TO: pjtremblay@hotmail.comTO:
To the Notes, Tables, Bibliography - On The prevalence of Homosexuality and Bisexuality in a Ramdom Community Survey....
To: Index Page for Bisexual, Gay, Queer Male Suicidality
Last Page! - On The prevalence of Homosexuality and Bisexuality in a Ramdom Community Survey....
At A Glance! The result of more than 30 studies in 4 categories.  
A GLBT Internet-Based Education with InfoSearch Pages
Homosexual and Bisexual Orientation 
In Young Adult Calgary Males:
A Demographic Study

Addendum 2000


The Bagley and Tremblay (1998) results indicate significant  "homosexuality" underestimates made in many demographic studies: from 200 to 800 percent for homosexual behavior and self-identification, depending on the type of study methodologies being compared.

A part of the Bagley and Tremblay (1998) demographic results are replicated in the CARDIA Cohort Study by Krieger and Sidney (1997).

The "homosexuality" underestimating magnitude suggested by Bagley and Tremblay (1998) for previously flawed study - about 400%, on average - is confirmed in the Turner at al.(1998) "homosexual behavior" study of young males ranging in age from 16 to 19 years. The web page - Caveat Alert! - was written on the basis of these study results.

One study reports that the wording used in a questionnaire soliciting homo-sexual behavior information can result in underestimates in the 400% range.

Demographic studies based on 'sexual oreintation / homosexual behavior' using flawed methodologies continue to be published.

Random Telephone Surveys In North America: 3 to 5 percent of the adult population acknowledging that they identify as gay, lesbian, or bisexual, this being an underestimate. Repeatedly, the representation of people of colour in this subpopulation has been about twice their representation in the mainstream population.

References: - Includes specific information related to some studies.

Bagley and Tremblay (1998) stated:

Demographers must begin to address seriously the issue of flawed methodology (in terms of populations sampled, the methods of asking questions, and the actual questions asked) in arriving at an estimate of the proportion of adult men who are homosexual. For example, if the homosexual population is undercounted, then the rate of HIV in such populations will be greatly overestimated. Studies producing 200 to 800 percent underestimates for male homosexual realities in our society should be unacceptable as social science research.
On the basis of our study's results, we would argue that it is unwise for demographers to seek data on sexual orientation using either telephone or face-to-face interviews. If face-to-face interviews are to be conducted, they should be carried out as part of ethnographic studies, as exemplified by the excellent work of the Kinsey Institute. Many homosexuals in our society remain hidden, and conventional survey techniques are unlikely to elicit frank responses about questions of sexuality.

Our computerized response format in which questions about taboo sexual activities (sexual contacts with minors) were answered with some frequency does show that less stigmatized behaviors, such as voluntary, adolescent and adult same-gender sexual contacts are likely to be revealed more frankly than in other methods.

Recent researchers seem to have ignored the realities of being homosexual in a homophobic culture [1]: thus Billy et al. (1993), Michael et al. (1994), and Binson et al. (1995) appear to have failed with regard to one of the "overarching principles" listed by a group of experts who were examining homosexuality- related issues: "Researchers must know the community they are researching." (Working Group, 1995)

Bagley and Tremblay (1998) were referring to a range of likely underestimates for male homosexual / bisexual realities ranging from the reporting of same-gender sexual activity by study subjects to the acknowledgement that one identifies as homosexual or bisexual. Two "homosexuality" demographic results of the Cardia Cohort study of young adult American males are similar to the Bagley and Tremblay (1998) results for a Calgary sample of similar aged males. Respectively, 15.3 and 14 percent of the males in these samples reported having had at least one same-sex partner, and 5.2 and 5.6 percent of the males may be deemed homosexual or predominantly homosexual because they reported having had only same-sex partners in the Cardia Cohort study, or they reported "homosexual" self-identification in the Calgary Study.

By 1998, the Turner et al. study of a large ramdom sample of 16- to 19- year-old males reported that, on average, a 400% underestimate occurred for adolescent males reporting various forms of same-gender sexual activity if the data is solicited via pencil-and-paper questionnaires as opposed to using computer technology (questions on screen, with audio via headphones. See Caveat Alert!). The results were in the range (200% to 800% underestimates) postulated by Bagley and Tremblay (1998) for studies using methodologies other than computer technology, and especially for face-to-face interviews. Infortunately, face-to-face demographic studies based on sexual orientation continue to be done as illustrated by the Cochran and Mays (2000) study of a NHANES sample of 17- to 39-year-old males which reported on the higher risk for lifetime suicide attempts by males reporting having engaged in same-gender sexual activity. [Summary in the Addendum to the Bagley and Tremblay (1997) suicidality paper.]

Only 2.2 percent of males (aged 17- to 39-years-old) in the Cockran and Mays (2000) study acknowledged having "any male sex partners in their lifetime" (p. 575), and similar methodologically flawed studies were then cited (given range of study results: "2% to 7%") to conclude that "this is consistent with the prevalence observed in NHANES III" (p. 577), thus creating the illusion of validity for their demographic results. Not mentioned is that this range of demographic results suggests a possible 350 percent underestimating error for lowest results compared to the highest cited result. Instead, it is asserted that "the willingness of men to report same-sex partners in a population-based survey such as NHANES is unknown; thus, the extent to which homosexually experienced men... declared no male sex partners cannot be determined" (p. 577). This assertion, however, is only made possible by not citing studies (e.g. Bagley and Tremblay, 1998; Turner et al., 1998) indicating the likelihood of producing significant underestimating errors when highly flawed methodology is used.

Bagley and Tremblay (1998) stated that "Studies producing 200 to 800 percent underestimates for male homosexual realities should be unacceptable as social science research" (Journal of Homosexuality, 36(2), p. 14), and the Turner et al. (1998) study confirmed what was indicated by the Bagley & Tremblay (1998) results. From the abstract: "As a group, however, there was a "fourfold" increase (400%) in males [age 15 to 19] reporting homosexual activity when the computer-audio system was used: from a 1.5% to 5.5% positive response rate.."  The implications of these results are monumentally important, especially with respect to the results of all studies comparing attributes (such as suicidality) of youth identified by their reported (acknowledged or admitted) "homosexual" attributes."

The Sedlacek and Kim (1996) study result for college students (not peer reviewed, but published) serves as a warning to all researchers seeking "homosexuality" information in mainstream population study samples.

"This study examined whether terminology used in surveys could affect respondents' answers... In the true-false statement, "I have had at least one homosexual experience during the last year," only 2.5% of the respondents marked it true. However, for the statement, "I have had at least one sexual experience with someone of my gender during the last year," 10.4% of respondents indicated this as true."
This result implies than "language used" on a questionnaire can easily result in a 400% underestimate in young adults reporting same-gender sexual activity in the past year. The "10.4%" Sedlacek and Kim (1996) study result for college students reporting same-gender sex partner(s) in the past year is also consonant with the Bagley and Tremblay (1998) report that 9.2% of young adult males in Calgary's maintream population reported having at least one same-gender sexual partner in the past 6 months.

Asking individuals if they are gay, lesbian, or bisexual in a world where such "labels" may be disliked - or these "labels" are NOT felt to apply to oneself - can create 'homosexuality" underestimating problems. In a survey of 721 students in four large human sexuality classes at California State University, Fullerton (287 men and 434 women, average age = 22 years, 42% White, 22% Hispanic, 21% Asian, and 15% "Others"), 2% of males (6/272) self-labeled as "gay," and 3% (8/272) as "bisexual." For women, 1% (5/412) self-labeled as "lesbian" and 3% (8/412) as bisexual. However, they were also asked about the "sexual attractions" in the following manner. Males were to respond to the following statement "I am sexually attracted to men" using a 7-point Likert Scale from "Strongly Disagree" to "Strongly Agree." That is:

(1) Strongly Disagree,
(2)  Disagree, (3)  Slightly Disagree,
(4)  Neither Agree nor Disagree,
(5) Slightly Agree (6) Agree (7) Strongly Agree.

There was 10% (29/284) of males who responded with "neither agree nor disagree" to "strongly agree" (#4 to #7) to the statement. For females, the statement was "I am sexually attracted to women," and 12% (51/429) responded in the "#4 to #7" categories. Homosexuality, however, is still strongly taboo in many cultures and it is known that some individuals who eventually identified as gay or bisexual had, in their past, strongly denied any homosexual attraction to anyone asking for related information and also denied having such attractions to themselves. Therefore, when asking subjects about same-sex sexual desires / attraction, it may be advisable to offer some leeway in terms of responses. Therefore, what would it mean if a male did NOT chose "strongly disagree" in reply to "I am sexually attracted to men"? There were 19% (54/284) of males who did not "strongly disagree" with the statement "I am sexually attracted to men." There were 24% of females who did not "strongly disagree" with the statement "I am sexually attracted to women." This research was carried out by Richard A. Lippa (2001), California State University: Gender-related traits in gay men, lesbian women, and heterosexual men and women: the virtual identity of homosexual-heterosexual disgnosticity and gender diagnosticity. Journal of Personality, 68(5), 899-926. The above cited information is located on pages 903-5.

Another anomaly with respect to North American demophraphic studies based on sexual orientation have become apparent, especially in telephone surveys. In this respect, Tremblay and Ramsay (2000) note:

"Random telephone dialing techniques also produce apparently representative study population, and demographic results may be significantly different from the composition of gay community samples commonly used in studies. This outcome was reported by Harry (1990) and summarized in the abstract:

[N]ational probability sample of [American] males interviewed by telephone and asked their sexual orientation. Of these males 3.7 percent reported that they were homosexual or bisexual. Homosexual / bisexual men were compared with heterosexual ones on the demographic variables. This sample produced larger numbers in those groups which appear to be underrepresented in the usual samples drawn from the gay world. These groups include those with little education, married men, older men, minorities, and those living in small towns. It is suggested that probability samples which do not draw directly or heavily from the gay world for homosexual respondents obtain a broader sampling of those having homosexual feelings or behaviors (p. 89).
Harry (1990) reported that 38% of the sampled male population self-identified as homosexual or bisexual was not Caucasian, compared to 18% for heterosexual males, and the same overrepresentation of gay and bisexual males from racial minorities (non-Caucasian) also the result of a 1998 Exit Poll. People of color  formed 32 percent of the sample versus an 18 percent representation in the general population, with 4 percent of the sample identifying as gay, lesbian, or bisexual (HRC, 1999). Such results are important for researchers to know, especially with respect to recognizing that gay community samples represent a subset of the total population of homosexually oriented males in society. Telephone sampling, however, only produce study samples of individuals willing to reveal their homosexual orientation to someone requesting the information, and a significant percentage of homosexually identified individuals may withhold relevant information: about 60 to 70 percent of them if the Bagley and Tremblay (1998) homosexual and bisexual identification result for young adult males (11.1%) is accepted as a reasonable approximation of reality for young adult males. Furthermore, more sensitive "identification" information, such as having engaged in same-sex sexual activity, is often not solicited, thus not permitting the identification of individuals who have been homosexually active but are not identifying as homosexual, gay, lesbian, or bisexual.

Referenced Information

Billy J, Tanfer K, Grady W, Klepinger D (1993). The sexual behavior of men in the United States. Family Planning Perspective, 25, 52-60.

Binson D, Michaels S, Stall R, Coates T, Gagnon J. Catania J (1995). Prevalence and social distribution of males who have sex with males: United States and its urban centers. Journal of Sex Research, 32, 245-254.

Harry, Joseph (1990). A probability sample of gay males. Journal of Homosexuality, 19(1), 89-104.

HRC: The Human Rights Campaign  (1999). 1998 Exit Polls Show Diverse Picture of Identified Gay, Lesbian, and Bisexual Electorate. Press Release, February, 16. Internet Page: - . The Voter News Service exit poll was conducted for ABC, CBS, NBC, CNN, Fox and the Associated Press. The data was purchased by HRC from Voter News Service.
The American Exit Polls: The {1992} exit poll conducted interviews with 15,488 voters, selected to provide a statistically accurate portrait of the voting population. Of those who answered the self-identifier question, 3.2 percent identified themselves as gay, lesbian, or bisexual. (Power at the polls: the gay / lesbian / bisexual vote  (The National Gay and Lesbian Task Force, 1996) by John D'Emilio:

The size of the G/L/B vote in 1998 was similar to the 1996 Voter News Service exit poll, which put the G/L/B vote at 5 percent. Exit poll - G/L/B Voters More Non-White And Younger Than Overall Electorate, According To Poll - More G/L/B voters are non-white, 32 percent vs. 18 percent, and the G/L/B constituency is younger, 88 percent under age sixty compared to 72 percent under age 60 among voters overall. The G/L/B vote consists of more men, 59 percent male among G/L/B voters, compared to 49 percent male among all voters. Gay, lesbian and bisexual voters also had lower family income, with 66 percent under $50,000 vs. 52 percent for the entire electorate. (

Krieger N, and Sidney S (1997). Prevalence and health implications of anti-gay discrimination: a study of black and white women and men in the Cardia Cohort. International Journal of Health Services, 27(l), 157-76. (PubMed "abstract" link with a document delivery services.)

From Tremblay and Ramsay (2000): "The CARDIA (Coronary Artery Risk Development in Young Adults) Cohort: At baseline (1985-86), the sample consisted of an enrollment of 5,115 young adults, 18 to 30 years of age and, about ten years later, a decision was made to solicit "homosexuality" information to investigate of health effects possibly associated with experiencing anti-gay discrimination. The male demographic results based on "ever having a same-sex sexual partner" (Table: 15.3% for male) was similar to the Bagley and Tremblay (1998) 14.0% percent result for young adult males in the same age category (18- to 27-years-old) and sampled 5 years later. The  CARDIA Cohort study reported only slightly fewer females (13.3%) who acknowledged having at least one sexual partner.

Table - Cardia Cohort: Lifetime Same-Sex Partners
(Totals - Weighted* Results)
White/Black Men
White/Black Women
Sex of Sexual Partner
N = 693
N = 1031
Only Same-Sex
Weighted %
5.7% & 0.5%
1.8% & 0.7%
Both Sexes
Weighted %
10.6% & 5.4%
12.3% & 5.8%
Weighted %
n = 90
14.1% & 6.6% 
*Calculations by author, assuming a U.S. population of only white and black individuals, with about 90% Caucasian and 10% African-American individuals. The results are therefore a reasonable "approximation" of the percentage of white and Black American individuals who report having had lifetime same-gender sexual experiences.

From the CARDIA Cohort, about 5.2 percent of young adult males could be classified "homosexual" given their history of exclusive same-sex partners, a result consonant with the Bagley and Tremblay (1998) report that 5.6 percent of males self-identified as homosexual. The remaining 8.8 percent of males (14% minus 5.6%) reporting lifetime same-sex partners since the age of 12 years in the Bagley and Tremblay (1998) study also replicates the CARDIA Cohort 9.7 percent of males reporting to have been behaviorally bisexual."

Michael R, Gagnon J, Laumann E, Kolata G (1994). Sex in America: A definitive study. New York: Little, Brown, and Company.

Sedlacek WE, and Kim SH (1996). Survey Terminology Related to Sexual Orientation: Does It Matter? Report No. 12-96. (Article availability given in ERIC "abstract" link.) Originally published as: Kim SH, and Sedlacek WE (1995). Survey terminology related to sexual orientation: Does it matter? Counseling Center Research Report #11-95. University of Maryland, College Park. (Citation)

Abstract: Over the years, many different labels have been used to describe untraditional students. This study examined whether terminology used in surveys could affect respondents' answers. Two forms of a questionnaire were designed and distributed randomly to college students. One form consistently employed the terms gay, lesbian, and bisexual while the other form employed the term homosexual. On 29 of the 30 items, there were no significant differences between the two groups. However, on one item, the difference was significant. In the true-false statement, "I have had at least one homosexual experience during the last year," only 2.5% of the respondents marked it true. However, for the statement, "I have had at least one sexual experience with someone of my gender during the last year," 10.4% of respondents indicated this as true. It is speculated that the term "homosexual" seemed to generate some negative reactions and that respondents did not want to see themselves as homosexual. It is concluded that if researchers wish to measure behavior, it may be preferable to avoid any labeling and to be as operational as possible. Contains 11 references. (RJM).

Turner C, Ku L, Rogers S, Lindberg L, Pleck J, and Sonenstein F (1998). Adolescent sexual behavior, drug use, and violence: Increased reporting with computer survey technology. Science Magazine, 280(5365-8), 867-73.

Working Group (1995). Working group of the workshop on suicide and sexual orientation. Suicide and Life-Threatening Behavior, 25 (Suppl), 82-88.

To the Notes, Tables, Bibliography - On The prevalence of Homosexuality and Bisexuality in a Ramdom Community Survey....
A GLBT Internet-Based Education with InfoSearch Pages
Last Page! - On The prevalence of Homosexuality and Bisexuality in a Ramdom Community Survey....

To The Top Of The Page!
To: Home Page - Gay and Bisexual Male Suicidality Studies