullough

Money, J., and Ehrhardt, A. A. (1972). Man and Woman Boy and Girl, New American Library, New York. Peabody, G. A., Rowe, A. T., and Wall, J. M. (1953). Fetishism and transvestism. J. Nerv. Ment. Dis. 119: 339-350. Person, E., and Ovesey, L. (1978). Transvestism: New perspectives. J. Am. Acad. Psychoanal. 6: 304-322. Plasek, J. W., and Allard, J. M. (1984). Misconceptions of homophobia. J. Homosex. 10 (Fall): 23-37. Prince, C. V., and Bentler, P. M. (1972). Survey of 504 cases of transvestism. Psychol. Rep. 31: 903-917. Randall, J. B. (1959). Transvestism and trans-sexualism: A study of 50 Cases. Br. Med. J. 2: 1148-1152. Schifter, J., and Madrigal, J. (in press). The transvestites lover; identity and behavior. In Bullough, B., Bullough, V., and Elias, J. (eds.), Gender and Transgender Issues, Prometheus, Amherst, NY. Schott, R. L. (1995). The childhood and family dynamics of transvestites, Arch. Sex. Behav. 24: 309-327. Stoller, R. J. (1971). The term "transvestism." Arch. Gen. Psychiat. 24: 230-237. Stotter, R. (1985). Presentations of Gender, Yale University Press, New Haven, CT. Whitam, F. L., and Mathy, R. M. (1986). Male Homosexuality in Four Societies: Brazil, Guatemala, the Philippines, and the United States, Prager, New York. Whitam, F. L. (in press). Culturally universal aspects of male homosexual transvestites and transsexuals. In Bullough, B., Bullough, V., and Elias, J. (eds.), Gender and Transgender Issues, Prometheus, Amherst, NY.

Archives of Sexual Behavior, Vol. 26, No. 1, 1997

Comorbidity of Gender Dysphoria and Other Major Psychiatric Diagnoses Collier M. Cole, Ph.D.,1.2 Michael O'Boyle, M.D., Ph.D.,1 Lee E. Emory, M.D.,1,2 and Walter J. Meyer III, M.D.1,3

Previous studies suggest that many transsexuals evidence an Axis I diagnosis according to the DSM-IV classification (e.g., psychoses, major affective disorder). The current study examined retrospectively the comorbidity between gender dysphoria and major psychopathology, evaluating the charts of 435 gender dysphoric individuals (318 male and 117 female). AH had undergone an extensive evaluation, addressing such areas as hormonal/surgical treatment, and histories of substance abuse, mental illness, genital mutilation, and suicide attempts. In addition, a subgroup of 137 individuals completed the MMPI. Findings revealed over two thirds were undergoing hormone reassignment, suggesting a commitment to the real-life cross-gender process. One quarter had had problems with substance abuse prior to entering treatment, but less than 10% evidenced problems associated with mental illness, genital mutilation, or suicide attempts. Those completing the MMPI (93 female and 44 male) demonstrated profiles that were notably free of psychopathology (e.g., Axis I or Axis II criteria). The one scale where significant differences were observed was the Mf scale, and this held true only for the male-to-female group. Psychological profiles as measured by the MMPI were more "normal" in the desired sex than the anatomic sex. Results support the view that transsexualism is usually an isolated diagnosis and not part of any general psychopathological disorder. KEY WORDS: gender dysphoria; transsexualism; MMPI; comorbidity of psychiatric disorders.

1Department

of Psychiatry and Behavioral Sciences, University of Texas Medical Branch, 301 University Boulevard, Galveston, Texas 77550-0133. 2Rosenberg Clinic, Galveston, Texas 77550. *To whom correspondence should be addressed. 13 0004-0002/97/0200-0013$12.50/0 c 1997 Plenum Publishing Corporation

14

Cole, O'Boyle, Emory, and Meyer INTRODUCTION

Gender dysphoria, or transsexualism, involves a long-standing and persistent feeling that one's sexual identity is incongruent with one's anatomic sex. The American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (DSM-IV) delineates this further by noting that such individuals often have the desire to change their bodies through hormonal therapy and surgery to fit their personal sexual identity. This phenomenon has been described clinically and researched since the middle of this century (Cauldwell, 1949; Benjamin, 1966). Before that, cases of gender incongruence can be found in the writings of early sexologists such as Krafft-Ebing (1894) and Hirschfeld (1945). Indeed, references to such behavior can be found in a variety of cultures dating back to antiquity (Green and Money, 1969). Much has been done in recent decades to clarify this disorder and develop treatment techniques to aid individuals experiencing these feelings (Pauly, 1990). An international medical society has been formed, the Harry Benjamin International Gender Dysphoria Association, and Standards of Care have evolved for professionals who work with transgendered persons (Walker et at., 1985). Articles have appeared in the literature suggesting that outcome following treatment can be positive, and that individuals can lead happy and productive lives (Lundstrom et al., 1984; Cole et al, 1994). Despite this, however, a suspicion continues to exist that such individuals harbor serious underlying psychopathology. As some researchers note, the persistent wish to mutilate one's body must be a sign of some devastating trauma or maladaptive developmental-familial process from childhood that has gone terribly awry (Stoller, 1968, 1975; Ovesey and Person, 1973; Lothstein, 1979a, 1979b, Morgan, 1978; Buhrich and McConaghy, 1978; Halle et al., 1980). Some type of psychosis is generally suspected (Lothstein, 1983). Estimates from the National Institute of Mental Health regarding the American population in general suggest that up to 25% may have identifiable psychiatric symptoms suggestive of anxiety disorders, depression, drug and alcohol abuse, and personality disorders (Robins et al., 1984; Weissman and Myers, 1978; Weissman et al., 1991). In light of this, it is of interest to examine psychiatric symptoms in a transsexual population. To what degree do individuals with gender dysphoria have significant psychiatric problems that could result in a comorbid diagnosis? How have they impacted or not impacted the sense of gender dysphoria? The present study sought to examine these questions.

Comorbidity of Psychiatric Disorders

15

METHODS

Individuals presenting with a self-diagnosis of gender dysphoria or transsexualism were examined for indicators of coexisting mental illness that could result in a comorbid Axis I or Axis II psychiatric diagnosis. The criteria employed to determine whether other psychiatric problems were present are described. The overriding philosophical stance used was "functional" in nature. That is, to be considered as a serious psychopathological problem, the individual would need to experience the disruption in mood or personality to such an extent that it had impacted one's life, work, and relationship in identifiable ways. Many individuals experience transient bouts of anxiety, depression, or other states that may temporarily affect them but do not have long-standing notable consequences. Of interest here were the more serious, long-standing conditions that did have a serious impact on mental health.

Procedure

The data came from a retrospective analysis of 435 individuals who presented to a gender clinic with a self-diagnosis of transsexualism. Every subject completed a minimum 1- to 2-hour clinical interview and an extensive biographical, medical-psychosocial questionnaire upon first contact with the clinic. The factors studied included the following: (i) hormone treatment (Have you undertaken any prescribed hormone therapy for your condition?); (ii) surgical treatment (Have you undergone any surgical procedures related to your condition? e.g., cosmetic, breast, genital); (iii) substance abuse history (Have you ever received treatment for a substance abuse problem? Lost a job or a relationship? Been arrested?); (iv) mental illness history (Have you ever received treatment for a mental health problem other than this condition? Been hospitalized for psychiatric problems?); (v) genital mutilation history (Have you ever attempted to cause physical damage to your genitals? breasts?); and (vi) suicide attempts (Have you ever made an attempt to end your life?). A subgroup of the original sample also completed the 400-question short form Minnesota Multiphasic Personality Inventory (MMPI). The MMPI might offer a less biased perspective on the coexistence of other mental illness than the clinical interview and has been employed previously in studies of transgendered individuals (Rosen, 1974; Finney et al., 1975; Roback et al., 1976; Fleming et al., 1981).

16

Cole, O'Boyle, Emory, and Meyer

Subjects Since 1980, 435 individuals presented to a gender clinic with a selfdiagnosed condition of gender dysphoria and were requesting treatment for their condition (e.g., counseling, hormones, surgery): 318 were maleto-female candidates with a mean age ± SD of 32 ± 9 and 117 were female-to-male candidates with a mean age of 30 ± 8 years.

RESULTS Subjects presented from a variety of educational and occupational backgrounds. To qualify these data in more understandable terms, the Hollingshead Index of Social Position was employed (Meyers and Bean, 1964). The Hollingshead scale scores of the male-to-female transsexuals are presented in Table I. Both groups' educational means placed them between high school graduate (4.0) and partial college (3.0). There was no statistical difference between these groups. In terms of the Occupational Scale, both groups' mean score placed them between 4.0 (clerical and scales workers, owners of small businesses) and 5.0 (skilled manual employees). Again, there was no statistical difference between these groups. Hollingshead social position scores, calculated from the educational and occupational scores, were not significantly different and suggest both groups were primarily middle class, between III and IV. One could reason that distance from the clinic kept all but the most functional away. For this reason we examined samples in terms of how far they traveled to the clinic (Table II). Approximately one third of the sample lived less than SO miles from the clinic, one third lived between 50-300 miles from the clinic, and the remaining third lived over 300 miles away. Table L Subject Characteristics Male-to-female Female-to-male Hollingshead scale (AT = 117) (N = 318) Educational Occupational Social positiona aIndex

3.5 ± 1.2 4.2 ± 1.3 43.7 ± 12.4

3.7 ± 0.9 4.5 ± 1.2 46.2 ± 10.7

of Social Position: Occupational scale x 7 + Educational Scale x 4 = Social Position Scale.

Comorbidity of Psychiatric Disorders

17

Table II. Profile for Distance (in Miles) from Clinic for Total Sample Studieda

Mileage

«

Patients with add'l psychiatric % of Sample diagnoses

0-50 51-300 >300

133 173 129

aX2(2, N

= 435) = 4.6, ns.

30.6 39.7 29.7

%

18

13.5

14 8

8.1 6.2

Clinical Interview and Survey Findings Table III lists pertinent psychosocial and medical factors endorsed by the population as obtained from the clinical interview and biographical questionnaire. Two thirds or more of both groups were undergoing their respective hormone therapies, suggesting a commitment and involvement in the "real life" transition process. Additionally, nearly one third of the male-to-female individuals and one half of the female-to-male individuals had undergone surgery specific to their conditions. In most cases, such surgeries were cosmetic to aid in appearance (e.g., to the face) or were procedures to the breasts (e.g., augmentation for the male-to-female or reduction for the female-to-male). Very few had completed the genital procedures related to sexual reassignment surgery, hence a chief reason why they applied to the gender clinic program. Twenty-nine percent of the male-to-female and 26% of the femaleto-males reported past substance abuse problems which included such elements as having received treatment from a substance abuse counselor, participated in Alcoholics Anonymous or Narcotics Anonymous, or experienced problems with job or relationships as a consequence of their substance usage. In virtually all of the cases, subjects described their substance Table III. Factors Studied with Sample (Percentage Having Characteristics) Factor Hormone treatment Surgical treatment Substance abuse history Mental illness history Genital mutilation history Suicide attempts

M-to-F (N = 318)

F-to-M (N = 117)

Total (N = 435)

65 28 29 9 8 12

83 49 26 9 1 21

70 34 28 9 6 15

18

Cole, O'Boyle, Emory, and Meyer Table IV. Associated Psychiatric: Illness

Axis I Axis II

Male-to-female (N = 318)

Female-to-male (N = 117)

n

%

n

%

18 12

6 4

5 4

4 3

abuse problems as associated with trying to deal with their gender dysphoria issues. In other words, prior to seeking specific gender treatment, or even fully recognizing the nature of their problem, usage of a variety of chemicals served to ease the pain of their dilemma. When questioned about past psychiatric treatment, only 9% of the total group indicated past treatment for diagnosed psychiatric conditions other than gender dysphoria or substance abuse. This finding was examined in further detail. Specifically, the question was asked as to what types of mental disorders had been identified. Table IV provides a breakdown of the frequency of Axis I and Axis II diagnoses in both groups. (This classification was based on usage of DSM-III-R which was employed during the period that the data were collected and examined.) An almost equal split occurred between the groups with respect to the Axis I and Axis II identified problems. Further examination of the specific diagnoses was conducted, and Table V indicates the types of problems reported. Depression was the most common Axis I diagnosis. Diagnoses of bipolar disorder and schizophrenia were also well-represented dominant Axis I problems. Borderline and schizoid personality disorder were most prominent in the Axis II category. Only one person in that entire sample, presenting as male-to-female, had multiple diagnoses, those being paraphilia, borderline personality disorder, and antisocial personality disorder. This individual was, therefore, recorded in each of these categories. Records indicated that he came from a very dysfunctional family of origin, had various run-ins with the law, and also had molested his young daughter, resulting in a lengthy prison term. Now released from prison, he was seeking treatment for his long-standing self-diagnosed problem of gender dysphoria. Nearly two thirds of the individuals in both groups with prior Axis I diagnoses were taking medication for these conditions (e.g., antidepressants, neuroleptics). The distance from the clinic was not significantly related to having additional psychiatric diagnoses (Table II) x2(2, N = 435) = 4.6, ns.

Comorbidity of Psychiatric Disorders

19

Table V. Specific Types of Comorbid Psychiatric Diagnoses

n

Diagnosis Male-to-female Axis I Major depression Bipolar disorder Schizophrenia Paraphilia Substance abuse

9 3 3 1 2

Axis II Borderline personality Schizoid personality Antisocial personality Mental retardation

6 4 2 1

Female-to-male Axis I Major depression Bipolar disorder Schizophrenia

2 2 1

Axis II Borderline personality Schizoid personality Explosive personality

2 1 1

Regarding past genital mutilation attempts, 8% of the male-to-female and 1% of the female-to-male individuals reported such behavior (Table III). In general this involved such activities as taping, hitting, or squeezing the genitals out of intense frustration. Only a few individuals had cut their genitals with a knife or other object. Finally, the question of suicide attempts revealed that 12% of the male-to-female and 21% of the female-to-male subjects had engaged in such activity. Further questioning revealed that in virtually all of these cases this behavior was attributed to intense frustration and exasperation over the gender dysphoric condition. This finally reached a breaking point over such reported issues as feeling isolated and not able to talk to others, being rejected by family or an intimate partner, or disgust with one's anatomic state and feeling that it could never change. AH of the suicidal attempts occurred prior to individuals' becoming involved in specific gender treatment. None of these patients has had a suicide attempt since beginning therapy for his/her gender issues.

20

Cole, O'Boyle, Emory, and Meyer

MMPI Findings The MMPI was administered to 137 individuals (31% of the total original sample of 435). Ninety-three male-to-females (29% of the original M-F sample) and 44 female-to-males (38% of the original F-M sample) were included. The MMPI was used, rather than the MMPI-2 because data were collected using this instrument prior to the development and release of the newer version. It was felt, in order to be consistent, that usage of the original form was a better choice. Table VI and VII denote the mean T scores for both subgroups. Results are presented for each subgroup based on anatomic sex and then on self-perceived sex. The most striking observation is the relative absence of psychopathology per se in these profiles (i.e., mean T scores above 70). This suggests no substantial indication of problems indicative of Axis I diagnoses (e.g., depression, schizophrenia, anxiety, mania, etc.). Of particular note, the Mf score of the male-to-female subgroup was significantly elevated (above T score of 70) when plotted on the male profile but fell within the normal range when plotted on the female profile, consistent with the self-perception of this group (Fig. 1). However, the Mf score of the female-to-male subgroup was not significantly different when plotted as female or male. Both scores were below a T score of 70 with a slight tendency for the Mf score to appear closer to the norm when Table VI. MMPI Results (Mean T Scores) on Male-to-Female Group (n = 93) Plotted on Both Profiles Female profile Male profile (self-perceived sex) (anatomic sex) MMPI scale x±SD %>70 x± SD %>70

L F K Hs D

Hy Pd Mf Pa Pt Sc Ma Si

52 ± 7.7 55 ± 7.4 57 ± 7.8 52 ± 58 ± 59 ± 63 ± 81 ± 58 ± 57 ± 60 ± 57 ± 52 ±

9.1 11.4 8.5 11.1 10.8 10.0 9.1 11.1 10.7 10.7

1.0 4.4 7.6

52 ± 7.7 55 ± 7.4 57 ± 7.8

6.0

48 ± 52 ± 55 ± 63 ± 47 ± 58 ± 52 ± 58 ± 57 ± 52 ±

12.0 8.7 25.7 89.3 13.8 11.7 16.2

8.5 7.4

7.2 10.5 8.3 11.1 10.2 10.0 7.4 9.3 10.1 10.7

1.0 4.4

7.6 2.4 7.6 6.5 25.7

2.2 13.8

2.2 10.8

8.5 7.4

Comorbidity of Psychiatric Disorders

21

Table VII. MMPI Results (Mean T Scores) on Female-to-Male Group (n = 44) Plotted on Both profiles

Female profile (anatomic sex)

Male profile (self-perceived sex)

x ± SD

%>70

x± SD

%>70

L F K

51 ± 6.4 53 ± 6.6 56 ± 8.1

0 0 4.6

51 ± 6.4 53 ± 6.6 56 ± 8.1

0 0

Hs D

52 ± 49 ± 51 ± 62 ± 65 ± 55 ± 53 ± 57 ± 59 ± 52 ±

2.3 4.6

56 55 56 62 62 55 58 59 59 52

MMPI scale

Hy Pd

Mf Pa Pt

Sc Ma Si

9.2 10.9 8.0 12.3 11.0 10.1 10.4 11.7 11.7 10.7

4.6 18.2 38.5

6.9 6.9 11.5 20.5

4.6

± ± ± ± ± ± ± ± ± ±

11.7 11.9 8.9 12.3 10.9 10.1 13.1 14.5 11.7 10.7

4.6

13.7 13.7 4.6 8.2 22.7

6.9 20.5 20.5 20.5

4.6

viewed from the self-perceived sex rather than the biologic sex for this group (Fig. 2). DISCUSSION

Our findings, based on retrospective analyses of clinical interview and questionnaire data and a subgroup of completed MMPIs, are consistent with our clinical experience over the last several decades. Specifically, gender dysphoric individuals appear to be relatively "normal" in terms of an absence of diagnosable, comorbid psychiatric problems. In fact, the incidence of reported psychiatric problems is similar to that seen in the general population (Robins etal.,1984; Weissman and Myers, 1978; Weissman et at., 1991). Similarities in incidence included depression, bipolar disorder, and schizophrenia. These are highlighted in Table VIII which compares our gender dysphoric sample and three surveyed general populations. In functional terms, the majority of such individuals are able to hold down employment, develop lasting friendships and relationships, and pursue leisure activities of interest. A "psychiatric" population, functionally speaking, typically exhibits significant difficulties in trying to accomplish these tasks. Although a small percentage of gender dysphoric individuals in this sample had prior identifiable psychiatric problems (7-10%), this is not inconsistent with the general population (Robins etal.,1984). Since the general popu-

22

Cole, O'Boyle, Emory, and Meyer

Fig. 1. MMPI testing of male-to-female transsexuals: A. mean scores plotted against male scale; B. mean scores plotted against female scale.

lation studies were based on the Diagnostic Interview Schedule, they detected individuals with symptoms who had not sought mental health or medical services. The lack of formal, standardized interview schedules or symptom checklists can be cited as a weakness of this study, as such techniques would have added to the measurement of possible psychopathology. This may account for the higher incidence in the general population than our gender dysphoric sample. Several studies have employed the MMPI around assessment issues involving transsexualism (Rosen, 1974; Finney et al., 1975; Roback et al., 1976; Fleming et al., 1981). Most notably, elevations have been identified on the Masculinity-Femininity (MF) and the Psychopathic Deviate (Pd) scales. One study went further by investigating both presurgical and postsurgical individuals, noting an interesting finding with respect to the MF scale (Fleming et al., 1981). Specifically, the male-to-female groups (both pre- and postsurgical) scored well above the mean on Mf as male but well within the normal range on Mf as female. This is consistent with the find-

Comorbidity of Psychiatric Disorders

23

Fig. 2. MMPI testing of female-to-male transsexuals: A. mean scores plotted against male scale; B. mean scores plotted against female scale.

Table VIII. Comparison of Lifetime Incidence of Major Psychiatric Diseases Gender dysphoric sample

General population"

% Male-female

% Female-male

%Male

% Female

Depression

2.8

1.7

2.3-4.4

4.9-8.7

Bipolar

0.9

1.7

0.8-1.1

0.5-1.3

Schizophrenia

0.9

0.9

1.0-1.2

1.1-2.6

aRobins et at., 1984.

ings of this study. On the other hand, the female-to-male groups (both preand postsurgical) scored within the normal range on Mf as female and well above the normal range on Mf as male. The current study's finding of most female-to-male patients scoring below 70 is different in this regard.

24

Cole, O'Boyle, Emory, and Meyer Table IX. MMPI Results (Mean T Scores) on Outpatient Psychiatric Sample Presented by Diagnostic Category" Borderline patients (n = 27)

Affective disorders (n = 67)

Other personality disorders (n = 47)

Adjustment and anxiety disorders (n = 64)

L F K

48 73 46

49 67 53

51 62 55

51 58 56

Hs D

70 75 73 81 60 64 78 64 70 65

71 75 74 77 61 63 78 64 71 63

66 69 70 74 61 60 70 50 69 61

66 66 67 67 60 56 67 57 64 61

MMPI scale

Hy

Pd Mf Pa Pt

SiSc

Ma Si aFrom

Lloyd et al. (1983).

How do these MMPI findings with the gender dysphoric sample compare to an identified psychiatric population? The MMPI findings for gender patients were near normal. Table IX illustrates the mean MMPI T scores reported by Lloyd et al., (1983) for four categories of patients: borderline, affective, other personality disorders, and adjustment-anxiety disorders (a "residual" group comprised of individuals with marital problems and the like). It can be seen that significant elevation (i.e., T scores above 70 occur on a number of scales, indicating problems with such areas as depression (D), sociopathy (Pd), anxiety (Pt), and somatic preoccupations (Hs). One would expect to see such elevations in a clinically identified psychiatric population. This is particularly notable for the borderline and affective subgroups as compared to the other subgroups. On the contrary, there appeared to be no elevations on the Mf scale across groups. Rather, these scores appear to be the most stable and least elevated of any indices on the profiles. A consistent theme presented by the individuals interviewed in this study was that they perceived themselves as happier, increasingly competent, and more productive in terms of vocational and avocational activities once they finally acknowledged their gender dysphoria. In other words, beginning the treatment process resulted in a sense of stability which was reflected by a decrease in self-destructive behavior (e.g., genital mutilation, suicide attempts, substance abuse). While contemporary gender treatment

Comorbidity of Psychiatric Disorders

25

does not result in a "cure" for transsexualism, developing insight into one's problem, meeting others with similar concerns, and taking behavioral steps to change (e.g., the real life test) do appear to lead to improvements in self-confidence and relating to the general environment at large. Conclusion This study should help to clear up certain misperceptions about gender dysphoria per se. Specifically, individuals presenting with gender dysphoria often do not have problems indicative of a coexisting psychiatric illness such as schizophrenia or major depression. Instead, these finding suggest that gender dysphoria is usually an isolated diagnosis. This study does not imply that contemporary evaluation and treatment approaches be abolished. The Standards of Care developed by the Harry Benjamin International Gender Dysphoria Association painstakingly evolved after much thought and discussion were shared among leaders in the field with many combined years of professional experience (Walker et al., 1985). Instead, it is felt that these elements (i.e., real life test, ongoing psychiatric counseling and support) are critical in helping individuals work through the multiple psychosocial, endocrine, and surgical issues associated with this diagnosis.

ACKNOWLEDGMENT The authors thank Nita Brannon for typing the manuscript.

REFERENCES American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders, 4th ed, American Psychiatric Association, Washington, DC. Benjamin, H. (1966). The Transsexual Phenomenon, Julian Press, New York. Buhrich, N., and McConaghy, N. (1978). Parental relationships during childhood in homosexuality, transvestism, and transsexualism. Aust. N. Z. I. Psychiat. 12:103-108. Cauldwell, D. (1949). Psychopathia transsexualis. Sexology 16: 274-280. Cole, C. M., Emory, L. E., Huang, T., and Meyer, W. J. (1994). Treatment of gender dysphoria (transsexualism). Tex. Med. 90: 68-72. Finney, J., Brandsma, J., Tondow, M, and Lemaistre, G. (1975). A study of transsexuals seeking gender reassignment Am. J. Psychiat. 132: 962-964. Fleming, M., Cohen, D., Salt, P., Jones, D., and Jenkins, S. (1981). A study of pre- and post-surgical transsexuals' MMPI characteristics. Arch. Sex. Behav. 10:161-170. Green, R., and Money, J. (1969). Transsexualism and Sex Reassignment, Johns Hopkins Press, Baltimore, MD.

26

Cole, O'Boyle, Emory, and Meyer

Halle, E., Schmidt, C, and Meyer, J. (1980). The role of grandmothers in transsexualism. Am. J. Psychiat. 137: 497-498. Hirschfeld, M. (1945). Sexual Pathology, Emerson Books, New York. Krafft-Ebing, R. von. (1894). Psychopathia Sexualis (Trans, by F. Graz in 1939), Pioneer Publications, New York. Lloyd, C, Ovberall, J., Kimsey, L., and Click, M. (1983). A comparison of the MMPI-168 profiles of borderline and non-borderline patients. /. Nerv. Ment. Dis. 171: 207-215. Lothstein, L. (1979a). Psychodynamics and sociodynamics of gender dysphoric states. Am. J. Psychother. 33: 214-238. Lothstein, L. (1979b). The aging gender dysphoria (transsexual) patient Arch. Sex. Behav. 8: 431-434. Lothstein, L. (1983). Female-to-Male Transsexualism: Historical, Clinical and Theoretical Issues. Routledge and Kegan Paul, Boston. Lundstrom, B., Pauly, I., and Walinder, J. (1984). Outcome of sex reassignment surgery. Acta Psychiat. Scand. 70: 289-294. Meyers, J. K., and Bean, L. L. (1964). A Decade Later: A Follow-Up of Social Class and Mental Illness, Wiley, New York. Morgan, A. (1978). Psychotherapy for transsexual candidates screened out of surgery, Arch. Sex. Behav. 7: 273-283. Ovesey, L., and Person, E. (1973). Gender identity and sexual psychopathology in men: A psychodynamic analysis of homosexuality, transsexualism, and transvestism. /. Am. Acad. Psychoanal. 1: 54-59. Pauly, I. (1990). Gender identity disorders: Evaluation and treatment. /. Sex Educ. Ther. 16: 2-24. Roback, N., McKee, E., Webb, W., Abramowitz, C., and Abramowitz, S. (1976). Psychopathology in female sex-change applicants and two help-seeking controls. /. Abn. Psychol. 85: 430-432. Robins, L N., Helzer, J. E, Weissman, M. M., Orvaschel, H., Gruenberg, E., Burke, Jr., J. D., and Regier, D. A. (1984). Lifetime prevalence of specific psychiatric disorders in three sites. Arch. Gen. Psychiat, 41: 949-958. Rosen, A. (1974). Brief report of MMPI characteristics of sexual deviation. Psychol. Rep. 35: 73-74. Stoller, R. (1968). Sex and Gender On the Development of Masculinity and Femininity, Science House, New York. Stoller, R. (1975). The Transsexual Experiment, Hogarth, London. Walker, P., Berger, J., Green, R., Laub, D., Reynolds, C, and Wollman, L. (1985). Standards of Care: The hormonal and surgical sex reassignment of gender dysphoric persons. Arch. Sex. Behav. 14: 79-90. Weissman, M. M., Bruce, M. L., Leaf, P. J., Florio, L. P., and Holzer III, C. (1991). Affective

Disorders. In Robins, L. N., and Regier, D. A. (eds.), Psychiatric Disorders in America: The Epidemiologic Catchment Area Study, Free Press, New York, pp. 53-80. Weissman, M. M., and Myers, P. S. (1978). Psychiatric disorders in U.S. urban populations. Am. J. Psychiat 135: 459-465.

Archives of Sexual Behavior, Vol. 26, No. 1, 1997

The Relationships Among Ejaculatory Control, Ejaculatory Latency, and Attempts to Prolong Heterosexual Intercourse Guy Grenier, M.A.,1.2 and E. Sandra Byers, Ph.D.1

Although premature ejaculation (RE) is considered the most common male sexual dysfunction, progress in understanding it has been hampered by the lack of a commonly accepted definition. Several different criteria have been used to assess RE with no attempt to validate the extent to which they are related. The current study assessed, in a sample of university men, the occurrence and relationships among four commonly applied RE criteria: perceived control over the occurrence of ejaculation, latency from vaginal penetration to ejaculation, satisfaction with perceived degree of ejaculatory control, and concern over the occurrence of rapid ejaculation. Other aspects of ejaculatory behavior were also assessed such as the thoughts and techniques men used to prolong intercourse and delay ejaculation. Results indicated that although the four RE criteria were significantly correlated, the magnitudes of these correlations were small. This suggests that these commonly used RE criteria are largely independent and are not interchangeable and that research in this area needs to adopt a multivariate approach to assessment. Men's erotophilia/erotophobia was not related to RE. While use of several ejaculatory delaying techniques were individually and jointly predictive of ejaculatory control and/or ejaculatory evidence, there was no strong support for any specific pattern of behavior that is related to better control and longer latencies. Further, experts were not able to distinguish the ejaculatory delaying techniques of the men with the poorest control and shortest latencies from those of the men with the best control and longest latencies. KEY WORDS: rapid/premature ejaculation; ejaculatory latency; ejaculatory control; ejaculation delaying techniques; heterosexual intercourse. 1Department

of Psychology, University of New Brunswick, Fredericton, New Brunswick, Canada E3B 6E4. 2To whom correspondence should be addressed.

27 0094-0002/97/0020-0027$12.5

C 1997 Plenum Publishing Corporation

28

Grenier and Byers

INTRODUCTION

Premature or rapid ejaculation (RE) has been identified as the most common male sexual dysfunction and is estimated to affect between one third and three quarters of men (Chesney et al., 1981; Kinsey et al., 1948; McCarthy, 1988; Nathan, 1986; Reading and Wiest, 1984; Spector and Carey, 1990). Clinical discussion and case reports indicate that ejaculating sooner than desired may, among other things, substantially detract from the enjoyment of intercourse, reduce the frequency with which sensual and sexual interactions are attempted, and negatively impact on self-esteem (Kaplan, 1989; McCarthy, 1988; Zilbergeld, 1992). Despite the pervasive nature of RE and its potential negative consequences, understanding of this condition has advanced little since Semans (1956) first proposed the start-stop technique (Grenier and Byers, 1995; St. Lawrence and Madakasira, 1992). If sex researchers and therapists are to make greater progress in understanding RE, it must be adequately described and defined. However, there is a continued absence of an empirically based and commonly accepted operationalized definition of RE. RE has been defined using any of the following criteria: the latency to ejaculation after intromission, the perceived degree of ejaculatory control, the occurrence of antiportal ejaculation, the degree of distress a man (or his partner) feels over ejaculating rapidly, the frequency with which rapid ejaculation occurs, the number of intravaginal thrusts accomplished prior to ejaculation, a female partners' experience of orgasm or "sexual satisfaction," and/or the amount of sexual stimulation required to bring on ejaculation. These criteria have been used individually or in various combinations to distinguish RE and non-RE men. However there has been little or no investigation of the relationships among these criteria. The use of different combinations of criteria by different investigators makes crossstudy comparisons problematic. The assumption among some researchers that the various criteria (or combination of criteria) are equivalent remains to be validated empirically. Consequently, our first goal was to examine the relationships between some of the most commonly applied RE criteria: latency to ejaculate, perceived ejaculatory control, satisfaction with ejaculatory control, and concern over rapid ejaculation. The occurrence of involuntary antiportal ejaculation was also used as a behavioral indicator. Another factor contributing to the RE definitional confusion has been the inconsistent application of the same criteria. One example is latency to ejaculate. Latency values which have been used to dichotomize men into RE and non-RE groups have ranged from 1 min or less (e.g., Cooper and Magnus, 1984; Segraves et al., 1993) up to 7 min (Schover et al., 1982). Clearly, some men in the RE group in one study would have been in the

Rapid Ejaculation

29

non-RE group in another study and vice versa. The relatively arbitrary manner in which cutoff values are selected, their inconsistent application, and the ensuing dichotomizing of samples further contributes to the current lack of comparability across RE studies. In keeping with the suggestions of Grenier and Byers (1995), we operationalized the RE variables as continuous rather than dichotomous variables, allowing for more statistically powerful and conceptually meaningful comparisons among criteria. Another aspect of RE, and ejaculatory behavior in general, which has not received sufficient attention, is the behavior of men during intercourse. Specifically, little data are available concerning exactly what men do in attempting to delay ejaculation and prolong intercourse. This applies equally to men who ejaculate rapidly and those who ejaculate after more protracted sexual interactions. Zilbergeld (1992) offered some information (albeit unsystematic and anecdotal) concerning what men who report good ejaculatory control do to prolong intercourse. In his discussion of RE, he contended that men without RE make adjustments to their thrusting movements/pace of thrusting during intercourse to specifically increase their ejaculatory latencies. However, he provided no evidence that men with RE do not use the same adjustments in thrusting movements/pace of thrusting behavior. Also, it is not known whether there are specific adjustments or behavioral techniques that are particularly effective at decreasing the occurrence of RE. Consequently, the second purpose of the current study was to gather systematic information concerning the intercourse/ejaculatory behavior and cognitions of men with varying latencies and reported degrees of control. We also investigated whether experts in the field of sexuality could distinguish the ejaculatory delaying technqiues used by men with longer ejaculatory latencies and good ejaculatory control from those reported by men with shorter latencies and poorer ejaculatory control. Fisher et al. (1988) proposed that ones' affective and evaluative disposition to sexual cues (i.e., degree of erotophobia/erotophilia) could be both a cause and an effect of sexual dysfunction. If true, men who are dissatisfied with their ability to prolong intercourse might also be less positively disposed to sexual cues (i.e., be more erotophobic) than are men who are more satisfied (i.e., be more erotophilic). Further, erotophilic men, because of their greater presumed comfort with sexual expression, should be more likely to have explored more sexual behaviors and consequently be more likely to have discovered a greater variety and greater number of effective ejaculation delaying techniques than would men who are more erotophobic. Therefore we explored the relationships between the erotophobia/erotophilia dimension and both the occurrence of RE and the use of ejaculation delaying techniques.

30

Grenier and Byers METHOD Subjects

In an Introductory Psychology course, 112 heterosexual male undergraduates completed a survey concerning their sexual and ejaculatory behavior and received extra credit towards their final grade. Two participants who were significantly older than the rest of the sample were removed from the analyses leaving a total sample of 110. Although 86% (95 men) of the sample elected to take home self-monitoring forms, only 57 men returned the questionnaires for analysis. Materials

A background questionnaire requesting basic information concerning demographics and sexual history was created for this study. To ensure common usage and understanding of sexual terms, definitions of basic concepts (i.e., sexual intercourse, ejaculation, "cumming," orgasm, foreplay, penetration, and withdrawal) were included in the questionnaire. As a search of the literature did not uncover any standardized instruments concerned with ejaculatory behavior and control, the Men's Sexual Behavior Questionnaire was developed. The questionnaire consisted of five parts, with each part focusing on a different aspect of sexual and ejaculatory behavior. (A copy of the Men's Sexual Behavior Questionnaire may be obtained from the authors.) In Part A, participants provided information concerning their most recent sexual intercourse experience. Participants provided information regarding the length of foreplay, latency to ejaculate, use of withdrawal, number of ejaculations, antiportal ejaculation, and the perceived imminence of ejaculation at the outset of intercourse during this experience. The following explanation of the concept of ejaculatory control was then provided: Some men ejaculate more slowly than they want to, some ejaculate more quickly than they want to, and some ejaculate exactly when they want to. Being able to control when you ejaculate means that during sexual activity, you are able to decide when your ejaculation is going to occur rather than it just happening no matter what you do. Different men have different amounts of ejaculatory control. There is no "correct" amount of control to have.

Following this, participants completed four questions concerning ejaculatory control "when you ejaculate for the first time during a sexual intercourse experience." Participants indicated on 8-point scales their degree of ejaculatory control (no control to absolute control), their concern

Rapid Ejaculation

31

over ejaculating faster than they wanted to (no concern to extreme concern), their concern over ejaculating slower than they wanted to (no concern to extreme concern), and their satisfaction with their ability to control when they ejaculated (extremely dissatisfied to extremely satisfied). In Part B, participants provided information concerning their overall intercourse experience. Participants reported their average ejaculatory latency, average number of ejaculations per intercourse experience, and longest ejaculatory latency. On two scales ranging from 0 to 100%, participants reported the percentage of time they experienced antiportal ejaculation and the percentage of time they felt they had control over when they first ejaculated in an intercourse situation. Finally, participants completed the same four questions concerning ejaculatory control used in Part A with respect to their average degree of ejaculatory control, concern over ejaculating faster or slower than they wanted to, and satisfaction with their ejaculatory control. Part C of the questionnaire assessed ejaculatory delaying techniques. To explain the concept of an ejaculatory delaying technique without unduly leading or contaminating participants' answers, participants were provided with an explanation of the physiological process of yawning along with a description of a number of techniques commonly used to try to stifle a yawn. They were then asked to list any techniques they had ever used to stifle or delay the physiological process of ejaculation. Additionally, subjects were asked to provide specific details as to the content of any distracting thoughts they used to attempt to prolong intercourse. Part D consisted of the short form of the Sexual Opinion Survey ((SOS); Fisher et al., 1988). The SOS short form is a 5-item measure of erotophobia-erotophilia or participants' positive-negative affective and evaluative disposition to sexual cues. The full SOS has been shown to be a reliable measure (Cronbach's alpha ranging between .85 and .90) and its validity is supported in that it accurately and consistently predicts sexual approach/avoidant behavior across samples and across cultures (Fisher et al., 1988). Part E contained a checklist of 28 common ejaculatory delaying techniques which were gathered from the literature as well as from an informal survey of men (see Table IV). Participants checked off all the techniques they had ever used to delay ejaculation. They also rated how effective each technique was in prolonging ejaculation on a 5 point scale (very ineffective to very effective). Participants were then asked to list any ejaculatory delaying techniques they used that were not included in the list and had not been mentioned in Part C.

32

Grenier and Byers

A self-monitoring form of the Men's Sexual Behavior Questionnaire was also developed in order to gather ejaculation information immediately after an intercourse experience. Participants provided basic information concerning their sexual partner during the experience (relationship status, etc.), and completed Parts A and E of the original questionnaire. Procedure

Participants were tested alone or in small groups, seated so as to maximize their and other participants' privacy. After reading and signing an informed consent document, participants were given the Men's Sexual Behavior Questionnaire. After participants had completed the questionnaire, they were given the option of earning a second experimental credit point by taking home and completing three self-monitoring questionnaires. Those who chose not to participate further in the study were given a debriefing form, provided an opportunity to record their mailing address to receive study results when they were available, and thanked for their participation. Participants who chose to complete the second part of the study were instructed to complete a self-monitoring questionnaire, to a maximum of three, after each time they had intercourse over the following 2 week period. At the end of the 2 week period or when the three questionnaires had been complete, whichever came first, they were to return the questionnaire to the experimenter and receive their additional credit. Participants who chose to self-monitor received the debriefing form and opportunity to receive study information after they returned their selfmonitoring questionnaires. The ejaculatory delaying techniques provided by participants in Part C were sorted by two independent coders into the 28 categories that were used in Part E. Coder agreement was 90%. Discrepancies were resolved through consultation and with the addition of two new categories: unspecified changes in behavior and the use of the squeeze technique. No other delaying techniques that were not included in the 28-item checklist were reported. Following initial analyses of the data, the delaying techniques reported by the 15 men with the greatest control and longest latencies and the 15 men with the least control and shortest latencies were arranged in random order and sent to the members of the Canadian Sex Research Forum (CSRF). The CSRF is an interdisciplinary group of sex educators, therapists, and researchers. Members were asked to read each of the 30 responses and make a judgment as to whether they believed the participant was likely to have reported having good or poor ejaculatory control.

Rapid Ejaculation

33

RESULTS Sexual Characteristics The average age of participants was 19.8 years (SD = 2.3). On average, participants began dating when they were 14.2 years old (SD = 1.8) and began having intercourse when they were 16.3 (SD = 1.7). The average score on the erotophobia/erotophilia scale was 11.3 out of a possible 25 (SD = 5.2). On average the men had had intercourse with one partner during the month previous to the study (SD = 1.1) and reported having intercourse an average of six times over the same period (SD = 10.9). To get a more complete understanding of the men's sexual interactions, they reported on a number of additional aspects of their sexual behavior during their most recent intercourse experience. They reported that, on average, the length of foreplay was 18 min (SD = 11.7). Nineteen percent of the sample reported ejaculating at least once during foreplay. Almost 7% of the sample reported that they were "extremely close" to ejaculating when intercourse began, while 26.9% reported that they were "Very close." The average number of ejaculations during intercourse was 1.7 (SD = 0.86).

Relationships Among Rapid Ejaculation Criteria Table I presents the ranges, means, and standard deviations for the four RE criterion variables: latency to ejaculate (Latency), perceived ejaculatory control (Control), satisfaction with ejaculatory control (Satisfaction), and concern over rapid ejaculation (Concern), as well as the average occurrence of involuntary antiportal ejaculation.3 According to the men's reports, the average latency to ejaculate following vaginal penetration was just under 14 min. The average amount of ejaculatory control men felt they had over the timing of their ejaculation fell in the middle of the 8point scale, as did their concern over rapid ejaculation and satisfaction with ejaculatory control. On average involuntary, antiportal ejaculation occurred during 1 out of every 10 intercourse opportunities. 3Participants provided

data on both their most recent sexual intercourse experience and their overall sexual intercourse experience. Due to the substantial correlations between each of the RE criteria from these two sources, correlations related to retrospective reports of the men's most recent experience were not analyzed further. Also, both Antiportal and Latency had univariate outliers. For subsequently analyses, these outliers were recorded to values that were just beyond 3 standard deviations beyond the mean and continuous with the variables' distribution.

34

Grenier and Byers Table I. Ranges, Means, and Standard Deviations for Rapid Ejaculation Criteriaa Sample range Measure

Average ejaculatory latency following vaginal penetration Average self-rated ejaculatory control Average concern over rapid ejaculation Average satisfaction with ejaculatory control Percentage occurrence of antiportal ejaculation an

Min

Max

X

1 1 1 1 0

45 8 8 8 60

13.6

SD

9.9 4.2 1.9 4.7 2.1 4.7 1.6 9.9 13.0

ranged between 107 and 110.

Table II presents the zero-order correlations among the RE criteria. All but two of the correlations were significant. Men with longer ejaculatory latencies reported more control over when they ejaculated, were more satisfied with their perceived degree of ejaculatory control, and experienced involuntary antiportal ejaculation less often than did men with shorter latencies. Similarly, men reporting more ejaculatory control were less concerned about rapid ejaculation, more satisfied with their ability to control the timing of ejaculation, and experienced involuntary antiportal ejaculation less often than did men reporting lesser perceived ejaculatory control. In addition, men who reported more concern about ejaculating rapidly felt they had less control over the timing of their ejaculation and were less satisfied with this degree of control than were men who were less concerned about ejaculating rapidly. Finally, as the incidence of involuntary antiportal ejaculation decreased, satisfaction with perceived Table II. Zero-Order Correlations Among Ejaculatory Latency and Control Variables (N = 110)

1 1. Latencya 2. Control* 3. Concernc 4. Satisfactiond 5. % Antiportal ejac.e aAverage

2

3

4

—

.31g -.14 .35* -22f

—g -.27 .68g -.22f

-— .38*

.16

—g -.36

latency to ejaculate following penetration. *Average degree of perceived ejaculatory control. cAverage concern with ejaculating faster than desired. dAverage satisfaction with ability to control the timing of ejaculation. ePercentage of intercourse experiences in which involuntary antiportal ejaculation occurred. fp < .05. gp < .005.

Rapid Ejaculation

35

ejaculatory control increased. However, the magnitude of almost all these correlations is, at best, moderate. With the exception of the relationship between Control and Satisfaction, which accounted for 46% of the variance, the other variables which were found to be related shared only between 5 and 14% of their variance. This strongly suggests that the concepts of ejaculatory latency, perceived degree of ejaculatory control, concern over rapid ejaculation, satisfaction with ejaculatory control, and the occurrence of involuntary antiportal ejaculation are relatively independent.

Traditional Approaches to Defining Rapid Ejaculation To assess the consequences of adopting various common approaches to identifying men with RE, the various RE criteria were dichotomized. Within the literature, reported ejaculatory latencies differentiating nonRE and RE men have ranged from 1 to 7 min. Applying these criteria, between 1 and 36% of the sample could be classified as having RE. Adoption of a 2-min cutoff, the RE criterion currently adopted most frequently, would result in identification of 4% of the current sample as rapid ejaculators. When the criterion of perceived ability to control the occurrence of ejaculation was dichotomized, 53% of the sample reported limited ejaculatory control (i.e., a score of 4 of less on an 8-point scale ranging from no control to absolute control). When only those participants who reported no control were included (i.e., a score of 1), 8% of the sample would be classified as having RE. In terms of concern over ejaculation, 59% of the sample registered considerable concern about ejaculating faster than they wanted to (i.e., score of 5 or more on an 8-point scale ranging from no concern to extreme concern), while 5% of the sample reported an extreme degree of concern. Only 12% repotted no concern. In addition, 39% of the sample reported considerable dissatisfaction with their ability to control when they ejaculated (i.e., a score of 4 or less on an 8-point scale ranging from extremely dissatisfied to extremely satisfied), while 3% reported being extremely dissatisfied. If the two most commonly used RE criteria, latency to ejaculate and perceived degree of ejaculatory control, are applied simultaneously using the most inclusive cutoff values, (i.e., a latency of 7 min or less and a 4 or less on the control scale), cross-tabulation indicates that 24% of the current sample would be classified as rapid ejaculators. However, if the most commonly used latency of 2 min is applied along with the criterion

36

Grenier and Byers

of little or no control, none of the men in the current sample would meet rapid ejaculation criteria. In sum, depending on how RE was operationalized, we found substantial variation in the prevalence of RE. Thoughts Used to Delay Ejaculation Ninety-seven percent of the sample were able to recall or generate at least one delaying technique in Part C of the questionnaire. However, these techniques were essentially redundant with those reported on the checklist from Part E, and consequently were not analyzed further. Participants were also asked to provide details concerning the content of any distracting thoughts they found helpful in delaying ejaculation. Seventy-four percent of participants provided information as to their specific ejaculation delaying thoughts. Two independent raters reviewed these thoughts, and identified six general themes. Thoughts were grouped into Sex Neutral, Sex Negative, Nonsexual Negative, Sex-Positive, Sex Incongruous, and RE Specific themes. Examples of thoughts making up each theme are presented in Table III. Some participants reported thoughts that fell under more than one theme. The most frequently reported theme was the Sex Neutral theme and was reported by 65% of participants providing information about their distracting thoughts.-There was a considerable degree of variety in the thoughts falling into this theme including thoughts about a snowy mountain, team sports, daily activities such as work or school, as well as singing the national anthem in one's head. The Sex Negative theme was the second most frequently reported theme. These types of thoughts were reported by 39% of participants. The Sex Negative theme involved thoughts including the negative consequences of sex or a perceived negative sexual experience such as getting AIDS or having sex with an unattractive partner. The Nonsexual Negative theme involved thoughts that distracted participants through unpleasant images that were not of a sexual nature such as thinking about war or financial problems. These were reported by 14% of those providing thought content. Distracting thoughts with a Sex Positive theme, reported by 7% of those providing content, were those stressing pleasant aspects of sexual interactions other than ejaculation and included thinking about the partner's pleasure or previous sexual interactions that included prolonged intercourse. Distracting thoughts of the Sexually Incongruous theme, reported by 7% of those providing content, were of images typically found to be antithetical to sexual arousal such as thinking of one's mother or of religious themes. Finally, the RE Specific theme included thoughts specifically or symbolically related to stopping ejaculation such as imagining

Rapid Ejaculation

37

Table III.Thoughts Used to Delay Ejaculation Sex negative Fat, ugly people/partner Ugly monster Ugly teacher Woman with unpleasant personality Unattractive aspect of partner Partner having sex with another man AIDS Pregnancy Stop thinking about intimacy Stop thinking of enjoyment of orgasm

Sex neutral Snowy mountain Books Sing National Anthem in head Count backwards from 100 Day's events Shopping TV show Planning for party School Work Later activities Sex positive Musical rhythm which does not match "We're in no hurry" breathing rhythm "Blank" thoughts How nice it would be to continue Postintercourse events Think of partners' pleasure "Lovemaking" rather than just "having sex" Various physical activities or sports Past prolonged intercourse episodes (exercises, hockey, baseball, salmon Other sex partner(s) fishing, etc.) Nonsexual, negative events Deceased parent Sad event Dead animals on side of road Vietnam war Disliked classes Unpleasant work experiences Solving financial problems Unpaid debts Arguments with partner Parents walking in When will parents/roommate return

Sexually/heterosexually incongruous Mother Calling mother on telephone Grandmother Lord's Prayer Priests and nuns Men RE specific Thinking specifically about stopping ejaculation Dam stopping water flow Try to block pleasurable feelings

a dam stopping a flow of water or specifically thinking about stopping ejaculation and was reported by 5% of participants. Ejaculation Delaying Techniques Participants indicated the techniques they had ever used to prolong intercourse from a list of 28 delaying techniques. The mean number of techniques reported was 14 (SD = 4, range = 0-22). Table IV displays these techniques ordered by average effectiveness of technique as well as the percentage of the sample using the technique. The average number of

38

Grenier and Byers

Table IV. Ejaculation Delaying Techniques as Reported by Entire Sample % of sample who have Correlations (r) Average used Technique effectivenessa technique Controlb Latencyc 4.3 Withdrawing for a few moments 65 4.3 47 Ejaculating before intercourse 12 4.3 Taking drugs before intercourse 4.0 62 Drinking alcohol before intercourse 4.0 02 Applying an anesthetic to penis/glans 3.9 86 Thinking distracting/nonsexual thoughts 3.8 80 Using a different intercourse position 3.7 03 Using more than one condom 3.7 91 Slowing down thrusting 3.7 Making a few thrusts then stopping 81 3.6 82 Alternating intercourse positions 3.6 10 Having a cold shower 04 3.5 Applying ice to penis 3.3 80 Using a condom 3.3 57 Contracting pelvic muscles 3.3 53 Relaxing pelvic muscles 67 3.1 Thrusting more shallowly 3.0 59 Talking during intercourse 3.0 20 Gritting teeth 2.9 66 Using oral sex as a "warm-up" 2.9 51 Thrusting in a circular motion 2.9 35 Taking deep breaths 2.9 16 Having intercourse at a specific time of day 23 2.7 Using more lubrication 2.5 22 Biting lip, tongue, or cheek 2.2 76 Thrusting more deeply 49 1.6 Speeding up thrusting 1.6 61 Thinking erotic/sexy thoughts aEffectiveness of technique in helping to delay ejaculation, ranging from to 5 (very effective). bReported perceived ejaculatory control, N = 109. cReported latency to ejaculate following vaginal penetration, N = 107. dp < .05. ep <. .01.

.01 -.25e .02 -.05 -.01 .05 .24d -.02 -.04 -.01 .18 -.10 .08 -.01 .06 .06 .11 .18 -.10 .05 .36e .08 .09 -.06 -.02 .10 .07 .05

-.06 -.25e -.05 -.12 -.07 .02 .34e -.10 .07 .04 .35e -.07 .03 -.02 .07 .04 .08 .19 -.04 .25e .27e .13 .16 -.15 -.06 .14 .11 .01

1 (very

ineffective)

techniques participants ranked as "very effective" was 3 (SD - 2). The techniques reported to be most effective in prolonging intercourse were withdrawing for a few moments, ejaculating prior to intercourse, and taking drugs before intercourse. The most commonly reported techniques (used by at least 80% of the sample) were slowing down thrusting, thinking distracting/nonsexual thoughts, alternating intercourse positions, making a few thrusts then stopping, using a different intercourse position, and using a condom. In general, the techniques reported most frequently and as most effective

Rapid Ejaculation

39

retrospectively (i.e., from Part E) were also reported to be the most frequently used and most effective when self-monitoring. The zero-order correlations between use of each of the 28 techniques and the two most commonly used RE criteria, Latency and Control, were examined to determine the relationship between the use of a technique and RE. These correlations are presented in Table IV. Only a few techniques were correlated with either criterion. Men who reported greater perceived ejaculatory control were more likely to report using a different intercourse position and thrusting in a circular motion and less likely to report having used ejaculating before intercourse to prolong intercourse. Men with longer reported ejaculatory latencies were more likely to report using a different intercourse position, alternating intercourse positions, using oral sex as a warm-up, and thrusting in a circular motion, and less likely to report ejaculating before intercourse to prolong intercourse. The self-monitoring data were used to examine the relationships between use of each technique and ejaculatory control/latency in a specific interaction. Although 86% of the sample had elected to provide self-monitoring data, only 52% of the sample actually returned self-monitoring forms. A discriminant function analysis comparing men who did and did not provide self-monitoring data revealed no differences on the four RE criteria (Latency, Control, Satisfaction, and Concern), x2(4) = 0.87, p = 0.93. None of the techniques that were significantly related to the retrospective reports were significantly related to self-monitoring reports in the same direction. Using the self-monitoring reports, only one technique, relaxing the pelvic muscles, was significantly positively correlated with both Control and Latency, r = .27, p < 0.05, and r = .39, p < 0.01, respectively. In addition, thinking distracting/nonsexual thoughts was significantly positively correlated and using more lubrication was significantly negatively correlated with Control, r = .31 and -.33, respectively, p < 0.05. Finally, three techniques, ejaculating before intercourse, taking deep breaths, and biting lip, tongue, or cheek were positively correlated with self-monitored Latency, r = .31, p < 0.05, r = .41, p < 0.01, and r = .41, p < 0.01, respectively. Relationships Between Ejaculatory Latency/Control and Ejaculation Delaying Techniques Multivariate analyses were conducted using the retrospective data only, as an insufficient number of participants completed the self-monitoring forms. The relationship between the Ejaculation Characteristics Set (Latency, Control) and the effectiveness of the participants' reported delaying techniques was analyzed using canonical correlation analysis (see

Grenier and Byers

40

Table V. Canonical Correlation Analysis Between Ejaculation Characteristics and Ejaculation Delaying Technique Characteristics (N = 106) Correlation with variate Variable set Ejaculation characteristics set Perceived ejaculatory control Ejaculatory latency Ejaculation delaying technique characteristics set No. of techniques ranked as "very effective" No. of reported techniques Average effectiveness of reported techniques Canonical correlation

.85 .77 .98 .57 .36 .33

Table V). The number of techniques ranked as very effective, the total number of reported techniques, and the average effectiveness of all the techniques reported by a participant were entered as predictor variables and are referred to as the Technique Characteristics Set. One significant pair of variates was found which accounted for 12% of the variance, Rc = .33, F(6, 202) = 2.19, p < 0.05. Both Latency and Control were significantly correlated with the ejaculation characteristics variate. All three predictor variables were correlated with the technique characteristics variate although the number of very effective techniques was most strongly related. Men who reported greater ejaculatory control and longer latencies were more likely to use more delaying techniques, to find those techniques more effective, and to rate more delaying techniques as very effective. A second canonical correlation analysis was conducted to explore the relationship between the Ejaculation Characteristics Set and whether participants reported using each of the ejaculatory delaying techniques, termed the Delaying Technique Set below. Only techniques reported by more than 25% of the sample were retained for analysis (see Table IV). Also, two delaying techniques (i.e., using a different intercourse position and altering intercourse positions) had a bivariate correlation of .80 and consequently altering intercourse positions was eliminated. Consequently, 18 of the original 28 delaying techniques were retained as predictor variables. The canonical correlation comparing the Ejaculation Characteristics Set with the Delaying Techniques set resulted in one significantly pair of variates that accounted for 42% of the variance, Rc = .65, F(36, 174) = 1.91, p < 0.01. Results of this analysis are presented in Table VI. Only variables that had a correlation of .30 or higher with the variate were interpreted (Tabachnick and Fidell, 1989). Ejaculatory control and latency

Rapid Ejaculation

41

Table VI. Canonical Correlation Analysis Between Ejaculation Characteristics and Use of Ejaculation Delaying Techniques (N = 107) Variable set

Correlation with canonical variate

Ejaculation characteristics set Perceived ejaculatory control Reported ejaculatory latency

.74 .87

Ejaculation delaying technique set Using a different intercourse position Thrusting in a circular motion Talking during intercourse Using oral sex as a "warm-up" Ejaculating prior to intercourse

.57 .58 .35 .30

Canonical correlation

-.48

.65

aOnly

techniques that correlated .30 or higher with the variate are reported.

were both strongly related to the ejaculation characteristic variate. Five of the delaying techniques were strongly related to the delaying technique variate. Men with greater reported ejaculatory control and longer reported ejaculatory latencies were more likely to report having used thrusting in a circular motion, using a different intercourse position, talking during intercourse, and oral sex as a warm-up and less likely to report ejaculating before intercourse to prolong intercourse. Erotophobia-Erotophilia and RE

Erotophobia-erotophilia was not significantly correlated with any of the RE criteria. It was, however, significantly correlated with the number of ejaculation delaying techniques ranked as very effective (r = -.33, p < 0.01), and the total number of ejaculation delaying techniques reported (r = -.22, p < 0.05), indicating the men who were erotophilic reported a broader range of ejaculation delaying techniques and found more of those techniques to be effective in helping to delay ejaculation. Sex Therapy Authority Identification of Effective Delaying Strategies

The 15 men who reported the highest degrees of ejaculatory control along with long latencies as well as the 15 men who reported having the poorest control along with short latencies were identified. To determine

42

Grenier and Byers

whether men with good control/long latencies could be distinguished from men with poor control/short latencies, authorities in the area of sexuality (members of the Canadian Sex Research Forum) were asked to examine verbatim transcripts of the methods each of these men described using to delay the occurrence of ejaculation. Twenty-two of the 54 members responded, 15 of whom reported having expertise in the area of sex therapy. These 15 raters correctly sorted 17 of the 30 men (57.7%) which was not statistically greater than chance, Z = .73, p > 0.10. Categorization success by the other 7 professionals was at a similar level. DISCUSSION The Relationship Among Rapid Ejaculation Criteria To determine the extent to which the various criteria used in the literature to identify men with RE are tapping the same phenomenon, RE was measured using four different criteria: perceived ejaculatory control, latency to ejaculate, concern over rapid ejaculation, and satisfaction with ejaculatory control. For the most part, these criteria were significantly related to each other. However, the magnitudes of these correlations were relatively small and accounted for only a small percentage of variance. For example, the two most commonly used components of RE definitions, latency to ejaculate after vaginal penetration and perceived degree of control over ejaculation, shared only 10% of their variance. The small amount of shared variance suggests that these two criteria, as well as the other assessed RE criteria, are relatively independent and represent different aspects of men's sexual and ejaculatory behavior. Further, the data demonstrate that estimates of the prevalence of RE, at least among young men, vary considerably depending upon which criteria are used and if and how those criteria are dichotomized. In the current study, estimates of RE ranged from 0 to 58% of the sample, depending on how RE was operationalized. General estimates reported in the literature range from approximately 30 to 75% of men (Kinsey et al., 1948; McCarthy, 1988, Zilbergeld, 1992). These results indicate that comparing the results of studies in which different criteria are used to operationalize RE may be inappropriate. They also highlight the importance of reporting exactly how RE is operationalized in both clinical and empirical investigations so that meaningful comparisons can be made across RE studies. Further, it may be that a number of different but related phenomena are being grouped together under the RE label and/or that there are a number of different etiologies for RE (Grenier & Byers, 1995; Vandereycken, 1986; Williams,

Rapid Ejaculation

43

1984). Thus, future research should use multiple measures of RE to facilitate identification of the subgroups of RE men. In addition to the measures used in the current study, a measure of men's self-definition as having a sexual problem related to rapid ejaculation should be included. For example, the findings that concern about rapid ejaculation was not related to latency to ejaculate and only shared 7% of the variance with men's perception of their ejaculatory control suggests that men's concern over the rapidity of their ejaculation, and thus perhaps their self-definition as having a sexual problem, is largely independent of the primary criteria used to diagnose RE. Unfortunately, as the current study did not query men (or their partners) specifically about their perceptions of RE as a sexual dysfunction or of their own RE status, the current data cannot be used to determine which dimensions or interactions of dimensions might best characterize a man's subjective identification of RE as a sexual dysfunction. We are currently researching these questions.

Use of Distracting Thoughts to Delay Ejaculation and Prolong Intercourse

The majority of participants reported using distracting thoughts in attempting to prolong intercourse. The men reported a variety of thoughts. Most commonly reported were thoughts that were nonsexual in nature such as thinking about sports, school, or work. Also, a substantial proportion of the men reported using thoughts that were specifically antithetical to sexual arousal or focused upon negative consequences of sexual activity. Only a minority of the men reported using thoughts that had a positive sexual image such as "lasting longer" or concentrating on their partners pleasure. We did not assess whether specific types of cognitions were related to ejaculatory control/latency. However, thinking distracting cognitions, in general, was not related to either ejaculatory control or ejaculatory latency. In addition, it is not known how specific types of thoughts relate to the quality of the overall sexual relationship. These are important topics for future research, given the pervasiveness of the use of distracting thoughts. In particular, it may be that use of distracting thoughts in general may decrease men's ability to attend to their partner and their partner's needs. Further, use of negative thoughts may interfere with men's sexual functioning in other ways, such as by decreasing their sexual satisfaction or sexual enjoyment or by creating a negative association with sexual intercourse (Cranston-Cuebas and Barlow, 1990).

44

Grenier and Byers

Use of Behavioral Techniques to Delay Ejaculation and Prolong Intercourse Eighty-eight percent of participants reported some concern about ejaculating faster than they wanted to. In fact, more than half (59%) reported considerable concern. Further, virtually all the men (99%) reported using at least one behavioral or cognitive technique to prolong intercourse and delay ejaculation, and most men reported using a large number of different techniques. This indicates that young men are quite concerned with the timing of ejaculation. The men identified a number of techniques they considered highly effective at prolonging intercourse. In general, the greater the number of techniques the men reported and the greater the rated effectiveness of the techniques they used, the greater men's perceived ejaculatory control and ejaculatory latency. However, use of only a few of the specific techniques was related to either ejaculatory latency or ejaculatory control and these correlations accounted for only a small percentage of the variance. Further, the techniques related to ejaculatory control/latency were not, in general, the techniques the men had rated as the most effective. That is, the men who reported poor control and/or short latencies were equally likely to report having found these techniques effective. Further, the techniques related to control and/or latency using the retrospective data were not the same techniques that were significantly correlated with these criteria when the men reported on their behavior following a specific sexual encounter. This suggests that, whereas there are behaviors that men can use to delay ejaculation, there may not be specific techniques that work for all men. Rather, the effectiveness of any specific technique at delaying ejaculation may be idiosyncratic to the man or to the particular situation. There may not be specific techniques that work for all or most men in all or most situations. This interpretation is also supported by the fact that a group of sex therapists were unable to discriminate the men with the best ejaculatory control/longest latencies from the men with the poorest ejaculatory control/shortest latencies based on the ejaculatory delaying techniques they reported using. Therefore, it seems that the use of a behavioral delaying technique or any particular pattern of techniques insufficiently accounts for the differences among men in their ability to prolong lovemaking by delaying ejaculation. There are several other possible explanations for these findings. It may be that men with poor control/short latencies do not use these techniques consistently to prolong intercourse. This could account for the discrepancies between the retrospective ("ever used") and self-monitoring (used in a particular sexual situation) data. Alternately, these findings may

Rapid Ejaculation

45

reflect the subjective nature of the effectiveness ratings. That is, the men were asked to evaluate how effective each technique was in prolonging intercourse relative to the length of intercourse when they do not use the technique, and not relative to some absolute standard or to other men's ejaculatory latency or ejaculatory control. Thus, if a technique allowed a man who usually ejaculates 10 sec after penetration to ejaculate 30 sec after penetration, he might rate the technique as effective but use of the technique would not change his latency/control relative to other men. Finally, it may be that the ratings of effectiveness lack validity or lack validity for men with poor control. For example, a man who has good control/longer latency may attribute this to a specific technique when, in fact, some other factor (e.g., little anxiety over sexual performance, greater somatic threshold to stimulation) is actually responsible for his ejaculatory control. Alternately, intercourse of longer duration may afford the opportunity to use a greater variety of techniques. Systematic manipulation of the use of techniques along with corresponding measures of latency and control would have to be undertaken to confidently draw conclusions concerning the effectiveness of any particular potential delaying technique. Multivariate analyses identified only four techniques that were associated with both greater perceived control and longer reported latencies and one technique that was associated with lesser control and shorter latencies. The men with greater perceived control/longer latencies were more likely to report having used different intercourse positions, thrusting in a circular motion, talking during intercourse, and oral sex before intercourse. These techniques can all be applied during lovemaking. On the other hand, men with lesser control/shorter latencies were more likely to use a preemptive technique, that is, ejaculating prior to intercourse to prolong intercourse. Erotophobia-Erotophilia and Ejaculatory Behavior

As predicted, more erotophilic men reported that they had used a greater variety of delaying techniques, and reported that a greater number of the techniques they used were very effective in their past sexual behavior than had more erotophobic men. This supports Fisher et al.'s (1988) contention that people who have more positive feelings about sexuality are more likely to seek out or engage in behaviors that have the potential to enhance their sex lives. It should be noted, however, that the size of correlations between erotophobia/erotophilia and both the number of techniques used and the number of techniques rated as very effective were relatively small and accounted for only about 10% of the variance. This suggests that other factors play a more significant role in the behavior a

46

Grenier and Byers

man engages in to prolong intercourse than does the man's affective evaluation of sexual cues. To the extent that poor ejaculatory control and short ejaculatory latencies is a sexual dysfunction, the results do not support the Fisher et al. prediction that erotophiles are less likely to have or to develop sexual dysfunctions. None of the RE criteria were significantly correlated with erotophilia-erotophobia. Thus, more erotophobic men did not report less ejaculatory control, shorter latencies to ejaculate, more concern over rapid ejaculatory, not less satisfaction with their perceived ability to control the timing of ejaculation than did more erotophilic men. However, the extent to which these results would apply to a clinical population or to a sample of men who are truly erotophobic rather than relatively more erotophobic as in the present sample, is not known. Conclusion

The men in this study were relatively young. Also, the current sample was not drawn from a clinical population. Thus, the extent to which these results can be generalized to older men and to men seeking help from professionals for RE is not known. Future research also needs to establish the reliability and validity of men's reports, perhaps by including reports by a partner about the partner's perception of the man's intercourse/ejaculatory behavior as well as the woman's own role and sexual satisfaction. Additional research is needed on the relationship between use of specific techniques and ejaculatory control. In particular, the relationship between partner behavior and men's ejaculatory control needs to be investigated. Nonetheless, these results demonstrate that RE is a multifaceted phenomenon and document the need for multivariate approaches to the assessment of RE. Further, they give some support to the notion that men's behavior in sexual situations is related to their ejaculatory control. ACKNOWLEDGMENT

The authors thank Kelli-an Lawrance, Lucia O'Sullivan, and Lisa Price for their feedback on the questionnaire, and Jennifer Meen for her assistance in coding the data. REFERENCES Chesney, A. P., Blakeney, P. E., Cole, C. M., and Chan, F. A. (1981). A comparison of couples who have sought sex therapy with couples who have not. /. Sex Marital Ther. 7:131-140.

Rapid Ejaculation

47

Cooper, A., and Magnus, R. (1984). A clinical trial of the beta blocker propranolol in premature ejaculation. J. Psychosom. Res. 28: 331-336. Cranston-Cuebas, M. A., and Barlow, D. H. (1990). Cognitive and affective contributions to sexual functioning. Ann. Rev. Sex Res. 1: 119-161. Fisher, W. A., Byrne, D., White, L. A., and Kelley, K. (1988). Brotophobia-Brotophilia as a dimension of personality. J. Sex Res. 25: 123-151. Grenier, G., and Byers, E. S. (1995). Rapid ejaculation: A review of conceptual, etiological, and treatment issues. Arch. Sex. Behav. 24: 447-472. Hunt, M. (1974). Sexual Behavior in the 1970's, Playboy Press, Chicago. Kaplan, H. S. (1989). PE: Haw to Overcome Premature Ejaculation, Brunner/Mazel, New York. Kinsey, A. C., Pomeroy, W. B., and Martin, C. E. (1948). Sexual Behavior in the Human Male, W. B. Saunders, Philadelphia. McCarthy, B. (1988). Cognitive-behavioral strategies and techniques in the treatment of early ejaculation. In Leiblum, S., and Rosen, R. (eds.), Principles and Practices of Sex Therapy: Update for the 1990's, Guilford, New York, pp. 141-167. Nathan, S. (1986). The epidemiology of the DSM-III psychosexual dysfunctions. J. Sex Marital Ther. 12: 267-281. Reading, A., and Wiest, W. (1984). An analysis of self-reported sexual behavior in a sample of normal males. Arch. Sex. Behav. 13: 69-83. Schover, L. R., Friedman, J. M., Weiler, S. J., Heiman, J. R., and LoPiccolo, J. (1982). Multiaxial problem-oriented system for sexual dysfunctions. Arch. Gen. Psychiat. 39: 614-619. Segraves, R. T., Saran, A., Segraves, K., and Maguire, E. (1993). Clomipramine versus placebo in the treatment of premature ejaculation: A pilot study. J. Sex Marital Ther. 19: 198-200. Semans, J. H. (1956). Premature ejaculation: A new approach. South. Med. J. 49: 353-358. Spector, I. P., and Carey, M. P. (1990). Incidence and prevalence of the sexual dysfunctions: A critical review of the empirical literature. Arch. Sex. Behav. 19: 389-408. St. Lawrence, J. S., and Madakasira, S. (1992). Evaluation and treatment of premature ejaculation: A critical review. Int. J. Psychiat. Med. 22: 77-92. Tabachnick, B., and Fidell, L. (1989). Using Multivariate Statistics, 2nd ed. Harper and Row, New York. Vandereycken, W. (1986). Towards a better delineation of ejaculatory disorders. Acta Psychiat. Belg. 86: 57-63. Williams, W. (1984). Secondary premature ejaculation. Aust. N.Z. J. Psychiat. 18: 333-340. Zilbergeld, B. (1992). The New Male Sexuality, Bantam, Toronto.

Archives of Sexual Behavior. Vol. 26, No. 1, 1997

Yohimbine, Erectile Capacity, and Sexual Response in Men David L. Rowland, Ph.D.,1,2,3 Khalid Kalian, M.D.,2 and A. Koos Slob, Ph.D.2

In a double-blind, placebo-controlled crossover study on a group of men with erectile problems and a sexually functional comparison group, the effect of yohimbine (up to 30 mg/day) was assessed on a number of objective and subjective measures of erectile response through the use of daily logs and psychophysiological laboratory procedures involving response to visual sexual stimulation (VSS). Sexual desire, arousal, and ejaculatory response were also assessed. Results indicated no effect of yohimbine on most aspects of sexual response in sexually functional men. Mixed effects were found on measures of sexual function in men with erectile dysfunction, with 3 of 11 men reporting strong positive effects. Under yohimbine, frequency of sexual activities increased, as did self-assessed genital response to VSS. Daily diaries indicated increased sexual arousal and erectile response during masturbation but not intercourse. A number of other measures, including NPT and retrospective summaries of erectile functioning at the end of drug phases, showed no effect. Results are discussed in terms of possible yohimbic effects on psychological factors that modulate overall sexual response, and consequently, erectile response. KEY WORDS: Yohimbine; men; erection; sexual dysfunction; sex; pharmacology.

This project was supported in part by grants from the Society Trust Fund, Erasmus University Rotterdam; the Foundation for Urological Scientific Research, SUWO, Professor F.H. Schroder, MD, Urologist; and the O.P. Kretzmann Fund, Valparaiso University. 1Department of Psychology, Valparaiso University, Valparaiso, Indiana 46383. Department of Endocrinology and Reproduction, Faculty of Medicine and Health Sciences, Erasmus University, 3000 DR Rotterdam, The Netherlands. 3To whom correspondence should be addressed. 49 0004-0002/97/0200-0049$12.50/0 C 1997 Plenum Publishing Corporation

so

Rowland, Kallan, and Slob

INTRODUCTION

Over the past 10 years, a multitude of studies have attempted to determine the efficacy of the putative aphrodisiac yohimbine (Mann, et al., 1996; Morales et al., 1987; Nessel, 1994; Riley, 1989; Riley et al., 1993; Sonda et al., 1990; Susset et al, 1989). Although the mechanism of action of this ot2-adrenergic blocker on sexual function is unclear (see Grossman et al., 1993; Riley, 1994), it has been used with varying degrees of success in the treatment of both psychogenic and organic impotence. In general, these studies typically conclude partial or complete efficacy in as high as one third of treated patients. Some of the variation in outcome of these studies may be related to differences in methodologies (e.g., rigor of study design, dose), in the type of subjects tested (unknown etiology, somatic, psychogenic), and in the method of assessment of sexual functioning. Among the issues included in this last point is the heavy reliance upon self-assessment by the subjects; few studies have provided objective, quantitative evidence that yohimbine increases erectile capacity within a sexual situation. In the following study we have attempted to determine the efficacy of yohimbine under rigorous testing conditions by using a prospective double-blind, placebo-controlled crossover design. A group of volunteer, sexually functional men was also included for purposes of comparison. Unlike many previous studies, which have relied upon simple global self-assessment of sexual function, the present investigation utilized a battery of tests to assess various components of the sexual response cycle: sexual desire, sexual and erectile arousal, and ejaculation. Since erectile response was the primary dependent measure of interest, this variable was further measured using multiple indices, including self-assessment of erectile response through daily logs and overall assessment at the end of each drug phase, waking erectile assessment (WEA) of response to erotic visual stimulation (Janssen et al., 1994), and nocturnal penile tumescence (NPT) measurements involving maximum penile circumference and rigidity. METHOD Subjects

The experimental (i.e., sexually dysfunctional) group consisted of 11 patients with erectile dysfunction recruited from the urology outpatient clinic of an academic hospital. For inclusion, patients had to meet the following criteria: (i) age 18 years or older (X = 48.6), (ii) erectile dysfunction

Yohimbine and Sexual Response in Men

51

(ED) defined through self-reported failure to obtain an erection sufficient for vaginal intromission at least 50% of the time without ejaculation, or loss of erection following intromission without ejaculation at least 50% of the time, (iii) duration of ED for at least 6 months (Mdn duration = 3.8 years), (iv) willingness to attempt intercourse or masturbation at least once per week, and (v) consent of sexual partner when appropriate. Prior to participation, all patients underwent extensive evaluation by medical specialists, including physical examination with special attention to vascular and neural function, serum testosterone levels, sexual and medical history, psychiatric evaluation, and blood analysis for metabolic disturbances. On all measures, patients were within the normal range. Based upon extensive medical evaluation (which when called for included such procedures as Doppler investigation of the cavernous arteries, papaverine injection, cavernosography, and endocrinological analysis) and clinical interview, and waking erectile assessment (WEA), erection problems in this group were judged primarily psychogenic. However, while major somatic (vascular and neural) etiologies were eliminated, some men with ED (7 of 11) had a concomitant disease, medical condition, or use of medication that might have placed them at minor or occasional risk for a somatic contribution to the dysfunction. Specifically, 4 men reported high blood pressure, 1 of whom used a beta-blocking agent; 2 reported chronic respiratory problems with occasional use of bronchodilators; and 1 reported controlled diabetes. Four men reported no medical problems. A group of healthy, sexually functional volunteers (n = 15), recruited through announcement to regional university and health profession communities, was also included. These subjects indicated no problems with sexual functioning, particularly erectile response, were at least 18 years old (X = 39.9), were willing to attempt intercourse or masturbation at least once per week, and had the consent of their sexual partner when appropriate. Experimental Design and Treatment

This prospective study consisted of a double-blind, placebo-controlled, crossover design with inclusion of a sexually functional comparison group. Prior to implementation of the study, a pretest session provided an initial evaluation and collection of data, and gave subjects the opportunity to adapt to the laboratory setting and procedures. Drug phases consisted of 4-week periods of placebo or yohimbine treatment. After the first 4-week phase, subjects underwent an additional 4-week crossover treatment, with placebo for those subjects previously receiving yohimbine, and yohimbine

52

Rowland, Kalian, and Slob

for those receiving placebo. Tablets of yohimbine hydrochloride and placebo were manufactured identical in appearance and the order of treatment (drug vs. placebo) for the subjects was determined by the Pharmacy Department of the University Hospital. Yohimbine treatment began at a dose of 3x5 mg/day for the first 2-weeks; for the final 2 weeks, the dose was increased to 3 x 10 mg/day. Subjects on placebo were given identical instructions. Following completion of the study, the unblinded code revealed that 6 patients began with yohimbine, and 5 with placebo; for control subjects, 8 began with yohimbine, 7 with placebo. Assessment of Sexual Function During the pretest and at the end of both drug phases, a series of self-assessments and objective measures of sexual functioning were undertaken. Self-assessment, based upon an extensive structured interview with a clinician, relied on an inventory of sexual function consisting of 56 items which dealt with quantitative and qualitative aspects of the sexual relationship and sexual activities, and the occurrence of sexual problems (Rowland and Slob, 1992; Rowland et al., 1994; test-retest reliability for this instrument is approximately .75). As part of this questionnaire, the various dimensions of sexual desire, arousal (including erectile function), and ejaculatory control were evaluated on 7-point scales. Objective assessment of erectile response was achieved using both NPT and WEA procedures. Specifically, NPT data were collected during 4 consecutive nights sleeping with erectiometers (test-retest reliability for NPT = .82): 2 nights with a yellow type (requiring approximately 250-g expansion force) and 2 nights with a green type (approximately 450-g expansion force) (see Rowland and Slob, 1992, and Slob et al., 1990, for descriptions of erectiometer response characteristics). WEA to visual sexual stimulation (VSS) was determined using (i) penile measures: Barlow strain gauge (Barlow et al., 1970) placed toward the tip of the penis to determine penile circumference, and a yellow erectiometer around the base to provide a measure of change in penile circumference and rigidity; and (ii) subjective assessment of arousal and genital response. During the WEA procedure, subjects were tested privately for erectile response to VSS. Penile devices were positioned by the experimenter, then the lap of the subject was covered to prevent visual genital feedback. Three different videos, each approximately 9 min long, were used for each of the different sessions (pretest and two drug phases). Because the two videos used during treatment periods were always presented in the same order, they were counterbalanced across

Yohimbine and Sexual Response in Men

53

drag and placebo conditions. All three videos depicted a man and a woman having sex in various ways: cunnilingus, fellatio, masturbation, female orgasm; and all ended with intravaginal ejaculation. Following the video, the subject completed a number of items assessing sexual and genital response during the stimulus period. During each drug phase, subjects also completed daily logs monitoring the quality and quantity of various aspects of sexual activity, including sexual fantasizing, arousal, and ejaculation. Subjects mailed these daily diaries at the end of each week to the investigators in prestamped envelopes. To maximize drug effects and minimize carryover effects, daily log data collected from only the last 2 weeks of each treatment period were used for analysis. Pharmacokinetic studies on yohimbine indicate drug inactivation and excretion within days or hours of administration (Owen et al., 1987; Hedner et al., 1992). General Procedure

Subjects initially met with the investigators following either volunteer or referral by the urology clinic, and after full discussion of the aims and procedures of the studies as approved by the Medical Ethics Committee gave written informed consent. At this time, subjects were scheduled for the pretest session and given explicit written instructions for obtaining NPT measurements. During the pretest session, all measures indicated previously were taken on the subject. Subjects were then instructed in the medication procedure and the use of the daily logs, and were provided with contact numbers in case questions arose about the procedure or the medication, side effects were experienced, or the desire to adjust the dose downward or to discontinue treatment occurred. At this time, subjects scheduled appointments for laboratory assessments, scheduled for the final week of each treatment phase so as to maximize drug effects and minimize carryover effects. Approximately 1 week prior to these appointments, subjects were mailed an NPT kit (containing yellow and green erectiometers, instructions, and data sheets) and instructed to bring the results of NPT testing collected on the 4 days immediately prior to their appointment with them to the laboratory. In addition to all the measures collected at the pretest, several additional measures were taken during drug phase tests. Blood pressure was measured as a precautionary step (Grossman et al., 1993). Subjects were also interviewed to obtain subjective impressions of their sexual experiences over the preceding weeks, to ascertain the accuracy and completeness of

54

Rowland, Kalian, and Slob Table I. Inventory of Sexual Function Prior to and During Treatment Phasesa Pretest

X

Yohimbine

SE

X

Placebo

SE

X

SE

14.9 13.0 13.1

2.1 1.1 1.3 1.4 1.9

Functional group (n = 15) Frequency per month No. night erections No. intercourse No. masturbation No. erections w. sex No. ejaculations

14.0* 5.1* 11.1* 14.9* 14.9*

2.2 0.7 2.7 2.4 2.8

13.3 12.5

2.1 1.1 1.4 1.8 1.8

5.3b

4.1

0.7 0.4

4.7 4.3

0.4 0.4

5.0 4.3

0.3 0.3

Erectile function Able to get erection Able to keep erection

6.7* 6.5*

0.2 0.1

6.9 6.5

0.1 0.2

6.4 6.5

0.1 0.1

Ejaculatory function Latency to ejac. (min) Ejac. when desired

7-10 6.2*

0.3

0.4

4-6 6.2

0.3

Sexual interest Interest in sex Sexual fantasies

17.5

5.3 8.5

7-10

6.8

4.5 7.3

ED group (n = 11) Frequency per month No. night erections No. intercourse No. masturbation No. erections w. sex No. ejaculations

8.1 3.2 3.3 3.1 3.2

1.9 1.0 1.5 0.7 1.1

8.4 3.4 3.4 5.1 6.0

3.2 1.1 1.7 1.4 1.6

9.1 2.4 3.2 2.6 3.7

3.2 1.3 1.3 0.7 1.1

Sexual interest Interest in sex Sexual fantasies

4.0 3.0

0.7 0.7

4.4 3.5

0.6 0.6

3.9 2.9

0.5 0.4

Erectile function Able to get erection Able to keep erection

2.2 2.4

0.6 0.5

2.6 2.2

0.5 0.5

2.4 2.5

0.6 0.5

Ejaculatory function Latency to ejac. (min) Ejac. when desired

7-10

4.1

7-10

11-15

0.6

6.2

1.1

4.6

0.7

aFor

items other than frequencies, 1 = never, none, or not at all; 7 = often, always, or very high. bIndicates significant difference (p < 0.05) from ED group during pretest only. Significant effects during treatment are reported in the text.

their daily logs and the inventory of sexual function, and to elaborate on items from the daily logs and inventory of sexual function. They were also queried about side effects and whether they and/or their partner had any

Yohimbine and Sexual Response in Men

55

notion regarding the particular treatment, yohimbine or placebo, that they were receiving at that time.

RESULTS Differences Prior to Treatment

Table I indicates differences between dysfunctional and functional groups on self-report measures of sexual functioning prior to treatment. As expected, frequency of sexual activities and erectile function were lower in dysfunctional subjects than in the comparison group, but there were no differences in measures of sexual desire. In addition, NPT was significantly lower in the ED group during pretest (15.5 vs. 24.5 mm), F(l, 23) = 6.74, power = .68, p = 0.016. WEA procedures indicated that in response to VSS penile response (as measured by the erectiometer, 7.7 vs. 15.7 mm), F(l, 25) = 3.68, power = .45, p = 0.066, and self-assessed genital response (penile stiffness, 28.3 vs. 64.0 mm on a 100-mm analog scale), F(1, 25) = 9.75, power = .85, p = 0.004, were lower in dysfunctional men than in the functional comparison group. Analysis of Drug, Group, and Drug x Group Effects

A 2 (Drug: Yohimbine vs. placebo) x 2 (Group: dysfunction vs. comparison) factorial ANOVA was used to determine effects for each of the measures used to assess sexual function: daily logs, inventory of sexual function, NPT, and WEA. Daily Logs (Table II). Overall level of sexual interest and activity (a composite score representing the sum of masturbation, intercourse, and sexual fantasizing) was not affected by yohimbine, F(l, 21) = 0.71, power = .09, p = 0.41, although there was a near-significant Drug x Group interaction, F(l, 21) = 3.97, power = .46, p = 0.06. Specifically, an increase in level of activity occurred in the ED group, whereas no change was seen in the comparison group; in posthoc analysis, however, the increase in the ED group did not reach significance (p > 0.20). Sexual arousal and erections were reported significantly stronger under yohimbine for masturbation, Drug: F(l, 16) > 7.25, power .71, p > 0.016, with nearly all of this effect occurring in the dysfunction group, Drug x Group: F(l, 16) > 5.33, power > .58, p < 0.03; rcocmiox: p = 0.16 and 0.10, respectively. There was no significant effect of yohimbine on these measures for intercourse, F(l, 17) < .79, power <, .15, p > 0.39.

56

Rowland, Kallan, and Slob Table II. Summary of Daily Diariesa

Placebo

Yohimbineb

x

SE

x

SE

Functional group (n = 15) Intercourse (No. per week) Sexual arousal

Firmness of erection Masturbation (No. per week) Sexual arousal Firmness of erection Sexual fantasies (No. per week)

3.8 7.8 9.1

0.9 0.2 0.4

3.5 8.0 9.2

0.9 0.3 0.3

3.5 7.7 9.2 8.1

0.6 0.3 0.3 1.1

3.5 7.6 9.1 9.2

0.7 0.3 0.3 1.2

ED group (n = 11) Intercourse (No. per week) Sexual arousal Firmness of erection

4.0 7.1 6.4

1.3 0.6 1.0

2.7 7.0 6.0

1.0 0.6 1.1

Masturbation (No. per week) Sexual arousal Firmness of erection

2.8 5.8 7.0

1.1 1.4 0.6

2.4 4.5 5.3

0.8 1.1 0.8

Sexual fantasies (No. per week)

4.8

1.3

4.0

0.8

aFor

items other than frequencies, values represent ratings on a 10-point scale, where 1 — little or none, 10 = very intense or high. bSignificant effects are reported in the text.

Inventory of Sexual Functioning. This retrospective inventory, administered at the end of each drug treatment period, indicated overall group differences in sexual function (Table I). In a pattern somewhat consistent with daily log data, significant yohimbine effects were seen on overall level of sexual activity (a composite score representing the sum of masturbation, intercourse, and erections during sex), F(l, 23) = 4.06, power = .48, p = 0.056, with a larger proportionate increase occurring in the ED group (42% increase in ED group, 14% increase in functional comparison group). Significant drug effects were also seen on ejaculation, with "greater ability to ejaculate when desired," F(l, 23) = 9.71; power = .85; p = 0.005, and increased latency to ejaculation, F(l, 23) = 8.32; power = .79; p = 0.008, under yohimbine treatment. No significant drug or interaction effects were found on retrospective measures of erectile response. Nocturnal Penile Tumescence. Although there were near group differences in NPT, F(l, 24) = 3.05, power = .46, p = 0.09, neither drug nor interaction effects were found on this measure, F(l, 24) < .31, power < .15, p > 0.59 (Table III).

Yohimbine and Sexual Response in Men

57

Table III. WEA and NPT During Treatment Phasesa Group Functional Yohimbine

x Waking Erectile Assessment Penile response (mm change from baseline) 18.5 Erectiometer 30.0 Barlow gauge Subjective response Strongest arousalb Strongest penile feelingb Penile stiffnessc Ease of getting erectionb Ease of keeping erectionb NPT (mm change)

SE

Dysfunctional (ED)

Placebo

x

SE

2.8 21.0 2.6 4.5 28.1 4.0

0.3 4.9 0.2 0.4 5.3 0.4 62.3 8.6 68.3 8.0 6.1 0.3 6.2 0.3 5.1 0.4 5.0 0.5 4.6 5.3

23.5

6.4

22.3 6.1

Yohimbine

X

SE

10.1 3.6 19.6 3.8 4.1 4.0 38.4 3.8 2.9

Placebo

0.5 0.5 7.4 0.5 0.5

16.3 5.3

X

SE

6.6

2.8 3.9

17.7

3.8 3.1

0.5 0.4 28.0 7.0 2.6 0.5 2.5 0.5 16.6

5.4

aSignificant

effects are reported in the text. items, 1 = none or not at all; 7 = very much or intense. cMeasured on 100 mm analog scale where 0 = no erection, 100 = full erection. bFor these

Waking Erectile Assessment. As expected, group differences were evident on WEA, F(l, 24) = 5.65, power = .87, p = 0.026 (Table HI). No overall drug effect occurred (p = 0.60), but a near significant interaction effect was detected on penile circumference/rigidity as measured by the erectiometer, F(1,24) = 2.6, power = .32, p = 0.06, one-tailed. Specifically, dysfunctional men showed higher penile response to VSS under yohimbine compared to placebo, whereas controls showed lower penile response, but these differences between yohimbine and placebo conditions did not reach significance for either group in post hoc analyses (p > 0.10). Identical (though much stronger) patterns were found on self-reported measures of arousal and genital response to VSS: Based on interaction effects and subsequent post hoc analysis, under yohimbine dysfunctional men found it easier to get an erection, F(1, 24) = 10.52, power = .88, p = 0.003; post hoc: p = 0.024, reported stronger feelings in their penis, F(l, 24) = 4.86, power = .56, p = 0.037; post hoc: p = 0.053, and rated their erections (penile stiffness) as somewhat stronger, F(l, 24) = 5.66, power = .56, p = 0.026; post hoc: p = 0.113. In contrast, controls showed no differences on these measures across the two treatment phases (p > 0.35 for all comparisons).

58

Rowland, Kalian, and Slob

Global Assessment Based upon blind global investigator evaluations of combined objective and subjective measures, the effect of yohimbine on erectile function was mixed. Of the 11 patients studied in our project, 3 showed an overall strong effect, 5 showed a partial effect (3 moderate, 2 weak), and 3 showed no effect. In contrast, 1 subject showed very weak improvement during the placebo phase, x2(l) = 2.93, p = 0.086. Other Effects At the doses used in this study, yohimbine did not significantly increase either systolic or diastolic blood pressure, F(1, 16) < 2.22, power < .30, p > 0.16. However, a number of side effects were reported, and during the final 2 weeks of yohimbine, 1 control and 4 dysfunctional subjects used doses lower than the maximum (30 mg/day). Among functional subjects, the most prevalent side effects included disturbed sleep (9) and lower sexual desire (4). Significant side effects were noted by 2 dysfunctional subjects, and these were minimal (slight diarrhea [1], frequent urination, and lack of energy [1]). Of the 15 subjects in the comparison group, 10 correctly discriminated yohimbine from placebo treatment. Two were incorrect in their attribution, and 3 did not know. Of men with ED, 3 of 11 correctly discriminated yohimbine from placebo, 6 erred in their discrimination, and 3 did not know. The difference in correct identification of treatment phases between controls and dysfunctional men was significant, x2(1) = 6.4, p = 0.02.

DISCUSSION The present results agree with and elaborate upon the findings of other studies; furthermore, they may help specify conditions under which yohimbine is likely to impact sexual response positively. Yohimbine had no enhancing effect on sexual desire, arousal, or erectile capacity in sexually functional men, a finding consistent with previous reports (Charney and Heninger, 1986; Danjou et al., 1988). Indeed, if the open comments of men in this group are to be believed, the side effects associated with yohimbine may have slightly decreased interest in sex in functional men. A mild enhancing effect on ejaculatory control was reported by this group, with ejaculation delayed slightly under yohimbine.

Yohimbine and Sexual Response in Men

59

In men with erectile problems, yohimbine also slightly postponed ejaculation, resulting in improved self-reported ejaculatory control. In addition, yohimbine may have exerted an enhancing effect on sexual desire, so suggested by the increased sexual activity exhibited particularly by the dysfunctional group under yohimbine. Although congruent with yohimbine effects previously reported on sexual motivation in both humans (Sonda et al., 1990) and nonhumans (Clark et al., 1984), the increase in our study may simply reflect the result of improved and/or successful sexual performance, an outcome that is likely to induce further interest in sexual activities. Based upon global evaluations, the effect of yohimbine on erectile response in ED men was mixed, with 3 of 11 patients indicating a strong effect, and 5 others reporting partial effect. This success rate parallels that of other studies which indicate anywhere from 20% or higher rates of success in achieving good erections (Morales et al., 1987; Riley, 1989; Sonda et al., 1990; Susset et al., 1989). Indeed, given the limited sample utilized in the present analysis, it is likely that some yohimbine effects went undetected. In the present analysis, the effect of yohimbine on erection appeared to depend on a number of factors, including the particular method used to assess the sexual response and specific type of sexual activity that was being assessed. For example, based on WEA tests, self-reported genital and arousal responses (and to a lesser extent, objectively measured erection) were enhanced to sexual stimulation under yohimbine treatment, suggesting a moderate though real effect on sexual response in men with erectile problems. In addition, daily log responses from our study indicated augmented erectile response for masturbation but not intercourse under yohimbine. These findings suggest that a positive response to yohimbine may be more probable when the sexual partner is not present Perhaps the presence of the sexual partner introduces significant psychological/relationship factors which in turn mitigate the arousal-enhancing properties of yohimbine. Such dyadic factors may, for example, induce performance-related anxiety which inhibits erectile response in dysfunctional men (Kaplan, 1981; Rowland and Heiman, 1991). Although the implication remains to be tested empirically, treatment of such relationship factors through counseling might significantly enhance the effect of yohimbine. In this respect, yohimbine might best be viewed not as a replacement to sex therapy (e.g., Montorsi et al., 1994) but rather as a supplement to it. An alternative explanation to the lack of self-reported improvement of erectile response during sexual activity with a partner may stem from different criteria applied to various sexual situations by the patients. The criterion for improvement with a sex partner may hinge upon attainment of sufficient erection for intromission. In contrast, sex without a partner requires no erectile threshold for successful performance, and as such the

60

Rowland, Kallan, and Slob

criterion for improvement may be tied more directly to incremental increases in penile circumference or rigidity, as occurred under WEA procedures. Nevertheless, given the modest, yohimbine-induced increase in erectile response demonstrated under laboratory WEA conditions, yohimbine might be most effective for men who are already capable of attaining at least partial erection. Indeed, we found moderate (though not quite statistically significant) correlations between global assessments of yohimbine efficacy (1 = no effect, 4 = strong effect) and pretest measures of self-reported penile response, r(10) = .48; p = 0.06, one-tailed, and actual erectile response (using erectiometer), r(10) = .43; p = 0.09, one-tailed, during WEA. This pattern is consistent with Susset et al.'s (1989) finding that yohimbine is most effective in men with milder degrees of impotence. Global assessment of yohimbine efficacy and the patient's own ratings of the severity of the erection problem during the pretest interview showed low correlations (r < .21), not surprising given that initial (i.e., prior to evaluation) anamnestic reports of sexual function often suffer from lack of reliability or accuracy (see Rowland and Slob, 1995). The lower NPT scores for the ED group may indicate a possible somatic contribution to the etiology of the erectile dysfunction. However, consistent with a number of previous reports (Condra et al., 1986; Mann et al., 1996; Riley, 1989), NPT measurements were unaffected by yohimbine in the present study. Previous theorizing posits a selective effect of yohimbine on stimulated (vs. nocturnal) erections (Riley, 1989). Why NPT should be unaffected by yohimbine is not clear, but it may be related to the special capacity of this agent to improve erectile problems that are psychogenic (Mann et al., 1996; Riley, 1994; Segraves, 1991). For example, through its action on the CNS, yohimbine may modulate psychological processes involved in sexual response, e.g., motivational (desire and arousal), affective, and/or cognitive components, and therefore may enable erectile response to reach its potential physiological maximum. Indeed, there is ample evidence to suggest yohimbic action on these psychological processes: affective systems (e.g., Charney et al., 1983), motivational and desire systems (Clark et al., 1984; Sonda et al., 1990), and subjective arousal systems (Price et al., 1984). These psychological factors, combined with greater ejaculatory control, may well enhance the overall sexual experience and thus maximize erectile response. To the extent that most somatic impotence includes a significant psychogenic component (LoPiccolo and Stock, 1986), patients categorized as such may show substantial improvement as well. In contrast, yohimbine may exert little or no direct action on penile physiology or on actual erectile capacity. And since NPT is minimally affected by psychological factors, this measure would likewise be unaffected by yohimbine treatment. Verification of such suppositions requires greater attention to

Yohimbine and Sexual Response in Men

61

the effects of yohimbine on the motivational, affective, and cognitive components of sexual response, an avenue of investigation that has largely been ignored. Finally, unlike their sexually functional counterparts, many men with ED had concomitant physical disorders and/or were taking other medications during the course of the experiment. Such disorders and/or medications appear to produce side effects sufficient to mask those of the yohimbine and thus to interfere with subjects' ability to discriminate between yohimbine and placebo treatment phases. So, while blind treatment conditions were effective for the dysfunctional group, they were not valid for functional subjects, most of whom correctly assumed which treatment they were undergoing. Future studies claiming blind treatment conditions may be well advised to validate this control procedure within the experiment, as without it the expectational factors of treatment are no longer eliminated.

ACKNOWLEDGMENT

The authors gratefully acknowledge the contribution of P. L. A. Smeenk, Pharmacist of GENFARMA, The Netherlands, who provided the yohimbine hydrochloride and placebo tablets. REFERENCES Barlow, D. H., Becker, R., Leitenberg, H., and Agras, W. S. (1970). A mechanical strain gauge for measuring penile circumference change. J. Appl. Behav. Anal. 3: 73-76. Charney, D. S., and Heninger, G. R. (1986). a-Adrenergic and opiate receptor blockade: Synergistic effects on anxiety in healthy subjects. Arch. Gen. Psychiat. 43: 1037-1041. Charney, D. S., Heninger, G. R., and Redmond, D. E. (1983). Yohimbine induced anxiety and increased noradrenergic function in humans: Effects of diazepam and clonidine. Life Sci. 33: 19-29. Clark, J.T., Smith, E. R., and Davidson, J. M. (1984). Enhancement of sexual motivation in rats by yohimbine. Science 225: 847-849. Condra, M., Morales, A., Surridge, D. H., Owen, J. A., Marshall, P., and Fenemore, J. (1986). The unreliability of nocturnal penile tumescence recording as an outcome measurement in the treatment of organic impotence. J. Urol 135: 280-282. Danjou, P., Alexander, Warot, D., Lacomblez, L., and Puech, A. (1988). Assessment of erectogenic properties of apomorphine and yohimbine in man. Br. J. Clin. Pharmacol. 26: 733-739. Grossman, E., Rosenthal, T., Peleg, E., Holmes, C., and Goldstein, D. S. (1993). Oral yohimbine increases blood pressure and sympathetic nervous outflow in hypertensive patients. J. Cardiovasc. Pharmacol. 22: 22-26. Hedner, T., Edgar, B., Edvinsson, L., Hedner, J., Persson, B., and Pettersson, A. (1992). Yohimbine pharmacokinetics and interaction with the sympathetic nervous system in normal volunteers. Eur. J. Clin. Pharmacol. 43: 651-656.

62

Rowland, Kalian, and Slob

Janssen, E., Everaerd, W., van Lunsen, R., and Oerlemans, S. (1994). Validation of psychophysiological waking erectile assessment (WEA) for the diagnosis of male erectile disorder. J. Consult. Clin. Psychol. 1222-1228. LoPiccolo, J., and Stock, W. (1986). Treatment of sexual dysfunction. J. Consult. Clin. Psychol. 54: 158-167. Kaplan, H. S. (1981). The New Sex Therapy, Bruner/Mazel, New York. Mann, K., Klingler, T., Noe, S, Roeschke, J., Mueller, S., and Benkert, O. (1996). Effects of yohimbine on sexual experiences and nocturnal tumescence and rigidity in erectile dysfunction. Arch. Sex. Behav. 25: 1-16. Montorsi, F., Strambi, L., Guazzoni, G., Galli, L., Barbieri, L., Rigatti, P., Pizzini, G., and Miani, A. (1994). Effect of yohimbine-trazodone on psychogenic impotence: A randomized, double-blind, placebo-controlled study. Urology 44: 732-736. Morales, A., Condra, M., Owen, J., Surridge, D., Fenemore, J., and Harris, C. (1987). Is yohimbine effective in the treatment of organic impotence? Results of a controlled trial. J. Urol 137: 1168-1172. Nessel, M. A. (1994). Yohimbine and pentoxiifylline in the treatment of erectile dysfunction. Am. J. Psychiat. 151: 453. Owen, J., Nakatsu, S., Fenemore, J., Condra, M., Surridge, D., and Morales, A. (1987). The pharmacokinetics of yohimbine in man. Eur. J. Clin. Phamacol. 32: 577-582. Price, L. H., Chantey, D. S., and Heninger, G. R. (1984). Three cases of manic symptoms following yohimbine administration. Am. J. Psychiat. 141: 1267-1268. Riley, A. J. (1994). Yohimbine in the treatment of erectile disorder. Br. J. Clin. Phamacol. 38: 133-136. Riley, A. J. (1989). Double blind trial of yohimbine hydrochloride in the treatment of erection inadequacy. Sex. Marital. Ther. 4: 17-26. Riley, A. J., Feet, M., and Wilson, C. (eds.). (1993). Sexual Pharmacology, Oxford, Claredon. Rowland, D. L., and Heiman, J. R. (1991). Self-reported and genital arousal changes in sexually dysfunctional men following a sex therapy program. J. Psychosom. Res. 35:1-11. Rowland, D. L., and Slob, A. K. (1992). Vibrotactile stimulation enhances sexual response in sexually functional men: A study using concomitant measures of erection. Arch.Sex. Behav. 21: 387-400. Rowland, D. L., and Slob, A. K. (1995). Understanding and diagnosing sexual dysfunction: recent progress through psychophysiological and psychophysical methods. Neurosci. Biobehav. Rev. 19: 201-209. Rowland, D. L., den Ouden, A. H., and Slob, A. K. (1994). The use of vibrotactile stimulation for determining sexual potency in the laboratory in men with erectile problems: Methodological considerations. Int. J. Impotence Res. 6: 153-160. Segraves, R. T. (1991). Pharmacological enhancement of human sexual behavior. J. Sex Educ. Ther. 17: 283-289. Slob, A. K., Blom, J. H., & van der Werff ten Bosch, J. J. (1990). Erection problems in medical practice: Differential diagnosis with a relatively simple method.J. Urol. 143:46-50. Sonda, L. P., Mazo, R., and Chancellor, M. B. (1990). The role of yohimbine for the treatment of erectile impotence. J. Sex Marital Ther. 16: 15-21. Susset, J., Tessier, C. D., Wincze, J., Bansal, S., Malhotra, C, and Schwacha, M. (1989). Effect of yohimbine hydrochloride on erectile impotence: A double-blind study. J. Urol. 141: 1360-1363.

Archives of Sexual Behavior, Vol. 26, No. 1, 1997

Classical Conditioning of Female Sexual Arousal Elizabeth J. Letourneau, Ph.D.,1.3 and William O'Donohue, Ph.D.2

The classical conditioning of subjective and physiological aspects of female sexual arousal was examined. Experimental subjects were run in a delayed conditioning design, where an amber light was paired with excerpts from erotic videos. Control subjects received presentations of the same amber tight and videos, but these presentations did not overlap with one another. Dependent variables included subjective ratings of arousal to the light and the videos as well as change in vaginal pulse amplitude assessed during exposure to the different stimuli. Experimental subjects evidenced increased arousal to the light over conditioning trials, as assessed by subjective ratings of sexual arousal. This finding is suggestive of a learned effect. However, results failed to indicate significant differences between experimental and control groups. Therefore, the increased arousal to the light evidenced by experimental subjects may not be due to classical conditioning. Suggestions regarding these findings, clinical implications, and future research are discussed. KEY WORDS: female sexual arousal; vaginal photoplethysmography; classical conditioning.

INTRODUCTION To date, no studies examining the relationship between conditioning procedures and female sexual arousal have been reported. Yet, an understanding of the effects of classical conditioning on female sexual arousal is important for two reasons. First, classical conditioning may be an etiological mechanism in the development of female sexual dysfunction, particularly 1Medical University

of South Carolina and National Crime Victims Research and Treatment Center, Charleston, South Carolina. 2Department of Psychology, University of Nevada, Reno, Nevada. 3To whom correspondence should be addressed at Department of Psychology, Augusta State University, 2500 Walton Way, Augusta, Georgia 30904. 63 0004-0002/97/0200-0063$12.50/0 C 1997 Plenum Publishing Corporation

64

Letourneau and O'Donohue

for victims of rape or child sexual assault (Becker et al., 1986, 1982; Courtois, 1979; Letourneau and O'Donohue, 1993; O'Donohue and Geer, 1993). Second, the more general assumption that sexual behavior is at least partially learned is common across numerous theories of sexual behavior. For example, behavioral accounts have viewed the acquisition of a sexual repertoire as resulting from both operant and classical conditioning procedures (Skinner, 1969; Watson, 1925). Learning processes are also influential in anthropological, sociological, developmental, and feminist accounts of sexuality, although these theories frequently do not postulate specific learning mechanisms responsible for the acquisition of sexual behavior (O'Donohue and Plaud, 1994). Given the specific hypothesis that classical conditioning plays an etiological role in the development of female sexual dysfunction, and the more general hypothesis that much sexual behavior is acquired through learning processes, basic research in the area of classical conditioning of human sexual arousal is warranted. Research has been conducted with male subjects that indicates male sexual arousal may be influenced by classical conditioning paradigms (Langevin and Martin, 1975; McConaghy, 1967; Rachman, 1966; Rachman and Hodgson, 1968). However, research on female sexual functioning has lagged substantially behind that of male sexual functioning (Beck and Baldwin, 1994). Thus, one goal of the present study was to extend these research efforts to include female subjects. A second goal was to determine appropriate parameters for the classical conditioning of female sexual arousal. In any conditioning experiment there are numerous variables that impact on the likelihood of achieving a conditioned response, such as the choice of the conditional stimulus (CS) and the un conditional stimulus (UCS). To the extent possible, variables employed in the present study were based on research with classical conditioning of male sexual arousal and research on other aspects of female sexual arousal. Every effort was made to maximize the salience and/or effectiveness of each of these variables. Rationales for decisions regarding conditioning parameters in the present study are presented below, in the Method section. The present study attempted to classically condition an arousal response in female subjects. The specific hypotheses follow. Hypothesis 1: Based on classical conditioning theory, it was hypothesized that vaginal pulse amplitude (VPA) and subjective ratings of arousal would significantly increase to a previously neutral stimulus over the course of delayed conditioning trials. Hypothesis 2: As traditional conditioning theory proposes that conditioning will not occur unless the CS and UCS are contiguous, it was hypothesized that neither vaginal pulse amplitude nor subjective ratings of

Conditioning of Arousal

65

arousal would significantly increase to a neutral stimulus during an explicitly unpaired paradigm. Hypothesis 3: The CS-UCS interval and/or the length of CS-only presentation times might impact on the magnitude of conditioning. Thus, the CS-UCS interval and the CS-only presentation time were altered for the experimental subjects, resulting in two experimental groups. Similarly, the CS presentation time was altered for the control subjects, resulting in two control groups. Analyses were conducted to determine whether these changes impacted on the responses to the stimuli.

METHOD Subjects Twenty-five women between ages 18 and 40 were recruited at a Midwestern university. Participants received $25 at the end of the study. Most subjects were Caucasian (76%), with 12% Hispanic and 12% Asian. Subjects were either dating (92%) or single (8.0%). The mean age of subjects was 21.04 (SD = 2.72) with a mean of 14.50 years of education (SD = 1.22). All subjects were heterosexual, between the ages of 18 and 40 years, experienced consistent menses (operationalized as lasting between 27 and 32 days with variations no greater than 3 days), and were neither pregnant nor currently infected with a sexually transmitted disease (STD). The first 13 subjects were randomly assigned to either an experimental group (Exper-1; n = 6) or a control group (Control-1; n = 7). As mentioned in Hypothesis 3, changes were made in the experimental treatments and the final 12 subjects were randomly assigned to either an experimental group (Exper-2; n = 6) or a control group (Control-2; n = 6). Apparatus and Materials

Conditioning Stimulus. The CS was an amber, incandescent, 25-watt light. This amber light was qualitatively different from the background lighting (i.e., colored and incandescent vs. fluorescent) and quantitatively different due to the lower wattage of the CS, presumably increasing the saliency of this stimulus. The choice of amber coloring was due in part to the fact that lay descriptions of erotic and/or romantic settings often specify background lighting that provides a soft orange glow. Unconditional Stimulus. The UCSs were 22 excerpts from heterosexual erotic videotapes. Video was chosen as the presentation medium as it

66

Letourneau and O'Donohue

appears to be more salient than audiotape or slides for generating sexual arousal in females (Heiman, 1980; Rubinsky et al., 1987). The scenes chosen for this study were rated as at least moderately arousing (5 on a scale of 1 to 10) by three raters. Half of these scenes depicted explicit acts of cunnilingus and half depicted explicit acts of intercourse. Each scene lasted approximately 2 min. This relatively long time frame was chosen as it is unknown how long it takes women to become subjectively and physiologically aroused to erotica. All excerpts were taken from female-oriented erotic videotapes, involving mutual, nondenigrating, and nonabusive sexual interactions. Research findings indicate that women become more aroused and experience fewer negative cognitions or affective responses when exposed to female-oriented versus male-oriented erotica (Heiman, 1977; Laan et al., 1994). To reduce habituation, each scene was different from the others (Meuwissen and Over, 1990). However, it was not cost effective to locate a novel scene for each of the 50 conditioning trials. Interstimulus Interval. The interstimulus interval (ISI) refers to the time between onset of the CS and onset of the UCS during a conditioning trial. The ISI is a critical parameter in conditioning (Domjan and Burkhard, 1986). To increase the opportunity for conditioning to occur the present study examined two values of the ISI. Thus, a relatively long ISI of 10 sec was used for the first experimental group and a shorter ISI of 1.0 sec was used for the second experimental group. Probe Trials. Probe trials occur when the CS is presented alone. Arousal is measured to determine whether any change from baseline has occurred. Again, because length of time required for manifestation of physiological and subjective changes in female sexual arousal is unknown, two different probe trial lengths were utilized. The initial experimental group received relatively long exposures to the CS of 130 sec (corresponding with length of exposure to the CS during conditioning trials). The second experimental group received probe trials of approximately 30 sec (corresponding with length of exposure to the CS during their conditioning trials). Three probe trials per session were employed to facilitate detection of patterns of change in arousal to the CS. Intertrial Interval, The intertrial interval (ITI) refers to the time between conditioning trials, the purpose of which is to allow physiological and subjective sexual arousal to return to baseline. Most studies that involve photoplethysmography differ in terms of ITIs. However, many researchers (e.g., Adams et al., 1985; Cerny, 1978; Geer et al., 1974; Hoon et al., 1976) have reported return to baseline arousal within 8 min or less. Therefore, ITI's were set for 8 min in the present study. Control Groups. Another critical decision concerned the type of control group to employ in the present study. Based on his contingency theory,

Conditioning of Arousal

67

Rescorla (1967) argued for the truly random design. A problem with the truly random procedure is the potential for conditioning to occur if the CS and UCS overlap (Kremer, 1974; Kremer and Kamin, 1971). Therefore, an alternative type of control procedure was utilized. This control procedure, known as the explicitly unpaired control procedure, involves presenting control subjects with randomly ordered CSs and UCSs so that they receive the same number of presentations as the experimental subjects. However, the CS and UCS never overlap (Domjan and Burkhard, 1986; Kremer and Kamin, 1971). Number of Conditioning Trials. Another important parameter for the present study was the number of conditioning trials. Review of the literature on conditioning of male sexual arousal suggested that 50 conditioning trials should be sufficient to achieve a conditioned response (CR) of increased sexual arousal (Rachman, 1966; Rachman and Hodgson, 1968). Therefore, SO conditioning trials were employed. Menstrual Cycle. Some researchers have found differences in levels of arousal corresponding with different menstrual cycle phases (Henson et al., 1977; Wincze et al., 1976), although Slob et al. (1991) reported that their subjects evinced similar levels of arousal across trials, despite being in different phases of their cycles. Given the inconclusive findings in this area, we attempted to run subjects at the same time in their cycle, 1 week after the last day of menstrual blood flow, during the follicular phase (or late follicular phase, depending on how long subjects took to complete the five experimental sessions). The Vaginal Photoplethysmograph. In the present study, A Farrall Instrument CAT-400UL model with the SVM-60UL Photoplethysmograph was employed for the assessment of vaginal pulse amplitude. The Farrall photoplethysmograph consisted of a light-emitting diode and a photoresistive cell, both housed in a small acrylic sphere. This sphere (about 1.5 inches in length and 0.5 inches in width) is inserted into the vaginal canal by the subject in private. Vaginal pulse amplitude was recorded 24 times per second over the entire course of each experimental session. Sterilization of the photoplethysmograph involved soaking in Cidex 7 (a commercial antibacterial solution) for a minimum of 1 hr and then thorough cleaning by hand with an antibacterial liquid soap. Subjective Arousal Ratings. A 100-point Likert scale was used to assess subjective level of arousal. This scale specified arousal to the amber light or to a video, depending on the stimulus trial. Personal History Questionnaire. A modified version of the Personal History Questionnaire (PHQ) developed by Heiman (1977) was used in the present study. This questionnaire contains items on demographics, sexual history, sexual orientation, heterosexual experiences, masturbation ex-

68

Letourneau and O'Donohue

periences, and experience with erotica. Additional items assessed for sexual problems and irregular menstrual cycle. Laboratory Set-Up. A sound-controlled chamber was used in which a recliner was against one wall, with the CS and video monitor approximately 6 feet from the chair. The CS was directly adjacent to the right of the monitor, on the same cart. Rating materials were to the right of the subject, on a small table. The examiner was in an adjoining room, with auditory but not visual contact. Design and Procedure

Experimental Sessions After an initial informational group session, subjects attended five separate experimental sessions on an individual basis. Upon arriving at the laboratory for these experimental sessions, subjects were greeted and then left alone in the experimental chamber where they inserted the vaginal photoplethysmograph and then relaxed in a recliner for 8 min. Subjects continued to sit quietly for an additional 2 min during which resting vaginal pulse amplitude readings were obtained and after which they subjectively rated sexual arousal. Following the 2-min resting period, subjects were presented with the CS for 2 min. Immediately following offset of the CS a 1-min ITI began, during which subjects rated their subject arousal to the CS. Subjects received three identical presentations of the CS in the first experimental session to assess baseline levels of VPA and subjective arousal to the CS. Details for the remainder of Session 1 and Session 2 through 4 are presented below by group. In addition, Fig. 1 depicts the relationship between the CS and UCS for each of the four groups. Exper-l Group. Following the final baseline measure in Session 1, Exper-1 subjects received the first conditioning trial. During conditioning trials, the CS was presented alone for 10 sec at which time a UCS was presented. The CS remained on during the length of the 2-min erotic video and then the CS and UCS were terminated together. Immediately following the conditioning trial, subjects provided ratings of their sexual arousal to the video. Conditioning trials were followed by 8-min ITIs, during the first 3 min of which subjects completed a backward-counting technique to reduce the likelihood that they would engage in sexual fantasies. There were 10 conditioning trials per session, resulting in SO conditioning trials over the course of the experiment. Three probe trials, during which the CS was presented alone for 130 sec, were interspersed with the 10 conditioning trials in each session. This resulted in 15 probe trials over the course of

Conditioning of Arousal

69

Fig. 1. Presentation times for CS and UCS, interstimulus intervals, and intertrial intervals for Exper-1, Exper-2, Control-1 and Control-2 groups.

the five experimental sessions. Subjects rated their arousal to the CS during the 1-min ITI following each probe trial. The protocol for Sessions 2-5 for the Exper-1 group was identical to that of Session 1, with the exception that the CS was not shown for three baseline presentations at the start of these remaining sessions. Each session took approximately 40 min. Exper-2 Group. The protocol for the Exper-2 group was nearly identical to that of the Exper-1 group with three exceptions. First, the ISI was only 1 sec (vs. 30 sec for the Exper-1 group). Second, the CS overlapped the UCS by only 30 sec, thus ending 90 sec earlier than in the Exper-1 group. Third, probe trials for the Exper-2 group lasted 30 sec (vs. 130 sec for the Exper-1 group). Again, each session took approximately 40 min.

70

Letourneau and O'Donohue

Control-1 Group. Many of the parameters employed with the Exper-1 group were the same for the Control-1 group and presentation order of the videos was identical for all groups. However, presentations of a CS or UCS were determined randomly and did not overlap. Following each CS presentation there was a 1-min ITI. Following each UCS presentation there was an 8-min ITI. Three CS presentations were designated as probe trials and these occurred after the same UCSs as in the Exper-1 protocol. Subjective arousal ratings were completed by Control-1 subjects after each UCS and after the three probe trials. Sessions for the Control-1 group were longer than for the experimental groups because the CS and UCS never overlapped. These sessions lasted approximately 70 min. Control-2 Group. The protocol for the Control-2 group was identical to that of the Control-1 group with one exception. All presentations of the CS, including the probe trials, were for 30 sec (vs. the 130-sec presentations received by the Control-1 group). Each session lasted approximately 60 min. Subjects in all groups completed the five sessions within a mean of 6.7 days (range = 5-12, SD = 3.92) after which they completed a postexperimental questionnaire assessing affect during the sessions, and were debriefed and reimbursed. Data Reduction

Physiological Data Collection Different researchers have employed their own methods for data reduction. The most common data reduction technique for VPA involves taking the mean VPA, either at predetermined points during the experiment (e.g., every 10 sec) or for a specific number of pulses in the set of pulses that includes the maximum VPA response to a stimulus (e.g., Heiman et al., 1991; Osborn and Pollack, 1977; Schreiner-Engle et al., 1981). The present study utilized the mean of the highest seven consecutive data points during a stimulus period. Furthermore, these were used in the determination of change scores because VPA frequently failed to return to baseline and because there is no uniform baseline pulse amplitude with which to compare different subjects. In the present study, data reduction involved a multistep process. First, the mean of the seven highest consecutive VPA peaks in a probe trial was computed by a computer program designed for this study. The mean from the seven highest consecutive peaks for the ITI immediately preceding that probe trial was also computed. The ITI mean was subtracted

Conditioning of Arousal

71

from the probe trial mean, resulting in a change score. The next step involved taking the mean of all three probe trial change scores per session, resulting in five probe trial data points per subject, one for each session. A similar process was employed to create mean change scores for the UCS stimulus periods. Due to experimenter error and equipment failure, there were some instances of missing pulse amplitude data. Specifically, 6 subjects had one session of data missing. Missing pulse amplitude data for subjects were replaced with the mean value from the subjects' group. This procedure of replacing missing data was utilized to limit the effects of unequal sample sizes on the analyses. Subjective Data Collection Changes scores were also utilized with the subjective data. Resting arousal was rated at the start of each session. This rating was subtracted from the rating of arousal to each probe trial, resulting in three change scores per session. The mean was taken of these three scores, resulting in five subjective change score data points per subject. A similar process was utilized to assess changes in subjective arousal to the videos.

RESULTS Sexual Experiences On average, subjects experienced masturbation at age 9, their first orgasm at age 16, and their first intercourse at age 16. They had a mean of nine sexual partners with whom they had experienced a variety of sexual activities including kissing, manual stimulation, oral-genital stimulation, and intercourse. One third of subjects had experienced anal intercourse. Subjects tended to have limited experience with erotica, with 80% having read 10 or fewer erotic magazines and having viewed 10 or fewer erotic videos. Although no subjects had ever been diagnosed with a sexual dysfunction, many subjects had experienced a sexual problem at least one time. The most common problems were lack of lubrication (66.7%) and inability to experience orgasm (62.5%). Other sexual problems included lack of pleasure (37.5%), painful intercourse (29.2%), lack of desire (29.2%), and inability of penis to enter vagina (12.5%). In addition, 9 subjects (36.0%) indicated they had some form of completed or attempted sexual assault

72

Letourneau and O'Donohue

during their lifetime. Results for these 9 subjects were not significantly different from the remaining subjects and their data were included in all analyses. Analysis of Sexual Arousal Analysis of Arousal to the Unconditional Stimuli It was assumed that subjects would experience significantly higher levels of arousal to the UCSs than at resting. To test this assumption a 4 (Group: Exper-1, Control-1, Exper-2, Control-2) x 2 (Time: mean of the resting pulse amplitude and mean of the pulse amplitude to the 22 UCSs) repeated measures ANOVA was conducted. Time was the within-subjects factor and Group was the between-subjects factor. Vaginal pulse amplitude was the dependent measure. The Group x Time interaction was not significant, F(3, 21), p > 0.05, nor was the main effect for Group, F(3, 21) = 2.32, p > 0.05. The main effect for Time was significant, F(l, 21) = 8.53, p < 0.01. The mean VPAs at resting for the Exper-1, Control-1, Exper-2, and Control-2 groups were, respectively, 10.86, 14.76, 12.84, and 15.68. The mean VPAs to the videos were, respectively, 14.53, 20.40, 20.51, and 21.57. Thus, pulse amplitude increased from resting to the videos for all four groups. A similar 4 (Group) x 2 (Time) repeated measures ANOVA was conducted with subjective ratings of arousal as the dependent measure. Again, neither the Group x Time interaction, F(3, 21) = .39, p > 0.05, nor the main effect for Group, F(3, 21) = 1.23, p > 0.05, was significant main effect for Time, F(l, 21) = 170.00, p < 0.001. The subjective ratings of arousal at resting for the Exper-1, Control-1, Exper-2, and Control-2 groups averaged, respectively, 0.00, 4.28,1.67, and 16.67. The corresponding mean subjective ratings to the 22 videos for each group were 55.58, 46.06, 49.45, and 57.94. Thus, subjective ratings increased from resting to the videos. These findings suggest that subjects found the videos sexually arousing. Analysis of Arousal to the Conditional Stimulus A second set of analyses was conducted to test the assumption that the amber light employed as the CS was initially a neutral stimulus that would not elicit arousal prior to the conditioning trials. A 4 (Group) x 2 (Time: resting pulse amplitude and baseline pulse amplitude to the CS) repeated measures ANOVA was conducted. The Group x Time interaction was not significant, F(3, 21) = 2.07, p > 0.05, nor was there a significant

Conditioning of Arousal

73

main effect for Group, F(3, 21) = 0.62, p > 0.05. There was a significant effect for Time, F(l, 21) = 27.34, p < 0.001. The VPA resting scores for the Exper-1, Control-1, Exper-2, and Control-2 groups were, respectively, 10.86, 14.76, 12.84, and 15.68. The VPA baseline scores for each group were 8.84, 9.77, 8.14, and 6.22. Thus, pulse amplitude actually decreased from resting to baseline. A similar 4 (Group x 2 (Time) repeated measures ANOVA was conducted utilizing the subjective arousal ratings as the dependent measure. The Group x Time interaction was not significant, F(3, 21) = .55, p > 0.05. Neither the main effect for Group, F(3, 21) = 2.74, p > 0.05 or Time, F(l, 21) = .08, p > 0.05 was significant. These findings suggest that subjects did not experience significant levels of arousal to the CS prior to the experiment. Analysis of the Conditioning Paradigm

Exper-1 Versus Control-1 It was hypothesized that the Exper-1 group would show an increase in arousal to the CS as a function of conditioning trials over and above any changes in arousal to the CS evinced by the Control-1 group. To test this hypothesis, a 2 (Group: Exper-1 and Control-1) x 6 (Time: baseline and the five probe trial scores from Sessions 1-5) repeated measures ANOVA was conducted. Vaginal pulse amplitude was the dependent measure. There was little change in VPA to the CS from baseline through the five probe trial scores. The Group x Time interaction was not significant, F(5, 55) = 1.26, p > 0.05. Neither the main effect for Time, F(5, 55) = .50, p > 0.05, nor the main effect for Group, F(l, 11) = 4.19, p > 0.05 were significant. A similar 2 (Group) x 6 (Time) repeated measures ANOVA was conducted utilizing subjective ratings of arousal as the dependent measure. Both groups evinced an increase from baseline to Session 1, although this difference was not significant. The Group x Time interaction was not significant, F(5, 55) = 1.64, p > 0.05, nor were the main effects for Time, F(5, 55) = 1.65, p > 0.05 or Group, F(l, 11) = 0.03, p > 0.05. Neither of these ANOVAs supported the hypothesis that the Exper-1 group would experience a conditioned effect. Exper-2 Versus Control-2 It was hypothesized that the Exper-2 group would show an increase in VPA from baseline as a function of the conditioning trials, and that this

74

Letourneau and O'Donohue

increase would be greater than changes in the Control-2 group. To test this hypothesis, a 2 (Group: Exper-2 and Control-2) x 6 (Time: baseline VPA and the five probe trial scores) repeated measures ANOVA was conducted. Vaginal pulse amplitude change score was the dependent measure. There was virtually no change in pulse amplitude from the initial presentations of the CS through the experimental sessions for either group. The Group x Time interaction was not significant, F(5, 50) = 1.18, p > 0.05. Neither the main effect for Time, F(5, 50) = 0.39, p > 0.05, nor for Group, F(l, 10) = 2.33, p > 0.05 was significant. A 2 (Group x 6 (Time) repeated measures ANOVA was conducted for the Exper-2 and Control-2 groups, with subjective ratings of arousal change score as the dependent measure. Again, the Group x Time interaction was not significant, F(5, 50) = 1.07, p > 0.05. Neither the main effect for Time, F(5, 50) = 0.48, p > 0.05, nor for Group, F(l, 10) = 1.90, p > 0.05, was significant. Neither of the ANOVAs conducted with the Exper-2 and Control-2 groups support the hypothesis that the Exper-2 group would show greater levels of sexual arousal to the CS. Effects of the CS Presentation Time Experimental Groups To determine whether the change in ISI and probe-trial presentation time significantly impacted on the results, analyses were conducted comparing physiological and subjective arousal levels between the two experimental groups. A 2 (Group: Exper-1 and Exper-2) x 6 (Time: baseline VPA and probe trial VPA scores) repeated measures ANOVA was conducted. Mean change in VPA was the dependent measure. These differences were not significant. The Group x Time interaction was not significant, F(5,50) = 1.16, p > 0.05, nor were the main effects for Time, F(5,50) = 0.52, p > 0.05, or Group, F(l, 10) = 2.41, p > 0.05. A 2 (Group) x 6 (Time) repeated measures ANOVA was conducted for the two experimental groups with subjective arousal rating change scores as the dependent measure. Both groups evinced an increase over the initial presentations of the CS. Neither the Group x Time interaction, F(5, 50) = 0.82, p > 0.05, nor the main effect for Group, F(l, 10) = 1.47, p > 0.05, was significant. The main effect for Time was significant, F(5, 50) = 2.74, p < 0.05. The significant time effect reflects the general increase in performance over time for both groups. However, these results do not support the hypothesis that the change in CS presentation time and

Conditioning of Arousal

75

CS-UCS interval made a significant impact on the arousal patterns of the experimental groups. Relationship Between Physiological and Subjective Measures Correlations were computed between these two methods of assessment for each CS data point (i.e., baseline and probe trials for Sessions 1-5). The correlations between physiological and subjective assessment of arousal were significant and positive for Session 2, r(23) = 0.46, p < 0.05, and Session 4, r(23) = 0.54, p < 0.01. None of the remaining four correlations were significant. In general, these results are similar to those reported in other studies where significant correlations are rare.

DISCUSSION The primary hypothesis in the present study concerns the pattern of arousal to the CS exhibited by the experimental versus the control subjects. The results from between-group comparisons fail to show differences between the experimental and control groups. Thus, the present study does not support the hypothesis that female sexual arousal may be classically conditioned. There are several possible explanations for these findings. The first is that female sexual arousal may be insusceptible to the effects of classical conditioning procedures. There is no previous literature that supports, or opposes, the view that female sexual arousal may be classically conditionable. Several studies appear to indicate that male sexual arousal is conditionable. However, these studies (e.g., Langevin and Martin, 1975; Rachman, 1966; Rachman and Hodgson, 1968) tended to report small effects and often lacked proper control groups. Consequently, at least some of these positive findings may be due to the effects of sensitization or pseudoconditioning. On the other hand, it has been suggested that the sexual arousal of male subjects may be more easily conditioned than that of female subjects, perhaps due to hormonal differences (Bancroft, 1988; Feierman, 1990). Future research conducted in this area should include both male and female subjects to determine whether conditionability is gender-specific. A second possible explanation for the lack of differences between the experimental and control groups is that the parameters of the present study may not have been optimal for facilitating a conditioned response. For example, the UCSs may have been too weak, the CS may been have easily

76

Letourneau and O'Donohue

habituated to as it was static, and the interstimulus intervals may have been too long or too short. One argument against the idea that the UCSs were too weak is that, in a postexperimental questionnaire, 96% of the subjects indicated they had been "moderately" to "considerably" aroused. Most subjects also indicated high levels of interest and enjoyment the study, although 36% reported moderate to considerable boredom (due, probably, to the 8-min ITIs about which subjects often complained). A third explanation regarding the null results is that the small number of subjects, and concomitant decrease in statistical power, limited the ability of the analyses to detect true differences between groups. A significant time effect was found for the two experimental groups. Specifically, both experimental groups evinced an increase in responding over time through the fourth experimental session. This increase is consistent with a learned response. Finding a significant, across sessions, result with these small samples is suggestive of a "real" finding. However, this effect did not emerge when these groups were compared with their respective control groups, possibly due to the small samples and lack of power. Future research should focus on maximizing statistical power by the use of larger samples. Another concern is that there was a total of 15 probe trials for the experimental subjects. This may have allowed extinction to influence the arousal patterns of these subjects. Also, many of these subjects indicated they had experienced sexual problems. Although none had been diagnosed with a sexual dysfunction, their rates of sexual problems were higher than might be expected (Laumann et al. (1994) reported between 14 and 32% of women ages 18-24 experienced sexual problems, whereas 2.5 to 66.7% of women in the present study endorsed similar sexual problems. Either extinction or sexual problems may have negatively impacted on the conditioning process. It is recommended that future studies employ a limited number of probe trials and (again) a larger sample to allow for comparisons between subjects with and without sexual problems. Several researchers and clinicians have suggested that female sexual arousal decreases after sexual assault as a result of conditioning (Becker et al., 1986; Falsetti and Resnick, 1994; Kilpatrick et al., 1982). Rape and other sexual assault may facilitate aversive conditioning, or inhibitory conditioning. It would therefore be useful to determine whether it is possible to condition an aversive response to a previously sexually arousing stimulus. This research would serve more specifically to address the question of whether classical conditioning accounts for sexual dysfunction following rape or sexual molestation. Female sexual arousal is a multifaceted construct. Research has only just begun to attempt to determine the impact of various procedures and experiences on female sexual arousal. Although fascinating, there remain

Conditioning of Arousal

77

more "unknowns" and "knowns" in this area of research and additional empirical studies are required before definitive answers can be made regarding the question of whether female sexual arousal can be conditioned. It is suggested that statements about variables and procedures that may affect female sexual arousal should be made cautiously; that while some explanations of changes in arousal may seem logical (e.g., classical conditioning), researchers and clinicians must temper their enthusiasm with which these explanations are endorsed until further research has provided more evidence regarding their utility. It is important that clinicians and researchers continue to study the feasibility of this explanation and avoid the assumption that female sexual functioning is easily explained.

REFERENCES Adams, A. E., Haynes, S. N., and Brayer, M. A. (1985). Cognitive distraction in female sexual arousal. Psychophysiology 22: 689-696. Bancroft, J. (1988). Sexual desire and the brain. Sex. Marital Ther. 3: 11-27. Beck, J.G., and Baldwin, L. E. (1994). Instructional control of female sexual responding. Arch. Sex. Behav. 23: 665-684. Becker, J. V., Skinner, L. J., Abel, G. G., and Cichon, J. (1986). Level of postassault sexual functioning in rape and incest victims. Arch. Sex. Behav. 15: 37-49. Becker, J. V., Skinner, L. J., Abel, G. G., and Treacy, E. C. (1982). Incidence and types of sexual dysfunctions in rape and incest victims. J. Sex Marital Ther. 8: 65-74. Cerny, J. A. (1978). Biofeedback and the voluntary control of sexual arousal in women. Behav. Ther. 9: 847-855. Courtois, C. (1979). The incest experience and its aftermath. Victimol. Int. J. 4: 337-347. Domjan, M., and Burkhard, B. (1986). The Principles of Learning and Behavior, 2nd ed., Brooks/Cole, Pacific Grove, CA. Falsetti, S. A., and Resnick, H. S. (1994). Helping victims of violent crime. In Freedy, J. R., and Hobfoll, S. E. (eds.), Traumatic Stress: Theory to Practice, Plenum Press, New York, pp. 20-21. Feierman, J. R. (1990). Introduction. In J. R. Feierman (ed.), Pedophilia: Biosocial Dimensions, Springer-Verlag, New York, pp. 1-8. Geer, J. H., Morokoff, P., and Greenwood, P. (1974). Sexual arousal in women: The development of a measurement device for vaginal blood volume. Arch. Sex. Behav. 3: 559-564. Heiman, J. R. (1977). A psychophysiological exploration of sexual arousal patterns in females and males. Psychophysiology, 14: 266-274. Heiman, J. R. (1980). Female sexual response patterns: Interactions of physiological, affective, and contextual cues. Arch. Gen. Psychiat. 37: 1311-1316. Heiman, J. R., Rowland, D. L., Hatch, J. PL, and Gladue, B. A. (1991). Psychophysiological and endocrine responses to sexual arousal in women. Arch. Sex. Behav. 20:171-186. Henson, D. E., Rubin, H. B., Henson, C, and Williams, J. (1977). Temperature changes of the labia minora as an objective measure of human female eroticism. J. Behav. Ther. Exp. Psychiat. 8: 401-410. Hoon, P. W., Wincze, J. P., and Hoon, E. F. (1976). Physiological assessment of sexual arousal in women. Psychophysiology 13: 196-204. Kilpatrick, D. G., Veronen, L. J., and Resnick, H. S. (1982). Psychological sequelae to rape. In Doleys, D. M., Meredith, R. L., and Ciminero, A. R. (eds.), Behavioral Medicine: Assessment and Treatment Strategies, Plenum Press, New York, pp. 473-497.

78

Letourneau and O'Donohue

Kremer, E. F. (1974). The truly random control procedure: Conditioning to the static cues. J. Compar. Physiol. Psychol. 86. 700-707. Kremer, E. F., and Kamin, L. J. (1971). The truly random control procedure: Associative or nonassociative effects in rats. J. Compar. Physiol. Psychol. 74: 203-210. Laan, E., Everaerd, W., van Bellen, G., and Hanewald, G. (1994). Women's sexual and emotional responses to male and female erotica. Arch. Sex. Behav. 23: 153-169. Langevin, R., and Martin, M. (1975). Can erotic responses be classically conditioned? Behav. Ther. 6: 350-355. laumann, E. O., Gagnon, J. H., Michael, R. T., and Michaels, S. (1994). The Social Organization of Sexuality: Sexual Practices in the United States, University of Chicago Press. Chicago, p. 371. Letourneau, E. J., and O'Donohue, W. (1993). Sexual desire disorders. In Geer, J., and O'Donohue, W. (eds.), Handbook of the Assessment and Treatment of Sexual Dysfunctions, Allyn & Bacon, New York, pp. 135-157. McConaghy, N. (1967). Penile volume change to moving pictures of male and female nudes in heterosexual and homosexual males. Behav. Res. Ther. 5: 43-48. Meuwissen, I., and Over, R. (1990). Habituation and dishabituation of female sexual arousal. Behav. Res. Ther. 28: 217-226. Nathan, S. G. (1986). The epidemiology of the DSM-III psychosexual dysfunctions. J. Sex Marital Ther. 12: 267-281. O'Donohue, W., and Geer, J. (1993). Research issues in sexual dysfunction. In Geer, J., and O'Donohue, W. (eds.), Handbook of the Assessment and Treatment of Sexual Dysfunction, Allyn & Bacon, New York, pp. 1-15. O'Donohue, W., and Plaud, J. (1994). The conditioning of human sexual arousal. Arch. Sex. Behav. 23: 321-344. Osborn, C. A., and Pollack, R. H. (1977). The effects of two types of erotic literature on physiological and verbal measures of female sexual arousal. J. Sex Res. 13: 250-256. Rachman, S. (1966). Sexual fetishism: An experimental analogue. Psychol. Record 16: 293-296. Rachman, S., and Hodgson, R. J. (1968). Experimentally-induced "sexual fetishism": Replication and development. Psychol Record 18: 25-27. Rescorla, R. A. (1967). Pavlovian conditioning and its proper control procedures. Psychol. Rev. 74: 71-80. Rubinsky, H. J., Eckerman, D. A., Rubinsky, E. W., and Hoover, C. R. (1987). Early-phase physiological response patterns to psychosexual stimuli: Comparison of male and female patterns. Arch. Sex. Behav. 16: 45-56. Schreiner-Engel, P., Schiavi, R. C, and Smith, H. (1981). Female sexual arousal: Relation between cognitive and genital assessments. J. Sex Marital Ther. 7: 256-267. Skinner, B. F. (1969). Contingencies of Reinforcement: A Theoretical Analysis, Appleton-Century-Crofts, New York. Slob, A. K., Emste, M., and Van der Werff ten Bosch, J. J. (1991). Menstrual cycle phase and sexual arousability in women. Arch. Sex. Behav. 20: 567-577. Spector, I. P., and Carey, M. P. (1990). The incidence and prevalence of the sexual dysfunctions: A critical review of the empirical literature. Arch. Sex. Behav. 19: 389-408. Watson, J. B. (1925). Behaviorism, Norton, New York. Wincze, J. P., Hoon, P., and Hoon, E. F. (1976). Physiological responsivity of normal and sexually dysfunctional women during erotic stimulus exposure. J. Psychosom. Res. 20: 445-451.

Archives of Sexual Behavior, Vol. 26, No. 1, 1997

Assessment of Sexual Functioning for Chinese College Students Catherine So-kum Tang, Ph.D.,1,2 Florence Duen-mun Lai, M. Phil.,1 and Tony K. H. Chung, MBBS, FRACOG1

Participating in this study were 305 Chinese college students in Hong Kong. Objectives included (i) examination of the psychometric properties of the Chinese version of the Derogatis Sexual Functioning Inventory (DSFI), (ii) exploration of the associations among various domains of sexual functioning, and (in) description of Chinese students' sexual behavior. The Chinese DSFI shows satisfactory internal consistency, except the Information, Drive, and Satisfaction subscales. In general, sex-related domains such as sexual information, experience, drive, attitudes, fantasy, and satisfaction were related to each other and to body image. Compared to their counterparts in the U.S., Chinese college students were relatively sexually inexperienced and conservative. Mean ages of students having their first sexual intercourse experience were 17.14 for males and 18.13 for females; 11% of the students had premarital sexual intercourse experience, and the mean frequencies were once weekly for males and once or twice per month for females. Compared to female students, twice as many males acknowledged masturbation (21.4 vs. 43.8%), 4.2-8.2% students had experienced oral-genital sexual stimulation and 0-1.2% anal sexual activities. KEY WORDS: Chinese sexuality; sexual behavior; premarital sex.

INTRODUCTION Cultural and social beliefs influence individuals' attitudes toward sex, which, in turn, affect their sexual behavior (Bhugra and de Silva, 1993; Gregersen, 1986; Unwin, 1934). Chinese sexuality is based mainly on the 1Chinese 2To

University of Hong Kong, Shatin, New Territories, Hong Kong. whom correspondence should be addressed.

79 0004-0002W/0200-007W12.50/0 C 1997 Plenum Publishing Corporation

80

Tang, Lai, and Chung

Confucian and Taoist traditions (Chang, 1977; Ng and Lau, 1990; Ruan and Matsumura, 1991; Tsang, 1988; Wen, 1995). According to the Confucian sexual philosophy, which emphasizes procreation and social order, sex for pleasure and outside wedlock is prohibited for both genders. The Taoist tradition, on the other hand, focuses on the balance between Yin and Yang, personal health, and longevity. Semen is valued as a source of vitality and needs to be preserved by avoiding ejaculation during intercourse. Masturbation and excessive sexual activities are viewed as causes for men's illnesses, as they lead to an excessive waste of semen (Wen, 1995). There is evidence that these two traditions persist in contemporary Chinese societies and affect Chinese sexual behavior, despite the influence of Western ideas and practices. In a study that examines sex-related folk beliefs in urban and rural regions of China and Taiwan, Chinese men still believe that controlled sexual activity protects health (34.7-74.2%), excessive masturbation weakens the body (31.6-65.5%), semen needs careful conservation (27.6-47.7%), menstruation is a dirty and unlucky thing (4-40.6%), and sex with a menstruating woman causes illness (11.2-64.1%) (Tseng et al., 1992). Similar sexual beliefs are also found in teenagers (Family Planning Association of Hong Kong, 1994), medical students (Chan, 1990), and men seeking help for sexual problems in Hong Kong (Lieh-Mak and Ng, 1981). With diverse cultural sexual beliefs and attitudes, individuals in different countries show variations in their sexual behavior. Among Hong Kong adolescents ages 14 to 17 years, 6.1% of the males and 4.3% of the females experienced premarital sexual intercourse (Family Planning Association of Hong Kong, 1994), whereas the percentages are much higher in other countries: 8.5% in New Zealand (Lynskey and Fergusson, 1993), 17.4% in South Africa (Fisher et al., 1993), 28.1% in Australia (Weisberg et al., 1992), 29% in Nigeria (Odujinrin, 1991), 51% in Cuba (Martinez et al., 1992), and 53-60% in the United States (Forrest and Singh, 1990; Sonenstein et al., 1991). Premarital sexual experience is also infrequent among Chinese college students in Hong Kong, 6.25% of the males and 3.45% of the females (Chan, 1990) when compared to their counterparts in Chile (78% of the males and 41% of the females, Repossi et al., 1994) and the United States (40-80%, DeBuono et al., 1990; Huang and Uba, 1992). Past studies on the sexual behavior of Hong Kong Chinese have methodological problems that make generalization to the Chinese population a cross-cultural comparison difficult. They often include biased samples of medical students (Chan, 1986, 1990) or self-constructed survey items that are of uncertain and questionable reliability and validity (Family Planning Association of Hong Kong, 1994). The present study represents the first

Chinese Sexuality

81

attempt to extend the Derogatis Sexual Functioning Inventory to Chinese college student samples. This inventory adopts a multidimensional approach to the assessment of sexual functioning (Derogatis and Melisaratos, 1979) and has an advantage over current global assessment of sexual behavior by providing information not only on the overall status of the individuals' sexual health but also on the patterns, strengths, and weaknesses of their sexual functioning. It measures 10 domains essential to sexual behavior: sexual information, sexual drive, sexual experience, sexual fantasy, sexual satisfaction, sexual attitudes, body image, gender role definition, affects, and psychological symptoms. Specific objectives of the present study include (i) to examine the psychometric properties of the Chinese version of the Derogatis Sexual Functioning Inventory, (ii) to explore whether the hypothesized associations among various domains of sexual behavior are evident in Chinese college student samples, and (iii) to describe sexual behavior and experience of Chinese college students in detail and compare results with studies in other countries.

METHOD Participants and Procedure The second author translated the Derogatis Sexual Functioning Inventory (DSFI) to Chinese with the meanings as to close to the original as possible. The original and translated versions of the DSFI were first examined in translation by two practicing clinical psychologists and an instructor to ensure trausliteral equivalence. Five college students were asked to complete the translated DSFI and take note of any difficulties in reading the items. Changes were made to the translated inventory according to the feedback. The Chinese DSFI was then administered to 160 male and 145 female Chinese college students at the Chinese University of Hong Kong. Their mean age was 20.5 years old (SD = 1.4). They came from different departments of the university and participated in this study for extra credit points in their introductory psychology courses. No names were included in the answer booklets to ensure confidentiality. Sexual Functioning The quality of sexual functioning was assessed by the DSFI (Derogatis, 1978). It is a self-report inventory and comprises 10 subscales, each

82

Tang, Lai, and Chung

measuring a distinctive domain essential to sexual behavior. The 10 subscales are Information, Experience, Drive, Attitudes (Liberalism and Conservatism), Psychological Symptoms, Affects (Positive and Negative Affects), Gender Role (Masculinity and Femininity), Fantasy, Body Image, and Satisfaction. High scores on the Information subscale indicate accurate knowledge on sexual matters, whereas high scores on Experience, Drive, Fantasy, and Satisfaction represent high levels of sexual functioning in these domains. The Attitudes subscale measures the balance between liberal and conservative attitudes regarding sexuality, with positive scores indicating more liberal balances. The Affects subtest represents the difference between positive and negative affects as reported by the respondents and high scores indicate positive affects balance. The Gender Role score is calculated as the difference between femininity and masculinity scores, with negative scores representing more masculine gender role orientation. High scores of the Body Image subscale are associated with poor body image. The DSFI subscales show satisfactory internal consistency, with Cronbach's alpha being .56, .97, .60, .81, .86, .77, .93, .94, .84, .76, .82, .58, and .71, respectively. Test-retest (2 weeks) reliability coefficients are > .90 for Experience, Attitude, Symptoms, and Fantasy subscales, > .80 for Affects and Gender Role subscales, and > .77 for Drive subscale. Factor analytic results show that the internal structure of the DSFI conforms to the hypothesized underlying domains of sexual behavior. The DSFI also discriminates between individuals with sexual disorders and individuals with normal sexual functioning (Cooper et al., 1993; Derogatis and Melisaratos, 1979; Fagan et al., 1988).

RESULTS Internal Consistency Estimates

Table I shows the internal consistency alpha values for the Chinese DSFI subscales. The Cronbach's alpha values for these subscales were similar across gender. The internal consistency estimates for the Experience, Attitudes, Fantasy, Symptoms, Affects, and Gender Role subscales were satisfactory (a. = .68 to .96). The Drive and Satisfaction subscales showed only marginal internal consistency (a. = .52 to .64). Compared to the original English DSFI, the Chinese version had higher internal consistency estimates for the Body Image (a = .60 to .71). The Information subscale showed low internal consistency for both English and Chinese versions (a = .24 to .41).

Chinese Sexuality

83

Table I. Internal Consistency of the Chinese DSFI Subscales Alpha No. Variable of items Male Female Total Information Experience Drive Attitudes-liberalism Attitudes-conservatism Symptoms Affects-positive Affects-negative Gender role-masculinity Gender role-femininity Fantasy Body image Satisfaction

26 24 5 15 15 53 20 20 15 15 20 15 10

.41 .84 .60 .73 .78 .96 .87 .95 .69 .84 .77 .71 .52

.24 .96 .64 .68 .70 .96 .94 .95 .68 .83 .80 .60 .55

.38 .89 .61 .74 .77 .96 .90 .95 .68 .84 .78 .67 .54

Table II. Intercorrelations Among Subscales of Chinese DSFI 1 2 3 4 5 6 Variable 7 8 1. Information .16b 2. Experience .31? .67c 3. Drive .31c 23* .34c 4. Attitudes — -.03 -.09 -.05 -.12a 5. Symptoms —C .05 -.26 .15b .10 .01 6. Affects — .15 b -.33c -.11 -.11 -.09 -.08 7. Gender role definition — .19c .20c .36c .38c .02 -.11 -.04 8. Fantasy -.06 -.24c -.25C -.22C .10 -.29c .37c -.06 9. Body image .17* -.09 .05 .10 .30c .18b .16b .05 10. Satisfaction ap bp cp

9

—C -.21

< 0.05 (two-tailed). < 0.01 (two-tailed). < 0.005 (two-tailed). Intercorrelations Among Subscales

Table II shows the intercorrelation results among the Subscales of the Chinese DSFI for the total sample. These results were similar across male and female students, thus, they were not reported separately. In general, sex-related variables such as sexual information, experience, drive, attitudes, fantasy, and satisfaction were positively related to each other (r = .16 to .67, p < 0.01). High scores on these variables were generally associated with good body image (r = -.21 to -.25, p < 0.005). Positive affects were related to sexual drive and satisfaction (r = .15 to .17, p < 0.01). The level of psychological symptoms was unrelated to sexrelated variables, except conservative sexual attitudes (r = -.12, p < 0.05).

84

Tang, Lai, and Chung

Gender role definition was also unrelated to the sex-related variables (p > 0.05). Chinese DSFI Subscales T tests were performed between male and female students on the 10 Chinese DSFI subscales to explore whether there were any gender differences. Results showed that, compared to female students, male students had more accurate sexual information, adopted more liberal sexual attitudes, and reported higher levels of sexual drive, fantasy, and satisfaction, (t = 4.31, 4.26, 2.89, 4.09, and 2.15, respectively,p < 0.05, Table III). Male students also had better body image than female students (t = -6.25, p < 0.001). There were no gender differences in sexual experience, psychological symptoms, and the balance between positive and negative affects (t = 0.61,1.06, and 0.06, respectively,p > 0.05). Figures 1 and 2 show the DSFI profiles for Chinese male and female students. American normative data from Derogatis and Melisaratos (1979) were also included in the two figures to provide a graphic comparison between profiles. T tests were also performed between students who had no sexual intercourse experience and those who were sexually active on the Chinese DSFI subscales. Results showed that sexually active students scored higher on the sex-related variables of information, experience, drive, attitudes, fantasy, and satisfaction as well as having a better body image (t = -2.56, -9.07, -7.00, -3.12, -2.41, -3.78, and 3.67, respectively, p < 0.01).

Table III. Mean and Standard Deviation of the Chinese DSFI Subscales Total (n - 305) Variable Information Experience Drive Attitudes Symptoms Affects Gender role definition Fantasy Body image Satisfaction

X

SD

17.04 2.64 4.29 6.79 8.64 4.90 1.98 11.27 1.06 0.70 0.74 0.93 3.14 7.75 2.81 2.77 28.89 8.29 6.29 1.99

Male (n = 160)

Female (n = 145)

X

SD

X

17.73 4.52 9.40 4.53 0.64 0.74 -0.20 3.42 25.91 6.53

2.65 7.23 4.78 11.56 1.01 0.92 8.00 2.79 7.42 1.97

16.26 4.05 7.48 -.81 0.77 0.74 6.95 2.14 32.17 6.01

SD 2.41 6.27 4.87 10.26

1.11 0.94 5.35 2.60 7.98 1.99

t value

4.31 0.61 2.89 4.26 1.06 0.06 -9.20 4.09 -6.25 2.15

P 0.00 0.54 0.00 0.00 0.29 0.96 0.00 0.00 0.00 0.03

Chinese Sexuality

85

Fig. 1. DSFI profile of male (Note: INF = information, EXP = Experience, DRIV = Drive, ATT = Attitude, SYMP = Symptoms, AFF = Affect, ROLE = Gender role, FANT = Fantasy, BODY = Body image, SATIS = Satisfaction.)

Fig. 2. DSFI profile of female. (Note: INF = information, EXP = Experience, DRIV = Drive, ATT = Attitude, SYMP = Symptoms, AFF = Affect, ROLE = Gender role, FANT = Fantasy, BODY = Body image, SATIS = Satisfaction.)

86

Tang, Lai, and Chung

Sexual Behavior of Chinese College Students The Drive subscale of the Chinese DSFI provided information about the sexual behavior of Chinese college students. Male students reported that they became interested in sexual activities at 13.94 years (SD = 2.95), whereas female students showed similar interests at 16 years (SD = 2.77). Mean ages of students experiencing their first sexual intercourse were 17.14 years (SD = 5.68) for male and 18.13 years (SD = 4.98) for female students. About 30% of the students engaged in kissing and petting with their partners, 87% of the male students reported having sexual fantasies weekly, and 59% of the female students reported having sexual fantasies monthly. About 11% of the students reported having sexual intercourse experience, and the mean frequencies were once weekly for males and once or twice per month for females. The Experience subscale of the Chinese DSFI represented a range of sexual behavior that students had experienced (Table IV). It could be broken down into five domains, including intimate and preliminary foreplay activities, anal activities, intercourse, and masturbation (Andersen and Broffitt, 1988). The percentages of students reporting experience in various sexual activities were similar across gender, except for masturbation (x2 = 16.19, p = 0.01). Compared to female students, twice as many male students acknowledged masturbation (21.4 vs. 43.8%). Few students engaged in oral-genital stimulation (4.2-8.2%) and anal sexual activities (0-1.2%). Male-superior was the most dominant position while side-by-side was the least used position in sexual intercourse.

DISCUSSION The present study attempts to extend the Derogatis Sexual Functioning Inventory to a sample of Chinese college students. Except the Information, Drive, and Satisfaction subscales, The Chinese DSFI shows satisfactory internal consistency comparable to the original English version. Similar to the results reported by Derogatis and Melisaratos (1979), the internal consistency coefficients of the Information subscale fall below the acceptable level, suggesting that it contains heterogeneous items that may not belong to the described domain. The internal reliability estimates of the Drive subscale of both Chinese and English DSFI are only marginally acceptable, whereas the estimates for the Satisfaction subscale of the Chinese DSFI are lower than the original version. Future studies should analyze the item composition of these subscales to improve its internal structure.

Chinese Sexuality

87

Table IV. Sexual Experiences of Chinese College Students % Male (n = 160)

% Female (n = 145)

23.8 21.9

23.4 17.9

Intimate foreplay activities Stroking and petting your partner's genitals Having your genitals caressed by your partner Mutual oral stimulation of genitals Oral stimulation of your partner's genitals Kissing of sensitive (nongenital) areas of the body Mutual petting of genitals to orgasm Kissing on the lips Breast petting while your are nude

7.5

8.2

10.0 29.4 14.4 36.3 23.1

12.4 28.3 11.0 33.2 24.8

Preliminary foreplay activities Your partner lying on you while you are clothed Erotic embrace while dressed Your partner kissing your nude breasts Breast petting while your are clothed Mutual undressing of each other Deep kissing

29.4 31.5 23.1 26.9 18.8 33.4

30.4 28.3 24.1 30.4 17.3 35.9

Anal activities Having your anal area caressed Caressing your partner's anal area Anal intercourse

5.0 8.2 1.2

4.2 4.9 0.7

Intercourse Vaginal entry from rear Side by side Sitting position Female superior position Male superior position

6.3 3.1 8.2 8.8

9.0 2.1 7.6

13.2

11.2 16.9

Masturbation Masturbation alone

43.8

21.4

The subscales of the Chinese DSFI show patterns of association as hypothesized by Derogatis and Melisaratos (1979). In general, sex-related variables such as sexual satisfaction, information, experience, drive, fantasy, and attitudes are related to each other and to body image. These results support the argument that healthy sexual functioning is related to accurate sexual information, liberal sexual attitudes, and positive body image (Andersen and LeGrand, 1991; Chan, 1990; Huang and Uba, 1992). Contrary to past studies on Western samples that show individuals with rigid gender role identity are more prone to sexual dysfunction (Derogatis, 1986) and have poor psychosexual adjustment after gynecologic treatment (Andersen and Wolf, 1986), the present study shows that gender role definition is unrelated to Chinese sexual expression such as sexual drive, fantasy, and sat-

88

Tang, Lai, and Chung

isfaction. Tsoi et al. (1984) also found that gender role stereotyping is unrelated to Chinese women's adjustment after surgical removal of benign gynecologic tumors. This study also describes sexual behavior of Chinese college students in Hong Kong. Compared to their counterparts in the United States (Davidson and Darling, 1993; DeBuono et al., 1990), Chinese college students are relatively sexually inexperienced, although the percentages of Chinese students having premarital sexual intercourse have increased from 6.25% of the males and 3.45% of the females (Chan, 1990) to about 11% across gender. Among those who are currently sexually active, their mean frequencies of sexual intercourse are also lower than U.S. students, weekly to monthly for Chinese versus one to three times per week for U.S. students. Such differences are also found between Chinese college students studying in the U.S. and U.S. students (Huang and Uba, 1992). The mean ages of first intercourse experience for Chinese students (17.14 years for males and 18.13 for females) are somewhat older than that of students in other countries. For example, most American males have their initial sexual intercourse by 16-17 years of age, and females by 17-18 years of age (Seidman and Rieder, 1994). The median age of South Africans at first intercourse is 15.1 years (Fisher et al., 1993), whereas most Cubans begin their sexual experience in early adolescence (Martinez et al., 1992). Several reasons account for these differences. First, Chinese parents have always focused on academic achievement, and universities in Hong Kong are very competitive and examination oriented. Students are often busy studying for lectures or preparing for examination. Thus, they have less time for dating and this, in turn, decreases chances for meeting sexual partners. Second, Chinese students may delay physical intimacy longer than their Western counterparts because they want to feel more certain that there is adequate emotional commitment (Huang and Uba, 1992). Third, most Chinese students live with their families during their college education and only about 20% live in university dormitories. Thus, there is more parental control as well as a lack of opportunity for sexual behavior. Chinese students are conservative in their sexual expression. Compared to students in other countries, fewer Chinese students report homosexual experience or experiment with alternate sexual experience beyond kissing and petting. For example, Seidman and Rieder (1994) reported that up to 20% of American adults have had homosexual experience, and approximately 25% of adults have had heterosexual anal intercourse. In Chan's (1990) study, no Chinese medical students reported sexual experience with prostitutes or homosexuality. The present results show that less than 2% of Chinese students have anal intercourse and only 4.2-8.2% engage in oral-genital sexual stimulation. The percentages of Chinese students

Chinese Sexuality

89

acknowledging masturbatory behavior are also much lower than those in the U.S. (21.4-43.8% vs. 77%; Davidson and Darling, 1993). This conservatism in sexual expression may be related to influences of Confucian and Taoist sexual philosophy that views sex as mainly for procreation and needs careful regulation to protect individual's health and social order. Chinese are thus discouraged from engaging in sexual activities that occur outside the marital relationship or do not fulfill procreation purposes. This study shows that the DSFI can be extended to Chinese samples. The Chinese DSFI yields comparable psychometric properties and associations among various domains as reported by the original inventory. Cultural beliefs and societal roles may affect the sexual expression of Chinese college students as they show different sexual behavior when compared to their counterparts in other countries. Future studies should further refine the Chinese DSFI subscale items to seek external validation of various underlying domains, and to extend its use beyond college samples.

ACKNOWLEDGMENT The authors thank Dr. Len Derogatis for permission to translate the DSFI into Chinese.

REFERENCES Andersen, B. L., and Broffitt, B. (1988). Is there a reliable and valid self-report measure of sexual behavior? Arch. Sex. Behav. 17: 509-524. Andersen, B. L., and LeGrand, J. (1991). Body image for women: Conceptualization, assessment, and a test of its importance to sexual dysfunction and medical illness. J. Sex Res. 28: 457-477. Andersen, B. J., and Wolf, F. M. (1986). Chronic illness and sexual behavior: Psychological issues. J. Consult. Gin. Psychol. 54: 168-175. Bhugra, D., and de Silva, R. (1993). Sexual dysfunction across culture. Int. Rev. Psychiat. 5: 243-252. Chan, D. (1986). Sex misinformation and misconceptions among Chinese medical students in Hong Kong. Med. Educ. 20: 390-398. Chan, D. (1990). Sexual knowledge, attitudes, and experience of Chinese medical students in Hong Kong. Arch. Sex. Behav. 19: 73-93. Chang, J. (1977). The Tao of Love and Sex: The Ancient Chinese Way to Ecstasy, Penguin, New York. Cooper, A. J., Cernovsky, Z. Z., and Colussi, K. (1993). Some clinical and psychometric characteristics of primary and secondary premature ejaculators. J. Sec Marital Ther. 19: 276-288. Davidson, J. K., and Darling, C.A. (1993). Masturbatory guilt and sexual responsiveness among post-college-age women: Sexual satisfaction revisited. J. Sex Marital Ther: 19: 289-300. DeBuono, B. A., Zinner, S. H., Daamen, M., and McCormach, W. M. (1990). Sexual behavior of college women in 1975,1986, and 1989. New Eng. J. Med. 322: 821-825.

90

Tang, Lai, and Chung

Derogatis, L. R. (1978). The Derogatis Sexual Functioning Inventory Manual, Clinical Psychometrics, Baltimore. Derogatis, L. R. (1986). The unique impact of breast and gynecologic cancers on body image and sexual identity in women: A reassessment. In Vaeth, J. M. (ed.), Body Image, Self-Esteem, and Sexuality it Cancer Patients, Karger, New York, pp. 1-4. Derogatis, L. R., and Melisaratos, N. (1979). The DSFI: A multidimensional measure of sexual functioning. J. Sex Marital Ther. 5: 244-281. Fagan, P. J., Wise, T. N., and Derogatis, L. R. (1988). Distressed transvestites: Psychometric characteristics. J. Nerv. Ment. Dis. 176: 626-632. Family Planning Association of Hong Kong. (1994). Youth Sexuality Study, 1991, The Family Planning Association of Hong Kong, Hong Kong. Fisher, A. J., Ziervogel, C. F., Chalton, D. O., Leger, P. H., and Robertson, B. A. (1993). Risk-taking behavior of Cape Peninsula high-school students. Part VIII. Sexual behavior. South Afr. Med. J. 83: 495-497. Forrest, J. D., and Singh, S. (1990). The sexual and reproductive behavior of American women, 1982-1988. Fam. Plann. Perspect. 22: 206-214. Gregersen, E. (1986). Human sexuality in cross-cultural perspective. In Byrne, D., and Kelly, K. (eds.), Alternative Approaches to the Study of Sexual Behavior, Erlbaum, Hillsdale, NJ. Huang, K., and Uba, L. (1992). Premarital sexual behavior among Chinese college students in the United States. Arch. Sex. Bekav. 21: 227-240. Lieh-Mak, F., and Ng, M. L. (1981). Ejaculatory incompetence in Chinese men. Am. J. Psychiat. 138: 685-686. Lynskey, M. T., and Fergusson, D. M. (1993). Sexual activity and contraceptive use amongst teenagers under the age of 15 years. New Zealand Med, J., 106: 511-514. Martinez, M. E., Salazar, R. M., Parada, R., and Cardoso, C. (1992). The sexuality of adolescents from 14 to 19 in a given population. Rev. Cubana Enferm, 8: 101-110. Ng, M. L., and Lau, M. P. (1990). Sexual attitudes in the Chinese. Arch. Sex. Behav. 19: 373-387. Odujinrin, O. M. (1991). Sexual activity, contraceptive practice and abortion among adolescents in Lagos, Nigeria. Int. J. Gynaecol. Obstet., 34: 361-366. Repossi, A., Araneda, J. M., Bustos, L., Puente, C., and Rojas, C. (1994). Sexual behavior and contraceptive practices among university students. Rev. Med. Chile, 122: 27-35. Ruan, F. F., and Matsumura, M. (1991). Sex in China: Studies in Sexology in Chinese Culture, Plenum Press, New York. Seidman, S. N., and Rieder, R. O. (1994). A review of sexual behavior in the United States. Am. J. Psychiat. 151: 330-341. Sonenstein, F. L., Pleck, J. H., and Kum L. C. (1991). Levels of sexual activity among adolescent males in the United States. Fam. Plann. Perspect. 23: 162-167. Tsang, A. K. (1988). Sexuality: The Chinese and the Judeo-Christian traditions in Hong Kong. Butt. Hong Kong Psychol. Soc. 20/21: 19-28. Tseng, W. S., Mo, K. M., Li, L. S., Chen, G., Ou, L. W., & Zheng, H. B. (1992). Koro epidemics in Guangdong, China: A questionnaire survey.J. Nerv. Mental Dis. 180:117-123. Tsoi, M. M., Ho, P. C., and Poon, R. S. (1984). Post-operative indicators and post-hysterectomy outcomes. Br. J. Clin. Psychol. 23: 151-152. Unwin, J. D. (1934). Sex and Culture, Oxford University Press, Oxford. Weisberg, E., North, P., and Buxton, M. (1992). Sexual activity and condom use in high school students. Med. J. Australia, 156: 612-613. Wen, J. K. (1995). Sexual beliefs and problems in Contemporary Taiwan. In Lin, T. Y., Tseng, W. S., and Yeh, E. K. (eds.), Chinese Societies and Mental Health, Oxford University Press, Hong Kong, pp. 219-230.

Archives of Sexual Behavior, Vol. 26. No. 1, 1997

BOOK REVIEWS Sex Errors of the Body and Related Syndromes. A Guide to Counseling Children, Adolescents, and Their Families (2nd ed.). By John Money. Paul H. Brookes Publishing Co., Baltimore, Maryland, 1994, 132 pp., $25.00. Reviewed by Heino F. L. Meyer-Bahlburg, Dr. rer. nat.1

This is the long-awaited second edition of the only book available on intersexuality and related categories of atypical genital development that is written for counselors, parents, and patients. In comparison to the original 1968 edition, the book's coverage has been expanded. It includes more conditions of genital disorders, has added a chapter on hormonal anomalies of puberty (chapter 10), and ventures into sexual problem areas, such as transsexualism and transvestism (chapter 11) and selected sexual dysfunctions (chapter 12). In addition, the text has been expanded and updated and the quality of graphs and photographs improved. The first chapter, "Verities and Variables of Sex," constitutes an introduction to the historical development of the field, including Money's own theoretical position on psychosexual differentiation. Chapters 2-9 concern anomalies of the sex chromosomes, gonads, fetal hormones, internal organs, external organs, hypothalamus (including a discussion of recent biological findings on homosexuality), and anomalies of assignment and rearing. These chapters cover descriptive material on each condition, provide the rationale for such condition-specific decisions as the sex of assignment and the timing of surgeries where called for, and point out counseling needs. A counseling guide (chapter 13) addresses some of the more broadly applicable counseling issues: hereditary transmission, infertility, chromosomal and gonadal status, sexual instruction, self-knowledge, and teenage autonomy and sexuality. Chapter 14 presents the parable technique, one of Money's specific interview techniques, designed to facilitate disclosure 1Department

of Psychiatry, New York State Psychiatric Institute, 722 W. 168th St., Unit 10, New York, New York 10032. 91 0004-0002/97/020&-0091S1Z5WO c 1997 Plenum Publishing Corporation

92

Book Reviews

of sensitive attitudes, thoughts, and behaviors. Chapter 15 is an exposition of Money's concept of the lovemap, i.e., an individual's mental schema of ideal partners, romance, and sex, and its development. This book is intended to be an educational text for the layperson. Thus, descriptions of medical issues are simplified and avoid technical jargon. The text is indeed highly readable. The downside is that descriptions are sometimes a bit superficial and that the detail given is not always satisfactory for highly educated patients. Money's interpretation of developmental mechanisms tends to be phrased as facts rather than hypotheses resting on a very limited empirical database. Although counseling lay people cannot easily avoid the problem of oversimplification, there is the risk that some may accept such statements as gospel and others attack them as easily refutable dogmatism. As is typical of Money's writings in general, there is also a lack of coverage of other sex researchers' psychological findings, and one may come away without any realization how controversial the whole area is. Future editions will hopefully include a discussion of the important role of peer counseling and syndrome-specific self-help groups that have sprung to life in recent years and are increasingly facilitated by the spread of the Internet. But these are minor criticisms in comparison to the uniquely rich educational and counseling material this book provides. Who should read this book? I personally do not recommend the entire book to patients with intersexuality or their parents. For their needs, the coverage is too broad and, therefore, confusing. Most of them find it difficult enough to absorb the medical and psychological information on a given specific condition, particularly in the beginning. This is especially so for those with lesser education. The many genital photos in this book add to this effect, although they are immensely useful for the counselor. On the other hand, I use some of the graphs regularly and individual pictures occasionally as counseling aids for my patients. For highly educated patients and parents, selected portions of the text might be complemented by additional syndrome-specific data from the literature which otherwise they are likely to garner, without guidance and discussion, from encyclopedias, literature searches by computer, and so on. For the most prevalent syndromes (Turner's, Klinefelter's, and congenital adrenal hyperplasia), there are now booklets available (Bock, 1993; Kirkland, 1990; Nielsen et al. 1991; Plumridge, 1987; Plumridge et al., 1982; Rieser and Underwood, 1989; Warne, 1989) that are specifically drawn up for such patients and their families. I like to combine Money's book as the basis for the professional counselor with recommendations of syndrome-specific booklets for the patients. In my view, the primary audience of Money's book are professionals. As intersex syndromes are rather uncommon, few clinicians (both medical

Book Reviews

93

and mental health staff) have had exposure to specialized training that prepares them to deal effectively with the psychosocial problems of such patients. The current takeover of clinical services by health maintenance organizations makes the development of professionals with some specialized knowledge even less likely. This book should be mandatory reading for any medical and/or mental health staff, especially trainees, who provide clinical services face-to-face and any form of counseling to intersex patients. (Nonmedical staff may find reading up on additional medical details necessary if they want to work in depth with a particular syndrome.) The book is also useful for anyone who is generally interested in a brief and highly readable introduction into the subject matter.

REFERENCES Bock, R. (1993). Understanding Klinefelter Syndrome: A guide for XXY males and their families, NIH Publication No. 93-3202, National Institutes of Health, National Institute of Child Health and Human Development, Bethesda, MD. Kirkland, R. T. (1990). Turner Syndrome, Office of Educational Resources, Texas Children's Hospital, Houston, TX. Nielsen, J., Naeraa, R., and Members of the Turner contact groups in Denmark (1991). Turner's Syndrome. Turner Contact Groups. An Orientation, 3rd rev. ed., The Turner Center, Aarhus, Denmark.. Plumridge, D. (1987). Good Things Come in Small Packages. The Whys and Hows of Tumer Syndrome, Rev. ed., Crippled Children's Division, The Oregon Health Sciences University, Portland. Plumridge, D., Barkost, C., End LaFranchi, S. (1982). Klinefelter Syndrome. The X-tra Special Bay, Crippled Children's Division, The Oregon Health Sciences University, Portland. Rieser, P. A., and Underwood, L. E. (1989). Turner Syndrome: A Guide for Families, Turner's Syndrome Society, Minnetonka, MN. Warne, G. L. (1989). Your child with congenital adrenal hyperplasia, Department of Endocrinology and Diabetes, Royal Children's Hospital, Parkville, Victoria, Australia.

The Development of Sex Differences and Similarities in Behavior. Edited by Marc Haug, Richard E. Whalen, Claude Aron, and Kathie L. Olsen. Kluwer Academic Publishers, Dordrecht, The Netherlands, 1993, 493 pp., $340.00. Reviews by Cheryl M. McCormick, Ph-D.,2,4 and Sandra F. Witelson, Ph.D.3 2Departmentof

Psychology, Bates College, Lewiston, Maine 04240. of Psychiatry, McMaster University, Hamilton, Ontario. 4To whom correspondence should be addressed. 3Department

94

Book Reviews

This volume is a collection of the major presentations from a NATO-sponsored workshop held in Gers, France, in July 1992. The chapters are written by an impressive array of researchers at the forefront of sex research. The general aim of the workshop was defined in the preface as the bringing together of biological and social scientists. This mix is not as balanced as one might want, at least in the book, in that the majority of the 28 chapters have a biological perspective. The "biological" and "social" chapters are in most cases too sophisticated to pit "nature versus nurture" or, in more current terminology, "essentialism versus constructionism." The chapter by Gladue offers a strong discussion of the limitations of these dichotomies. The role of social forces, experience, and environment are not ignored within the volume by the biological scientists. It is true that there is overwhelming evidence for the role of prenatal sex hormones in the process of sexual differentiation of the brain and behavior (in nonhuman animals, at least), and much of this evidence is reviewed in the book. However, the psychobiological model of the behavioral development of the sexes that emerges from the book is more sophisticated and more inclusive than the doctrine of sexual differentiation, outlined in earlier times by Jost (1953), that chromosomal sex determines gonadal sex which in turn determines phenotypic sex. Male and female phenotypes are not represented as ends of a continuum, but as orthogonal and multidimensional. Female development is no longer described as a passive or a "default" process: Evidence is accumulating that shows estrogens actively shape female brain development and behavior (see chapters by Dohler et al.; Arai et al.; and Olsen). Sex differences are often characterized as qualitative: For example, Williams and Meck's chapter indicates that male and female rats differ in how they use environmental cues for spatial navigation. The form and intensity of a behavior is viewed as best understood through consideration of both situational determinants and physiological constraints (e.g., see the chapters by Palanza et al.; Mos and Oliver; and Simon et al. on aggression in mice). The ecological niche and social organization of a species are explored as factors in the behavior of males and females: Keverne outlines how the opiate system may be involved in mediating social interactions which differ for males and females. The chapter by Gaulin indicates relationships among mating systems, territorial ranges, and sex differences in spatial ability through comparisons of monogamous and polygynous species of voles. Many of these ideas are not new to this volume, as indicated by some of the chapters: Whalen outlines some of the key studies since Phoenix et al.'s (1959) paper on the organizational effects of testosterone on female guinea pigs that have led to a change from a linear model of sexual dif-

Book Reviews

95

ferentiation to an orthogonal model in which masculine and feminine are separate dimensions (Whalen, 1974). In a chapter on the neural bases for sex differences in quail, Balthazart and Foidart reference several studies that indicate that the effects of a steroid on behavior are determined by factors such as condition of rearing, experience, age, and sex. The chapters of the book clearly illustrate that biological models of sex differences and similarities are complex and multivariate. Chapters such as those by Diamond and Green show that cross-cultural analyses and the complex variability in human psychosexual differentiation are in fact significant means by which psychobiologists can generate and test hypotheses about sex differences and similarities. This is important in a volume that aims to engage diverse disciplines, particularly given the admonitions in the chapters by social scientists against simplistic views of sex. Only three chapters represent social science perspectives (by Kail; Eagly; and Unger), in that they focus primarily on sociocultural determinants of sex differences in behavior. The three chapters differ in the degree to which the concept of sex (or gender), particularly as a dichotomized variable, is considered a useful or relevant variable for understanding behavior. Kail's chapter is a difficult read, mainly because so many methodological and conceptual issues related to sex as a variable in studies of language performance are addressed. She uses analysis of the linguistic use of gender across cultures to reinforce the hypothesis that sex is a social construct. Kail argues that the variable of sex has little or no predictive value for language performance and acquisition. On the other hand, she criticizes studies that have not included sex as a variable. Eagly provides evidence from meta-analytic studies to indicate that sex differences in human social behavior are more prevalent than suggested in the early comprehensive review by Maccoby and Jacklin (1974). Metaanalysis has been used increasingly as a tool by social scientists, and it would have been useful to have included some of the limitations of the technique given the varied audience for whom the book is intended. Eagly stresses the importance of gender role expectations and situational factors as moderators of sex differences in behavior. For example, the degree and direction to which a sex difference in helping behavior would be expected is said to depend on whether the situation required an assertive intervention, which would necessitate skills consistent with the stereotypic male gender role, or required more interpersonal activity, which would necessitate skills consistent with a stereotypic female gender role. Unger's chapter also stresses the importance of experience and situational factors as modifiers of behavior. She highlights the difficulty in drawing causal relationships between endocrinology and behavior in various species, including the human species, and provides a good critique of studies of hormonally/genetically

96

Book Reviews

atypical populations. The chapter provides a theoretical complement to the more empirical review of such populations from a psychobiological perspective in the chapter by Hines. Unger emphasizes that sex is a multidimensional, culturally-bound, system of categories, and uses transsexualism and alternative sexual categories from other cultures to illustrate these concepts. Much of the character of the book is undoubtedly related to it being a collection of the proceedings from a workshop. For example, the approaches and styles of the chapters are varied, with some focusing primarily on the exposition of research findings in a manner similar to a journal report, whereas others offer broad reviews of the research literature on a topic and place greater emphasis on theoretical issues. The plan underlying the order of the chapters is not obvious, nor was the book divided into sections with specific themes. Sections that included either an introduction to, or commentaries on, a set of papers would have provided the opportunity for integration of the diverse works. However, an index is provided which allows the reader to find related topics and issues among the chapters. The majority of the chapters are well-written and engaging. Unfortunately, the high number of typographical errors can be distracting. In sum, the book provides an excellent selection of psychobiological approaches to the study of the sexes: evolutionary (e.g., chapter by Petrinovich), neuroanatomical (e.g., chapters by Juraska and Yahr), neuropsychological (e.g., chapter by Benbow and Lubinski), physiological (e.g., chapter by Clemens et al.) and molecular (e.g., chapter by Baum). It is reminiscent of, and a welcome update to, the outstanding volume of Progress in Brain Research (DeVries, 1984) devoted to sexual differentiation of the brain. As such, the book will appeal primarily to researchers in the neurosciences. Although alternative positions represent only a small fraction of the included works, the chapters by the social scientists raise important issues for, and provide valuable critiques of, studies of sex differences in people. These chapters underscore the need for multiple levels of analysis in the study of sex, and stress that human cultural and linguistic systems need to be fully incorporated into models of the development of the sexes. This volume suggests that biological scientists agree.

REFERENCES DeVries, G. J. (ed.) (1984). Progress in Brain Research, Vol. 61: Sex Differences in the Brain,

Elsevier, Amsterdam, Jost, A. (1953). Problems of fetal endocrinology: The gonadal and hypophyseal hormones. Recent Prog. Horm. Res. 8: 379-418.

Book Reviews

97

Maccoby, E. E., and Jacklin, C. N. (1974). The Psychology of Sex Differences, Stanford University Press, Stanford, CA. Whalen, R. E. (1974). Sexual differentiation: Models, methods, and mechanisms. In Friedman, R. C., Richart, R. M., and Van de Wiele, R. L. (eds.), Sac Differences in Behavior, Wiley, New York.

Biographies of Gender and Hermaphroditism in Paired Comparisons. Clinical Supplement to the Handbook of Sexology (Series Editors, J. Money and H. Musaph). By John Money. Elsevier, Amsterdam, 1991, 375 pp., $60.00 (softcover), $217.25 (hardcover). Reviewed by Heino F. L. Meyer-Bahlburg, Dr. rer. nat.5

Intersexuality is one of the orphan domains of human behavioral research. Everyone is fascinated by the phenomenon, but very few have the opportunity to specialize sufficiently to generate systematic, useful, and informative data. The reasons are the rarity of the syndromes, their biological diversity, and the medical knowledge required for interdisciplinary collaboration. Much of what is known about the psychological development of individuals born with an intersex condition comes from John Money's extensive work in this area. Money is seen by many as the father of the psychoendocrinology of sex and gender, much as Lawson Wilkins, who originally brought Money to Johns Hopkins Hospital in Baltimore, was the father of pediatric endocrinology. There is no one anywhere in the world who has such a command of medical information on the many and diverse syndromes of intersexuality, so comprehensive a knowledge of their psychological development—based on the systematic and detailed collection of behavioral data over a span of up to over 40 years—and so much experience with principles and techniques of counseling for these conditions, most of which he himself developed. And there is no one else who has had such a strong influence on our conceptual understanding of the development of gender and sexuality in intersex conditions and on the policy governing psychosocially relevant decision making. This applies especially to the policy of gender assignment; it was the influence of the Hopkins school, and in particular Money's leadership, that unshackled the assignment decisions from the dictates of simplistic biological determinism. 5Department of

Psychiatry, New York State Psychiatric Institute, 722 West 168th St., Unit 10, New York, New York 10032.

98

Book Reviews

It does not diminish Money's crucial role in the development of this field if one acknowledges that the evolution of our understanding of psychological development of intersexuality and the formation of policy in this area is by no means completed— if not alone by virtue of the fact that the definition of gender role itself undergoes societal change which, in turn, is likely to influence the ontogenesis of gender-role behavior and identity, especially in those who do not fit the mold of the majority. Also, as additional psychological data accumulate on the various syndromes and new medical techniques are developed, both medical and psychosocial management is likely to become more syndrome-specific. The book's first chapter is introductory and addresses three broad issues: the logic of the presentation of case reports in the form of matched pairs, the assessment methodology Money uses for these clinical case studies, and the exposition of Money's "exigency theory"—somewhat cumbersome to digest because of his idiosyncratic anglocentric coinage of novel terms—as a solution to the limits of ideological thinking on the one hand and reductionist thinking on the other in this area of research. The interview method Money uses in the assessment of his cases is uncommon in clinical research. It is neither a fully structured or semistructured interview of the kind that, in recent decades, has become so important in psychiatric/psychologic research, nor is it an unstructured clinical interview. It may best be described as a combination of a biographical and qualitative interview administered in a clinical setting, covering a standard set of topics, and focusing on the interviewee's experiences and behaviors without attempting dynamic/analytic interpretations during the interview process. It uses a number of discrete techniques such as beginning a topic with open-ended questions and progressing to more structured probes, the "sportscaster technique" of concrete detailed reports of events, or the "parable technique" and "Catch-22 query" to facilitate disclosure of sensitive material. An extraordinary feature of the case reports is the fact that most of the interviewees have been followed over many years, often from childhood to adulthood, with interviews conducted at various stages in their development. These interviews are conducted in a medical setting, and are usually followed by or interspersed with clinical counseling. The assessment interviews are tape-recorded in full and transcribed, although the interviewer may ask the patient to summarize certain aspects of the interview in order to reduce transcription time and costs. As in anthropologic/ethnographic research in general, Money makes use of additional informants and also of other material, such as letters by patients and parents, school records, and hospital charts. All these materials combined make up the "consolidated, developmental history" which constitutes the

Book Reviews

99

database for outcome research. The resulting text material can be summarized for case reports as a biographical narrative, enriched by extensive direct quotations, as in this volume, or can serve trained raters in group studies as the basis for systematic ratings suitable for statistical analysis, as employed in many of Money's journal publications. The case reports include some of the counseling material as well; it provides very informative illustrations of the counseling procedures Money has been using. Money does not address the technical issues of interview-based material: reliability, respondent bias, retrospective reconstruction, memory problems, potential risks of the summarizing technique, and suggestions introduced by the parable technique. He acknowledges that his methodology is not as rigorous as recommended by many methodologists for hypothesis-testing studies, but he is aware of the fact that many of the systematic interview and questionnaire "instruments" in use have very limited validity and even less theoretical and/or clinical utility. He sees his case histories as predominantly exploratory in nature, quite appropriate for the current status of research in this field. Chapters 2-12 contain the case reports. The cases are presented as matched pairs, i.e., usually as pairs that are concordant with respect to some salient medical and/or psychological characteristics and discordant in others. After a "synopsis" of the matched pairs, each chapter describes in detail for each person the diagnostic and clinical biography, gender-coded social biography, and lovemap biography. It closes with an "exposition" in which Money discusses the specific biological and psychosocial mechanisms that may have affected the development of the individuals described. From his extraordinary collection of case histories, Money was able to contrast an astonishing diversity of cases: in chapter 2, two 46,XY cases of complete androgen insensitivity syndrome (cAIS), discordant in their sex of rearing; in chapter 3, two female patients lacking a vagina, one 46,XY with cAIS, the other 46.XX with vaginal atresia; in chapter 4, a 46,XY patient with gonadal dysgenesis versus a 46,XY patient with hypogonadism secondary to an LH abnormality, both raised female; in chapter 5, a set of two pairs of undermasculinized 46,XY persons, all assigned to the female gender at birth, but two reassigned to the male gender in infancy, with one of the boys starting to feminize and one of the girls to masculinize during puberty, in both cases leading to change of gender; in chapter 6, two 46,XY siblings assigned to the female gender one of whom changed gender in adulthood; in chapter 7, two 46.XY cases born with micropenis, one assigned and reared male but struggling with considerable gender ambivalence during his development, the other early reassigned to the female gender and remaining so without apparent gender problems; in chapter 8, two patients with 45,X/46,XY, one reared female, the other male; in chap-

100

Book Reviews

ter 9, two 46,XX cases with classical congenital adrenal hyperplasia (CAH), one assigned female at birth, reassigned male in infancy, and re-reassigned female in adolescence, the other with the exactly reverse three-step gender assignment history; in chapter 10, two patients with CAH, one female, the other male; in chapter 11, a male 47,XXY patient versus a male 47,XYY patient. In the final chapter, Money describes the lovemap development in two physically normal 46,XY males, with the intent of illustrating "the sexuoerotic difficulties and pitfalls that confront children" (p. 344) growing up in a sex-negative society, even in the absence of genital ambiguity, and thus presenting a "norm against which to compare the other cases in this book" (p. 344). Intersexuality involves the study of rare syndromes. Thus, case studies have been characteristic of the beginnings of this field, both in the medical and the psychological area, and significant progress may still take this form as recently exemplified by single-case reports of the new endocrine syndromes of aromatase deficiency and of estrogen receptor insensitivity (Ito et al., 1993; Morishima et al, 1995; Shozu et al., 1991; Smith et al, 1994). Most early case studies of a new syndrome are, in Yin's (1994) terminology, exploratory and descriptive. Subsequent pooling of individually published rare cases allows a preliminary characterization of the phenotypic spectrum of a given syndrome, until epidemiological studies of the condition reveal the full spectrum. The cases presented by Money in this volume go beyond exploratory and descriptive purposes. They have been deliberately combined in the form of contrasting pairs to force upon the reader the realization how variable the behavioral phenotypes are that may be associated with a given genotype or endocrinotype, and vice-versa. In reading these contrasting pairs of cases, it becomes obvious how difficult it is to predict at birth for individual cases the long-term outcome in terms of gender and sexual orientation and how much there is an interaction between the developmental effects of the genital status and the social environment. In the most extreme situation, case studies of this sort can be used to invalidate existing simplistic hypotheses of gender development such as that the presence of a female or male karyotype dictates the development of gender (and, therefore, ought to be the final criterion for gender assignment). The most clear-cut example of a syndrome that provides "a counterinstance for notions that are considered to be universally applicable" (Kazdin, 1992, p. 154) is cAIS in genetic (46,XY) males. Given the inconspicuously feminine appearance of their external genitalia at birth, such individuals are usually assigned to the female gender (since a genital anomaly is usually not expected) and elect to stay female even after the

Book Reviews

101

intersex condition is diagnosed, which often does not occur before puberty or even later. While it is relatively easy to dismiss one-factor explanations of psychosexual differentiation, it is much more difficult to identify positively and definitively among the many biological and psychosocial factors those that do contribute, and to delineate how they do so in the context of the other influences. Collections of case studies such as these suggest answers, but cannot transcend considerations of plausibility and compatibility with theoretical expectations. The selection of cases to make a specific theoretical point, and even more so the selection of contrasting pairs of cases, is, of course, anything but a random selection procedure and, therefore, unlikely to yield samples approximating representativeness. With regard to their theoretical purpose, the presentation of multiple cases, or of multiple pairs of cases, draws on replication logic (Yin, 1994, pp. 45-50), not on the logic of statistical generalization. Of course, the utility of case reports of intersex patients goes beyond the particular theoretical points Money wants to make. They make us realize how a person experiences his/her intersex condition in the context of everyday life as it unfolds through the various stages of development, how the parents respond to it, and how medical management influences the outcome. All the more, we need to keep in mind that unrepresentativeness is a common characteristic of case reports. More often than not, it is the "interesting" cases that are reported, i.e., those that are unusual, that lend themselves to telling a fascinating story, and, as part of it, those who have stayed long enough with the clinician to provide all the material. Individual case reports demonstrate what may occur, but not necessarily what does occur typically or on average. Such biases are also evident in Money's case selection. For instance, Money describes three cases of 46,XX genetic females with classical CAH of whom two change gender in adolescence, a ratio unmatched in any of the existing group studies of this syndrome, Money's own work included. This observation does not invalidate the purpose of Money's selection and the utility of the book. It only points out the necessity to complement collections of case studies of this sort by group studies that give a more representative picture of the spectrum of behavioral phenotypes associated with a given genotype or endocrinotype, as Money himself has done in many others of his large number of publications. It is a pity that this book ended up with the high-priced publisher Elsevier. Hopefully, some day it will be available in a less expensive edition that students can afford. The book should be must reading for anyone who has an interest in the development of human sex and gender, regardless

102

Book Reviews

of their theoretical preconceptions, especially for the many students of the behavioral sciences who rarely or never have an opportunity to get to know an individual with an intersex condition. The reader will be introduced to the medical aspects and the theoretical implications and, at the same time, gain insights into the life issues people with intersex conditions are confronted with. The text is very well written, even the case discussions are highly readable, and the biographical sections are captivating. REFERENCES Kazdin, A. E. (1992). Research Design in Clinical Psychology, 2nd ed., Allyn and Bacon, Boston, MA. Ito, Y., Fisher, C. R., Conte, F. A., Grumbach, M. M., and Simpson, E. R. (1993). Molecular basis of aromatase deficiency in an adult female with sexual infantilism and polycystic ovaries. Proc. Natl. Acad. Sci. (U.S.) 90: 11673-11677. Morishima, A., Grumbach, M. M., Simpson, E. E., Fisher, C, and Qin, K. (1995). Aromatase deficiency in male and female siblings caused by a novel mutation and the physiological role of estrogens. /. Clin. Endocrinol. Metab. 80: 3689-3698. Shozu, M., Akasofu, K., Harada, T., and Kubota, Y. (1991). A new cause of female pseudohermaphroditism: Placental aromatase deficiency. J. Clin. Endocrinol. Metab. 72: 560-566. Smith, E. P., Boyd, J., Frank, G. R., Takahashi, H., Cohen, R. M., Specker, B., Williams, T. C., Lubahn, D. B., and Korach, K. S. (1994). Estrogen resistance caused by a mutation in the estrogen-receptor gene in a man. New Engl. J. Med. 331: 1056-1061. Yin, R. K. (1994). Case Study Research: Design and Methods. 2nd ed.. Sage, Thousand Oaks. CA.