copy, Dec. 1996. Published as "Chapter 12" in C. Bagley & R. Ramsay,
Eds. (1997) Suicidal Behaviours in adolescent and adults: taxonomy,
understanding and prevention. Brookfield, Vermont: Avebury. A shorter
version of this paper was published as Suicidal Behaviors in Homosexual
and Bisexual Males in Crisis, The International Journal of Suicide
and Crisis Studies, Vol. 18(1), 1997, pp. 24-34 (PubMed "abstract"
link with a document delivery service. The abstract is also available on
webpage.). The results of this study, combined with the another
major demographic study also published in 1998, and additional
information have been used to produce a monumentally important CAVEAT
ALERT! In 2000, a CDC-Based study - from an NHANES sample
taken between 1988 and 1992 - produced a subsample of young homosexually
oriented young adult males (age = 17 to 29 years) comparable in age range
of the sample taken for this study (age = 18 to 27 years, sample taken
in 1991) with "suicide attempt result - compared to same-age 'heterosexual'
males - which replicated the results of this study. See Addendum.
ABSTRACT: A stratified random sample of 750 males aged 18 to 27 in Calgary, Canada included questions on sexual activity and orientation. Mental health questions included a measure of suicidality, and of acts of deliberate self-harm. A computerized response format (established as a good method for eliciting sensitive personal data) ensured anonymity. Almost 13 percent of males were classified homosexual or bisexual on the basis of being currently homosexually active and/or by self-identification. Significantly higher rates of past suicidal ideas and action were reported by homosexually oriented males compared to heterosexual males. These findings, indicating that homosexual and bisexual males are 13.9 times more at risk for a serious suicide attempt, are consonant with previous findings. Homosexually oriented males accounted for 62.5 percent of suicide attempters. We speculate that the predominant reason for the suicidality of these young males may be linked to the "coming out" process occurring in a highly homophobic society. Related issues are resolved for some, but may not be for others having high levels of current depression. These results underscore the need for qualified services rarely available to homosexually oriented youth.
In addition to these studies, the Bell & Weinberg (1978) study for the Kinsey Institute analysed a 1969 volunteer sample of 575 white predominantly homosexual San Francisco males  with a mean age of 37 years, reflecting the youth situation existing between 1930 to 1969. In the Kinsey Institute study, results were compared with those from a stratified random sample of predominantly heterosexual males taken in the same area, controlling for age, education, and occupational level. By the age of 20 years, homosexual males were shown to have been 13.6 times at greater risk for a suicide attempt, based on suicide attempt rates of 9.6 and 0.7 percent for homosexual and heterosexual males, respectively.
From the various North American studies, the lifetime attempted suicide rate for adolescent homosexual males appears to have tripled (9.6% to 31.3%) from about 1950 to 1990, which is similar to the two- to four-fold increases in the suicide rate for Canadian and American adolescents and youth during the same period (Berman & Jobes, 1995; Lipschitz, 1995; Health Canada, 1994). Compared to a recently reported lifetime attempted suicide rate of 3.2 percent for male adolescents , gay and bisexual male adolescents associated with North American gay communities are about 10-times more at risk (31.3% vs. 3.2%) for a suicide attempt than adolescent males not differentiated on the basis of sexual orientation.
The probability that homosexually oriented youth are at high risk for suicidal ideas and behaviors (or 'suicidality' as we term it) has occasionally been noted in the media, with requests for study and intervention reaching the U.S. Congress. This led to a 1994 workshop followed by a special issue of Suicide & Life-Threatening Behaviour (Moscicki, Muehrer, & Potter, 1995). The major emphasis of these papers was on the biased nature of all samples indicating high suicidality levels for homosexually oriented people, apparently limiting the reliable conclusions which could be made (Moscicki et al, 1995; Moscicki, 1995; Muehrer, 1995). A more constructive analysis of the earlier studies showing greatly elevated rates of suicidality in homosexually oriented males is offered by Tremblay (1995).
Despite suggestive evidence from community studies, homosexuality issues have been largely ignored in suicide intervention and prevention programmes, and in youth suicidality research (Remafedi, 1994; Tremblay, 1995), reflecting a phenomenon which exists in almost all fields where youth problems have been studied (Herdt, 1989a). This systematic ignorance also applies to the training of mental health professionals (Murphy, 1992; Steel & Gyldner, 1993). Resistance to addressing youth homosexuality issues is reflected in responses to relevant papers published in many fields: "Unfortunately and tragically, few have listened" (Savin-Williams, 1994. p. 266).
Homophobia has been a serious social problem in North America , Stigmatization of gay men and women by most members of the US public is reflected in the behaviors of some researchers, educators and therapists (Harbeck, 1992a; Steel and Gyldner, 1993). Homophobia may a major reason explaining why the methodologically excellent study of Bell & Weinberg (1978) has been largely ignored by suicidologists, and why no attempt has been made to replicate (or refute) the finding that young homosexual males are at much greater risk for experiencing suicidal crises.
The aims of this secondary analysis of the data set are:
(a) to determine whether self-harm and suicidal attempts vary significantly on the basis of sexual orientation;
(b) to see whether the earlier results of Bell and Weinberg (1978) would be replicated in a random sample of a community notselected on the basis of sexual orientation.
(c) To integrate our findings with those of other studies to produce a clearer perspective of the North American youth suicidality problem.
For purposes of this analysis, data in our study are given for several categories of males, based on being currently sexually active (had heterosexual and/or homosexual sexual contacts in the last six months) and being currently celibate (Table 12.1). Celibate males are divided in two categories: predominantly homosexual and predominantly heterosexual males, based on self-identification. Presenting the data for these categories permits key comparisons to be made; our grouping approximates the categorization of males studied by Bell & Weinberg (1978).
Bell & Weinberg (1978) studied two groups of predominantly homosexual and heterosexual males, defined on the basis of the Kinsey (1948) seven point scale (0 to 6) from exclusive heterosexuality (0) to exclusive homosexuality (6). Males in the white homosexual sample had a mean age of 37 years and 'almost three-quarters ... considered themselves exclusively homosexual (or 6 on the Kinsey Scale) in their current sexual behaviour.' (p. 54). Nearly 90 percent of homosexual respondents were rated five or six on the combined behaviour and feelings scale. About three-quarters of their heterosexual male sample were rated "0" on the same scale.
The mean age of our sample is 22.7 years, and the group best approximating the Bell & Weinberg sample of predominantly homosexual males includes males currently and exclusively homosexually active (in the past 6 months) classified as homosexual (4.3%: 32/750); celibate self-identified homosexual males (1.7%: 13/750) who had been homosexually active in the past (a few also self-identifying as bisexual); and bisexual males (4.9%: 37/750) currently sexually active with both genders. This grouping produces a total of 10.9 percent of males (82/750) in the homosexual category, 55 percent (45/82) being more or less exclusively homosexual, compared to 75 percent in Bell & Weinberg's homosexual male sample. This difference exists (and must existfor the two samples to be truly comparable) for reasons related to the difference between the mean age for our sample and their sample (22.7 vs. 37 years).
Binson, Michaels, Stall, Coates, Gagnon & Catania (1995) report that, for homosexually oriented males in the age groups 18-29, 30-39, and 40-49 years, living in the larger American cities, the proportion of homosexually oriented males in each age category was similar (6.4%, 7.1%, and 5.1%, respectively); but the percentage ratios of exclusively homosexually active males to those having sexual contacts with females (bisexual males) changes in the direction of exclusive homosexuality with increasing age: 51.3 : 48.7 (18-29 years); 78.7 : 21.3 (30-39 years); and 81.0 : 19.0 (40-49 years). Our sample of males, with a mean age of 22.7 and 55 percent of homosexually active males categorized as bisexually active males, is therefore expected to change so that, by a mean age of about 37 years, a higher proportion would be exclusively homosexually active. Bisexuality as a transition is documented (Gochros, 1989; Klein & Wolf, 1985), and it begins before the age of 30 (McKirnan, Stokes, Doll, & Burzette 1995).
1. Generally young families living in detached or row-houses, basically middle-class;
2. Generally unmarried young adults living in apartments, usually with white collar jobs, basically middle- to lower middle-class;
3. Generally adults who are single, married or cohabiting, living in low rent and public housing.
Two neighbourhoods (defined by postcodes related to the census data used) within each of the three strata were randomly selected. The sampling area was restricted to the northern half of Calgary, but we do know that the regions sampled were representative in demographic profiles of the types of neighbourhoods in the southern half of the city. Respondents were then randomly sampled within six neighbourhoods using the reverse telephone directory. An initial telephone call established whether anyone in the required 18 to 27 year age group was resident; then a request was made for a personal interview in the respondent's home for a study of 'childhood events, current adjustment and outlook on life.'
Respondents were paid $20 for participation, regardless of whether or not they completed the computerized questionnaire, the interviews occurring from 1991 to 1992. Sexual issues were not initially mentioned, being only asked at the very end of the computerized questionnaire, since we assumed that some types of mental health responses might be triggered if these questions followed those on sexuality. Sampling continued until approximately 75 individuals in each of the ten age groups 18 to 27 were obtained. Sixteen individuals did not complete the computerized questionnaire, and of those in the requisite age groups 72.9% both agreed to participate, and completed the computerized questionnaire, which took between 40 minutes and 1.5 hours, in the respondent's home setting.
Within each neighbourhood we interviewed between 31 and 52 percent of the total population of males aged 18 to 27 identified in the 1991 Federal Census. Sampling continued until 750 completed questionnaires were obtained. Substituting newly sampled individuals for those declining an interview could have been a source of bias. However, the 27% of the sample who were recruited as 'substitutes' did not differ on any of the demographic, childhood and current behavioral and emotional profiles from the remainder of the sample. The districts sampled did not include Calgary's city centre where the highest concentration of homosexually oriented males is located and is referred to as 'the gay community area.'
Respondents were advised that no record of identity would be made, and that the responses were completely anonymous. After an initial tutorial in using the portable computer, questions appeared on screen, and the computer's track-ball was used to key in the chosen response. Data for each respondent was stored in a random block, so that the order in which questionnaire data was saved on the hard drive did not allow the researchers to identify any individual.
The research assistants introducing the computerized questionnaire were male, casually dressed, in the same age range as the respondents. While the respondent completed the interactive program, the research assistant read a book, or watched TV, and declined (because of the ethical requirement that researchers should be blind to individual responses) to answer any specific questions about the response system. This was emphasized during the tutorial to increase the subjects' confidence that the information to be given was truly on an anonymous basis.
The measure of "Current sexual interests and activities" consisted of selected items on heterosexual and homosexual activity (voluntary and non-voluntary) at various ages, devised by Langevin (1985), and included questions about whether the sexual activity was wanted or unwanted, taken from a national Canadian survey (Bagley, 1989). Pilot testing of the computerized response format in comparison with a conventional pencil-and-paper format indicated that the computerized questionnaire yielded a higher proportion of young males who responded positively to questions on the sensitive theme of sexual abuse in childhood (Bagley & Genuis, 1991).
Using this method, one percent of men reported having had sex with girls and/or boys less than age 12 years, and four to five percent reported having desires of this kind (Bagley et al. 1994). Open acknowledgement of such highly taboo sexual information would not be expected in telephone or face-to-face interviews, thus making the computerized method probably the best available to date for soliciting other guarded information, including often threatening facts related to one's homosexual nature, given that North American societies are still highly homophobic .
The homosexual contacts reported in this paper were, according to the respondents, all voluntary, regardless of the age (12 years and over) at which these contacts occurred (non-voluntary contacts are discussed in Bagley et al., 1994). The age of 12 was chosen because at that age Canadian law allows a young person to exercise some degree of free choice in sexual relationships (Wells, 1989). The measures included frequency of sexual contact with someone of either sex, and responses to the question: "Do you consider yourself to be Heterosexual (Yes/No) ... Homosexual (Yes/No) ... Bisexual (Yes/No)?
Various mental health measures were employed: those reported here are those measuring suicidal ideas (in the past 6 months), suicidal behaviour (in lifetime) and depression (in the past two weeks). These measures are reported in this further analysis because of the known high incidence of suicidal ideas and behaviour in adolescents and young men struggling with homosexual identity issues, within an oppressive and stigmatizing climate of homophobia (Martin, 1982; Schneider, 1988, Herdt, 1989; Harbeck, 1992; Blumenfeld, 1992; Savin Williams, 1994, Unks, 1995).
The Suicidal Ideas & Behaviour scale was scored: 0, no thoughts or behaviours of this kind in past 6 months; 1, any thoughts or ideas about suicide in past 6 months; 2, plans for a suicide attempt in past 6 months; 3, any act of deliberate self-harm in lifetime; 4, any act of intentional (but failed) self-killing in lifetime. The face validity of this scale in terms of its psychosocial correlates was established by Bagley & Ramsay (1985) and Bagley & Ramsay (1993). Respondents also completed the Centre for Epidemiological Studies Depression scale (Radloff, 1977). This scale is well-validated, and is widely used in epidemiological studies (Roberts & Vernon, 1983; Zimmerman & Coryell, 1994). A score of 19 or more on the scale indicates a degree of psychological distress which might benefit from counselling.
The figures on suicidality and suicidal behaviours (Table 12.1) confirm earlier findings that homosexual and bisexual males have a higher incidence of suicidal thoughts and actions than heterosexual males. Celibate heterosexual men also have a high score on the suicidality index and on the measure of current depression, but no reported serious suicide attempts. The most at risk group in terms of actual suicidal behaviours, suicidal ideation in the past six months, and depression in the past two weeks, are celibate self-identifying homosexual males. They had the highest proportion of individuals in the "self-harm" category (46.1%: 6/13), followed by celibate heterosexual males (17.7%: 22/124), compared with 10.8% (4/37), 9.4% (3/32), and 2.8% (15/544) rates, respectively, for their sexually active bisexual, homosexual, and heterosexual counterparts.
Within the sexually active male groups, males classified homosexual and bisexual were nearly three times more likely (risk ratio 2.94 : 1) to have engaged in self-harm at some point in their lifetime, than heterosexual males. Homosexual and bisexual men are six (3.1% vs. 0.5%) to eleven times (5.4% vs. 0.5%) more likely to have made a life-threatening suicide attempt than heterosexual males. Celibate homosexual men had the highest serious suicide attempt rate (2/13: 15.5%). For the 10.9 percent of males classified as homosexually oriented (currently homosexually active males, and celibate homosexual males), the risk ratio for a life-threatening suicide attempt was 13.86 : 1; that is, these males were almost fourteen times (5/82: 6.1% vs. 3/688: 0 .44%) more likely to have made a serious suicide attempt at some point in their lives than their heterosexually oriented counterparts. They also accounted for 62.5% (5/8) of the serious suicide attempters (X2 = 17.69 p less than .001, df = 1).
Introduction / Study Results • Discussion • End Notes, Tables • Bibliography