Homosexual Demography: Implications for the Spread of AIDS.

This paper presents a meta-analysis of the demographics of male homosexuality in the United States during the postwar period, and examines the numbers, age distribution, life expectancy, marital status, and fertility of men who have sex with other men. The data are from both national surveys and surveys of the gay population. A typology, which is used to explain why so few women appear to have been infected with HIV by bisexual men, is included.

Numerous surveys of gay men have been carried out by academics, journalists and gay activists. These surveys are of convenience samples usually collected by distributing questionnaires in gay bars and publications, or through homophile organizations. Some are national in scope (i.e. Jay & Young, 1979; Spada, 1979), others restricted to a single city. The first nationally representative survey was carried out by the Kinsey Institute in 1970, although the results were not published until much later (Klassen, Williams, & Levitt, 1989). In 1988, Harris Polls conducted a face-to-face survey using male interviewers. That same year the National Opinion Research Center (NORC), as part of its General Social Survey, began to include questions about the number and gender of sex partners. Harry (1990) interviewed a national probability sample by telephone. Other academic studies include one conducted by University of California San Francisco (Bower, 1991) and another by Diamond (1992). The 1992 New York Times exit poll of voters was the first to ask a question on sexual orientation. The results of two other studies (Brecher, 1984; Hunt, 1974), while not based on truly random samples, appear credible. The most ambitious and sophisticated survey is the study by Michael, Gagnon, Laumann, and Kolata, (1994).

Prevalence of Homosexuality

The prevalence of homosexuality can be measured over various periods of time (last year, last five years, lifetime, etc.), and can be defined behaviorally or in terms of an individual's sexual identity. Table 1 shows the results of several national surveys using both criteria.

Table 1. Percentage of Males Who Define Selves as
Gay/Bisexual or Report They Had Sex With Men

                                               Current/Last Year

Study

Kinsey (1970)          male sexual partner           1.6-2.0

Hunt (1972)            male sexual partner             2.0
Brecher (1978)         male sexual partner             2.3
Harris Poll (1988)     male sexual partner             3.5
Harry (1988)           gay/bisexual identity           3.7
NORC (1988)            male sex partner                2.4
NORC (1989)            male sex partner                1.2
NORC (1990)            male sex partner                1.9
UCSF (1990)            male sex partner                2.0
Diamond (1992)         gay/bisexual identity           3.0
NYT (1992)             gay/bisexual identity           3.2
Michael et al (1994)   gay/bisexual identity           2.8
                       male sex partner                2.0

                       Lifetime

Study

Kinsey (1970)          11.9 since 15
                       6.7 since 21
Hunt (1972)            10.0-11.0 since 15
Brecher (1978)         13."ever"
Harris Poll (1988)         --
Harry (1988)               --
NORC (1988)                --
NORC (1989)            4.9 since 18
NORC (1990)            4.8 since 18
UCSF (1990)                --
Diamond (1992)             --
NYT (1992)                 --
Michael et al (1994)       --
                       9.0 since puberty
                       5.0 since 18

Note. Dates refer to date study was performed.

There is a remarkable agreement among the studies. The proportion describing themselves as gay or bisexual averages slightly more than 3%, while those who say they had sex with a man in the previous year averages about 2%. The behavioral figures are noticeably higher if we consider lifetime incidence. The difference between the two measures suggests that a large proportion of those men who have sex with other men do so infrequently and/or episodically. There is no obvious trend over time, but this may be a result of differences in survey methodology and the small numbers involved.

Only a handful of reputable studies suggest a higher number. The Los Angeles Times, reporting the results of a 1985 survey, stated that 10% of the national sample identified themselves as gay (Kirk, 1989). In fact, while the Los Angeles Times poll did ask several questions about gays, an examination of the questionnaire at the Roper Center Archives found that respondents were not asked about their own sexual orientation. The San Francisco Examiner reported in 1989 that 6.2% of those surveyed were gay or bisexual. However, the numbers cited in the description of the poll are inconsistent with this claim, and suggest a much lower figure of 2.4% among men.(1) In a recent survey by the Yankelovich Monitor, 5.7% of the respondents described themselves as either gay, lesbian, or homosexual (Elliot, 1994). Rex Briggs of the Yankelovich Monitor has suggested that their procedures, by guaranteeing anonymity, produce a more accurate figure than previous polls (personal communication, June 14, 1994). However, until details of the latter survey are available for academic scrutiny, an agnostic stance on its findings seems appropriate.

Homosexuality is very much an age-linked phenomenon, and is far more common at younger ages. If we consider the adult population, the proportion identifying themselves as gay/bisexual, or reporting sex with a man in the previous year, is highest for those in their twenties. Thereafter, according to the New York Times survey(2), there is a steady decline, although both the Kinsey and NORC data cited by Fay, Turner, Klassen, & Gagnon (1989) reveal a more erratic pattern (see Table 2).

Table 2. Homosexuality by Age

Age     Kinsey 1970(a)    NORC 1988(b)    NYT 1992(c)

21-29       4.1               3.8             4.5

30-44       1.0               2.2             3.6

45-64       2.3               2.2             2.2

65+         0.0               2.9             1.8

Note. Results of polls shown in percents. Dates refer to date study was performed.

(a) Kinsey Institute, male sexual partner in previous year.

(b) National Opinion Research Center, male sexual partner in previous year.

(c) New York Times, gay or bisexual identity.

Surveys of gay men also indicate that homosexuality is more frequent among young men. Thus, the median age of those surveyed was 37 in Bell and Weinberg (1978), 31 in Jay and Young (1979), and 36 in Weinberg and Williams (1974). Even in the McWhirter and Mattison (1984) study of gay couples, which might have produced an older sample, the median age was 36. Gay samples draw upon a different group from that found in national surveys, with the peak years being the late twenties and early thirties. Presumably this population, consisting of those who are socially involved with the gay community, excludes those who engage in homosexual activity infrequently and episodically.

There are no direct data on the frequency of adolescent homosexuality, but indirect evidence suggests it is high. Questions which ask if individuals have ever had a homosexual experience result in a much higher figure if they include early ages. Brecher (1984), who asked if respondents had ever had a homosexual experience, found 13% who reported that they had, while Hunt (1974), who used age 15 as a cut-off, found 10% to 11% and NORC, which asked about activity since 18, found less than 5%. This implies not only that a lot of homosexual activity occurs during adolescence, but also that a large amount is confined to that period. This conclusion is supported by Michael et al. (1994), who found that, of those men who had ever had sex with another man, 40% had this experience before they were eighteen and never engaged in homosexual behavior again.

The data on cumulative incidence provided by Rogers and Turner (1991) are very interesting since they show that very few males start having homosexual sex after their twenties. The proportion reporting that they had ever had a homosexual experience was 5.6% of those in their twenties, 5.6% of those in their thirties, and 5.5% of those in their forties. This is supported by the gay surveys, which find that almost all gay men started having sex with men in their teens or twenties. Manosevitz (1972) reports the average age of "coming out" was 18, while Berger (1982) found that the first homosexual experience was usually in early puberty or adolescence, with the latest being 21. Spada (1979) found that only 2.1% had their first experience after 25. In the McWhirter and Mattison (1984) study nearly all had their first experience by age 22.

Mortality and Morbidity

The reduction in homosexuality after 30, particularly at the older ages, could be explained not only by behavioral changes, but also by differential mortality. Since the outbreak of AIDS, the mortality and health status of gay men have received considerable attention. Yet, even before this time, their lifestyle appears to have exposed them to greater health risks than heterosexual men. Venereal disease, hepatitis B, and alcoholism are found at much higher rates among gay men than the general population (Bell & Weinberg, 1978; Darrow, Barrett, Jay & Young, 1981; Ziebold, 1982).

The only data available on life expectancy among gay men are the obituaries in gay newspapers. A study by Cameron (1992) of 16 gay publications over an 11 year period (1981-92) found that the median age of death was only 39. Excluding AIDS deaths improved the picture to a small degree, and for non-AIDS deaths the median age was 42.

Marital Status and Fertility

A large number of men who engage in homosexual behavior are, or have been, married. The distribution by marital status is calculated by applying the rates given in the 1970 Kinsey survey and the 1988 NORC survey to the number of males in each category, as given in the Statistical Abstract for the relevant years (see Table 3). In both years a majority of those who reported having sex with men were currently married. Harry's 1988 survey found that 42% of those who described themselves as sexually attracted to men were married.

Table 3. Estimated Distribution by Marital Status of Men
Who Reported Homosexual Contact in Previous Year

Marital Status        Kinsey (1970)    %     NORC (1988)   %

Married                 651,000       (54)     827,000     (54)
Single                  506,000       (42)     542,000     (35)
Previously Married       50,000        (4)     176,000     (11)
Total                 1,207,000              1,545,000

Note. Dates refer to when study was performed.

However, samples of the gay-identified population suggest a very different pattern of marital status. In Bell and Weinberg's San Francisco study (1978), conducted in 1970 (the same year as the Kinsey survey), only 20% of those surveyed reported that they were, or ever had been, married. This discrepancy indicates that married homosexuals are much less likely to be involved in the gay community, or to define themselves as gay. Bearing in mind that the gay population is only a subcategory of those who engage in homosexual behavior, what do we know of their fertility? Only a small minority (about 10%) are fathers (see Table 4). This is a result of two factors: few gay men marry and only about half of those who marry have children.

Table 4. Marital History and Fertility Among Gay Men

Study                          % Ever     % With
                               Married    Children

Bell and Weinberg (1978)          20         10
Saghir and Robbins (1973)         18          8
Jay and Young (1979)              16         13
McWhirter and Matison (1984)      15          8

A MODEL OF HOMOSEXUAL BEHAVIOR

It is clear that men who engage in homosexuality vary in the extent, duration, and timing of such activity. Ethnographic accounts suggest five types, and most individuals can be placed in at least one of them. The typology is based upon an individual's life history and is, therefore, retrospective in nature. In terms of Kinsey's scale of sexual preferences, types I and II would be 5 or 6 (with a strong preference for male sexual partners). Type III would rate 3 or 4 on the scale, (equally attracted to both men and women), while types IV and V would be 1 or 2 (prefer women but also mildly attracted to men).

Type I: Open Preferential Homosexuals

Some individuals of this type report that they have been exclusively homosexual throughout their lifetime, and strongly reject the idea of having sex with women.

   I have never had a sexual experience with a woman and have absolutely no
   desire to have one. I most certainly do not hate women. I think that they
   are fine, enjoyable to work with, socialize with, but would never want to
   live with one, much less have sex with one.... I'd no more be sexually
   intimate with a woman than I would with a reptile. Utterly repulsive idea.
   I've never even come close to sex with a woman. (Jay &Young, 1979, p.125)

Type II: Repressed Preferential Homosexuals

Others, however, describe attempts to lead a heterosexual life, and a significant minority get married. Their accounts usually indicate that this was due to social pressure, and that they tried to suppress their homosexual desires.

   In college I had three male one-night stands and no sex with girls. These
   three sessions caused me much grief--I struggled to forget the fact that
   they had really happened. I would not accept the fact that I might be a
   (horror of horrors) homosexual.... I married in my mid-20s still struggling
   with the guilt of those three one-night stands. I thought that marriage
   would erase all gay desires and that I would somehow escape the guilt
   feelings (actually the gay urges and yearnings I was struggling with
   subconsciously and trying to suppress). Sex with my wife was not terribly
   exciting. (Jay & Young, 1979, p. 119)

Type III: Bisexuals

Although many of the above groups may be, in a behavioral sense, bisexual, they are very different from those who prefer a blend of homosexual and heterosexual activity. Saghir and Robins (1973) quote a good example of this type.

   I am certainly enjoying the heterosexual and homosexual contacts. It is
   like apples and oranges. I like them both. I would object to giving any of
   them up. I get out and seek homosexual partners in my off hours and
   sometimes I make excuses that I have to work late. After 5 years of
   marriage, I still have intercourse with my wife 3-4 times a week.... [and]
   at least four homosexual contacts a week. (p. 99)

Type IV: Experimental Homosexuals

Many males experiment with homosexuality during adolescence, and in their twenties. Then they start having sex with women and stop having sex with men. Harry and DeVall (1978) argue that these men are really heterosexuals "who for a short period of time dabble in homosexual behavior" (p. 73). One of Brecher's respondents describes such experimental homosexuality (1984): "When I was about 15, the six or eight boys who hung together indulged in a summer of group masturbation, oral sex, and attempts at anal sex. This passed as soon as we discovered girls" (p. 217).

Type V: Situational Homosexuals

Men in this group would include Humphreys' (1970) "tea room trade," made up of "men with heavy emotional commitments to families and jobs [who] may not be able to afford investment in other than the most transient and impersonal types of extramarital sex," (pp. 153-154). This type of homosexuality is linked to the unavailability of female partners, and is common in prisons and other all male settings (Brecher, 1984; Humphreys, 1970). Whitam (1986) notes that this involves heterosexual men "using" homosexual men, and typically the men involved do not define themselves as gay. Experimental and situational homosexuals differ in age and marital status, with the former being young and unmarried and the latter usually older and married. Humphreys (1970) found that 89% of the trade category were married with a mean age of 38. Figure 1 summarizes the characteristics of each type.

Figure 1. A Typology of Homosexuality

                            Type I            Type II

                                                Repressed
Characteristics         Open Preferential       Preferential

Self Identity           Gay/bi                  Denial-Claims to
                                                be heterosexual

Social Disclosure       Open Participation      Closeted
                        in Gay Community

Age/Marital Status      Unmarried               Often Married

Homosexual Behavior     Frequent                Intermittent and
                                                Infrequent

Heterosexual Behavior   Very Rare               Moderate

                        Type III                Type IV

Characteristics         Bisexual               Experimental

Self Identity           Gay/bi                 Uncertain

Social Disclosure       Varies                 Varies

Age/Marital Status      Often Married          Young Unmarried

Homosexual Behavior     Frequent               For a short duration

Heterosexual Behavior   Frequent               Infrequent

                        Type V

Characteristics         Situational

Self Identity           Heterosexual

Social Disclosure       Hides Behavior

Age/Marital Status      Middle-aged
                        Usually married

Homosexual Behavior     Intermittent

Heterosexual Behavior   Infrequent

FINDINGS

A reasonable estimate is that currently about 3% of males are preferential homosexuals or bisexuals, and that another 4% are experimental or situational homosexuals. This distribution by type is consistent with two separate surveys. According to Michael et al. (1994), the number of men who identify themselves as homosexuals or bisexuals is almost the same as the number who report that they have had sex with men, but identify themselves as heterosexuals. Presumably the latter group is largely made up of situational homosexuals and those who experimented with homosexuality in the past. It is also compatible with the 1970 Kinsey data on lifetime incidence of homosexual contacts since the age of 21, with 1.4% reporting fairly often, 1.9% reporting occasionally and 3.4% reporting rarely or only once or twice. (Klassen, Williams, & Levitt, 1989). Given that some of these will go on to have more extensive experiences, and since preferential homosexuals and bisexuals have a more extensive history of homosexuality, the breakdown corresponds to our 3% preferential/bisexual group and 4% experimental/situational.

The simplest explanation for the distribution of homosexuality by age--with the highest rate among those in their twenties, followed by a marked decline--is that it reflects homosexual experimentation. A number of young men are confused about their sexual identity, but eventually decide that they prefer women to men and cease their homosexual behavior. The decline of homosexuality after age 50 could be attributed to a shortened life-span and/or an inability to find partners. Whereas Brecher (1984) found that 98% of heterosexual males in their fifties, and 91% in their sixties were sexually active, Pope and Schulz (1990) found that for homosexual males the figures were only 79% and 66%.

Changing Patterns Since 1960

To what extent have patterns of homosexual behavior changed over time? Two related factors presumably had significant effects: the emergence of gay communities in most large cities, and more tolerant social attitudes.

Throughout the 1960s, several cities experienced a buildup of gay neighborhoods, containing bars, restaurants, clubs, and other institutions catering to homosexuals, (Gagnon, 1990). One indicator of the growth of the gay community is provided by the number of gay establishments listed in Bob Damron's Address Book which rose from 690 places in 1964 to 5800 in 1983 (Darrow, Gorman, & Glick, 1986).

In the 1940s and 1950s American society was highly intolerant of homosexuality, and homosexuality was illegal in most states. Illinois decriminalized homosexual acts in 1962, and by 1983, a majority of states had revoked their sodomy laws. Although public attitudes towards homosexuality are still generally negative, there is a clear trend to greater tolerance with regard to employment issues (Gallup Poll Monthly, June 1992).

Decreased legal and social sanctions against homosexual behavior, in conjunction with the increased opportunities presented by the growth of the gay community, produced several changes in homosexual behavior. First, as Table 3 shows, the proportion of previously married men among those who acknowledged having sex with men increased dramatically, from 4% to 11%, between the 1970 and 1988 surveys. Presumably this was a result of homosexual husbands coming out of the closet. Second, the proportion of homosexuals who never marry increased over time. Using the studies of gay men which have been conducted at different times, we can construct an historical series by treating the different samples as quasi-cohorts. According to the Statistical Abstract of the United States (1997), in 1980 only 2.3% of first marriages were to husbands aged over 34, while in 1990 the figure was 5.3%. Since almost all men who marry do so by their early thirties, and since all samples have a median age greater than this, we can calculate the date at which the median individual in each sample reached 30. As Table 5 shows, there is a noticeable increase in the proportion never married.

Table 5. Estimated Proportion of Never-Married Men in Gay
Samples by Year.

Sample Median Age = 30 in (%)       Study

    1945 (54)                       Brecher (1984)
    1953 (71)                       Berger (1982)
    1962 (83)                       Weinberg and Williams (1974)
    1963 (80)                       Bell and Weinberg (1978)
    1968 (82)                       Saghir and Robbins (1973)
    1976 (84)                       Jay and Young (1979)
    1976 (85)                       McWhirter and Mattison (1984)
    1977 (83)                       Spada (1979)
    1989 (93)                       Woods (1993)

Experimental homosexuality might also have increased, at least initially. Gagnon (1990) argues that:

   The publicity afforded same-gender erotic desire, both in the majority
   media and in the existence of a gay and lesbian cultural apparatus, has
   changed the very character of the process of "coming out" (acquiring a gay
   or lesbian self-definition), ... Being gay or lesbian is now a more
   reasonable possibility, often a positive alternative to what
   heterosexuality offers, and there is now the opportunity for both
   anticipatory socialization in and practical experimentation with gay and
   lesbian role enactments that require relatively modest levels of erotic
   commitment. (p. 197)

Gagnon's position is plausible for the period from 1970 until the early 1980s. However, the gay world would have appeared far less appealing to the erotically uncommitted after the emergence of AIDS. One poll which asked what precautions were being taken to avoid AIDS found that 58% of those surveyed said they were taking more care in choosing their sexual partners, while 18% said they were avoiding people they suspected of being homosexuals. ("Growing Concern, Growing Precautions," 1986). It seems likely, therefore, that experimental homosexuality would have declined significantly in the latter half of the 1980s. Presumably there would also be a reduction in the amount of homosexual activity among bisexual individuals.

There is some fragmentary evidence to suggest an increase in experimental homosexuality and bisexuality prior to the outbreak of AIDS. A Playboy survey conducted in early 1982 found a high proportion of its readers (11.5%) had a predominantly heterosexual history but some homosexual experience (Lever, Kanouse, Rogers, & Carson, 1992). Surveys of the gay community, carried out from 1968 to 1977, found a high proportion of men who scored low on the Kinsey scale. The proportion rated as less than 5-6 on the Kinsey scale was 20% in Weinberg and Williams (1974), 9% in Bell and Weinberg (1978), 4% in Saghir and Robbins (1973), and 24% in Harry and DeVall (1978). In the Jay and Young survey (1979), 20% described themselves as bisexual.

The Implications of Homosexual Demographics for the Spread of AIDS

The increased attention paid to gays and homosexuality is partly due to the growing visibility and militancy of the gay community. Equally important is the emergence of AIDS, and the historical connection in the United States between AIDS and homosexuality. AIDS was initially labelled GRID (Gay Related Immune Deficiency), and regarded as a disease that attacked only gays. However, although HIV quickly spread from homosexuals to injecting drug users, and then to their sexual partners, it has made only slight inroads into the general population. (Gould 1993). Initial predictions that AIDS would spread rapidly from gays into the general population due to the large numbers of men who were behaviorally bisexual have been falsified. (See Fumento, 1990, for a description of how the media emphasized the role of bisexual men in spreading HIV to women). Only a handful of women diagnosed with AIDS are classified as having gotten the virus from bisexual men. According to the Center for Disease Control, by the end of 1992 the total number of women who had become infected through sex with a bisexual male was only 1,389. Anal intercourse is far more dangerous than vaginal intercourse in spreading the AIDS virus from an infected person to his partner, but the risk is low in both cases: approximately one in a hundred in the former case, and one in a thousand in the latter case. Thus, the risk for each group depends on both what they do, and how often they do it (Fumento, 1990).

The original locus of the HIV/AIDS epidemic in the early 1980s was among the preferential homosexuals, particularly those most involved in the gay community. The sexual lifestyle of this group could have been designed to facilitate the spread of AIDS. A 1989 National Research Council study on AIDS notes that:

   Before the viral cause of AIDS was known or detectable, behavioral research
   had shown that being sick with AIDS or showing laboratory signs of immune
   deficiency was associated with (1) a large number of different sexual
   partners; (2) receptive anal intercourse; (3) the use of bathhouses for
   sexual contact; (4) frequent infection with sexually transmitted diseases,
   particularly gonorrhea, syphilis, and enteric parasites (p. 11).

Gay liberation in the 1970s was accompanied by an explosion of gay sexuality, especially in tolerant cities such as New York and San Francisco, where gays concentrated. The director of the New York City Department of Health describes the situation as follows:

   By their own reports, many men had large numbers of sexual partners
   annually, often numbering in the hundreds and even in the thousands.
   Frenetic casual and anonymous sex was widespread among homosexual and
   bisexual men. Bathhouses, back rooms of bars and clubs, and other public
   settings such as erotic bookstores and movie "heaters were, in effect, wide
   open. Sexual practices such as anonymous group sex, sado-masochistic
   fantasies enacted with physical trauma, penetration of rectal orifices with
   penises, fists, and blunt objects - all of these practices and more were
   accompanied by extremely high rates of sexually transmitted diseases and
   set the scene for the rapid transmission of HIV once it appeared in the
   late 1970s. Rates of rectal and oral gonorrhea in males soared (Joseph,
   1992, p. 98).

But what of the other groups of men who had sex with men? How likely were they to become infected, and to infect their female partners? Situational homosexuals are a very low risk group, in part because they have sex with men only intermittently. Most important, if Humphreys (1970) is correct, they rarely engage in the most dangerous behaviors, such as receptive anal intercourse. They are also unlikely to transmit the HIV infection to their wives since they rarely have sex with them, which is why they resort to sex with other men.

Repressed gays, if they completely repress their desires are, of course, at no risk at all. Hunt (1974) reported very low annual frequencies of five to six encounters among closeted homosexuals. Bell and Weinberg (1978) found, however, that some married gays take several years before leaving their wives, and during this time they often have sex with men. The median length of marriage was 3 to 5 years, but 25% lasted more than 10 years. Since there is generally an inverse relationship between homosexual and heterosexual activity, with the frequency of marital sex declining dramatically by the last year of marriage, this considerably reduces the risk of the husbands infecting their wives.

The concept of experimenting with homosexuality is somewhat ambiguous, and may include anything from a single occasion to repeated activity for two or three years. Homosexual experimentation usually involves only a few experiences, if we accept the 1970 Kinsey survey data (Klassen et al., 1989). Therefore, these young men are unlikely to become infected, although any that are seem likely to transmit the virus to their subsequent female partners.

In the early period, bisexuals were likely to have a high risk of becoming infected and, since they also had an active heterosexual life, were very likely to infect their female partners. However, behavioral changes among this group, after they became aware of AIDS, are very marked. One study comparing sexual behaviors among bisexuals in 1983 and 1988 found a dramatic decline in the number of same-sex partners, anonymous sex, and group sex. Most important, those reporting that they never engaged in receptive anal sex rose from 32% in 1983 to 82% in 1988 (Weinberg, Williams, & Pryor, 1994).

By paying attention to the social and behavioral differences between men who have sex with men, rather than treating homosexuality as a single phenomenon, it is possible to estimate more precisely the likely spread of HIV among different groups of men, and from these men to their female partners.(3)

(1) The survey was carried out by Teichner and Associates, who "started with the premise that it would take 4,000 calls to reach 400 gays, both nationally and in the Bay Area. That was based on the idea that 10 percent of the population is gay--the generally accepted figure resulting from Alfred C. Kinsey's groundbreaking studies in the late 1940s. Instead, it took close to 27,000 calls to find the 800 people needed to make both gay samples statistically significant" (Hatfield, 1989, p. A20). Since in the Bay Area sample, 10% acknowledged being gay, this implies that 23,000 calls (27,000-4,000) had to be made to find 400 gays nationally, or that the rate was 1.7% for both sexes combined. Adjusting for the higher numbers of male homosexuals compared to lesbians found in the survey (287 vs 113), this suggests a rate of 2.4% for men and 1.0% for women.

(2) The data on the age distribution of self-identified gay men in the survey was provided by the New York Times.

(3) See Hewitt and Peverly (1996) for a computer simulation of the sexually-transmitted spread of AIDS within the United States, that divides the population into different categories on the basis of their sexual behavior.

REFERENCES

Bell, A. P., & Weinberg, M. S. (1978). Homosexualities: A study of diversity among men and women. New York: Simon and Schuster.

Berger, R. M. (1982). Gay and gray: The older homosexual man. Urbana, IL: University of Illinois Press.

Bower, B. (1991, August 31). Risky sex and AIDS. Science News.

Brecher, E. (1984). Love, sex and aging. Boston: Little Brown.

Cameron, P. (1992). The homosexual lifespan. Colorado Springs, CO: Family Research Institute.

Darrow, W., Gorman, E., & Glick, B. (1986). The social origins of AIDS: Social change, sexual behavior and disease trends. In D. Feldman & T. Johnson (Eds.), Social dimensions of AIDS. New York: Praeger.

Darrow, W., Barrett, D., Jay, K., & Young, A. (1981). The gay report on sexually transmitted diseases. American Journal of Public Health, 71, 1004-1010.

Diamond, M. (1992). Homosexuality and bisexuality in different populations. Archives of Sexual Behavior, 2, 291-310.

Elliott, S. (1994, June 9). A sharper view of gay consumers. New York Times, pp. D1-3.

Fay, R. E., Turner, C. F., Klassen, J. H., & Gagnon, J. H. (1989). Prevalence and patterns of same gender sexual contact among men. Science, 243, 338-348.

Fumento, M. (1990). Myth of heterosexual AIDS. New York: Basic Books.

Gagnon, J. H. (1990). Gender preference in erotic relations: The Kinsey Scale and sexual scripts. In D. P. McWhirter, S. A. Sanders, & J. M. Reinisch (Eds.), Homosexuality/heterosexuality. New York: Oxford University Press.

Gallup Poll. (1992, June). Tolerance of homosexuality is on the rise among the public. Gallup Poll Monthly. Princeton, N J.

Gould, P. (1993). The slow plague. Oxford, England: Blackwell.

Growing concern, greater precautions. Newsweek, (1986, November 24), 108, 32-33.

Louis Harris and Associates. (1988). Harris Poll Survey for Project Hope. New York.

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

Harry, J., & DeVall, W. (1978). The social organization of gay males. New York: Praeger.

Hatfield, L. (1989, June 5). Method of Polling. San Francisco Examiner, A3, A20.

Hewitt, C., & Peverley, J. R. (1996). The spread of AIDS into the general population: A simulation. Population Research and Policy Review, 15, 311-325.

Humphreys, L. (1970). Tearoom trade: Impersonal sex in public places. Chicago: Aldine.

Hunt, M. (1974). Sexual behavior in the 1970s. Chicago: Playboy Press.

Jay, K., & Young, A. (1979). The gay report. New York: Summit.

Joseph, S. (1992). Dragon within the gates. New York: Carroll & Grof.

Kirk, M. (1989). After the ball. New York: Doubleday.

Klassen, A. D., Williams, C. J., & Levitt, E. E. (1989). Sex and morality in the U.S. Middletown, CT: Wesleyan University Press.

Lever, J., Kanouse, D., Rogers, W., & Carson, S. (1992). Behavior patterns and sexual identity of bisexual males. The Journal of Sex Research, 29, 141-167.

Manosevitz, M. (1972). The development of male homosexuality. The Journal of Sex Research, 8, 31-40.

McWhirter, D., & Mattison, A. (1984). The male couple. Englewood Cliffs, NJ: Prentice-Hall.

Michael, R., Gagnon, J., Laumann, E., & Kolata, G. (1994). Sex in America: A definitive survey. Boston: Little, Brown.

National Research Council (1989). AIDS: Sexual behavior and intravenous drug use. Washington, DC: National Academy Press.

Pope, M., & Schulz, R. (1990). Sexual attitudes and behavior in midlife and aging homosexual males. Journal of Homosexuality, 20, 169-177.

Rogers, S. M. & Turner, C. F. (1991). Male-male sexual contact in the USA. The Journal of Sex Research, 28, 491-519.

Saghir, M., & Robbins, E. (1973). Male and female homosexuality: A comparative investigation. Baltimore: William and Wilkins.

Spada, J. (1979). The Spada report: The newest survey of gay male sexuality. New York: Signet.

U.S. Dept. of Commerce. (1997). Statistical Abstract of the United States. Washington: DC.

Weinberg, M., & Williams, C. (1974). Male homosexuals - Their problems and adaptations. New York: Oxford University Press.

Weinberg, M., Williams, C., & Pryor, D. (1994). Dual attraction: Understanding bisexuality. New York: Oxford University Press.

Whitam, F. L. (1986). Male homosexuality in four societies. New York: Praeger.

Woods, J. (1993). The corporate closet. New York: Free Press.

Ziebold, T. (Ed). (1982). Alcoholism and homosexuality. New York Haworth Press.

Manuscript accepted July 2, 1998

Christopher Hewitt

University of Maryland

Address correspondence to Dr. Christopher Hewitt, Department of Sociology/Anthropology, University of Maryland Baltimore County, 1000 Hilltop Rd., Baltimore, MD. 21250; e-mail: [e-mail]