in the context of his book, Understanding Homosexuality
by Tim Fisher
Tim Fisher
Minneapolis, MN
timrfisher@christian.net
November
1, 2006
David Glesne has insisted that if we want to argue that his statistics are not representative, we will need to present other studies that refute them. Of course, Glesne’s approach here is illogical, since bad science (or a bad use of science) is simply that: bad. Unrepresentative studies are . . . unrepresentative. But to respond to his challenge . . . Here are some statistics that do not support the claims about promiscuity that he makes in his book. Before I get to those, however, I have some preliminary comments.
II. THE STATISTICS: General Comments
III. STATISTICS SUPPORTING THAT GAY MEN ARE NOT MORE PROMISCUOUS
2. Vierod et al
5. Fay, Billy
6. My Own Research: Some Comparisons Based on the GSS
---------------
1) It is often said that “statistics don’t lie.” In a sense, this statement is true. (Unless, of course, the statistics come from Paul Cameron, in which case it is not true). But the question we need to ask of all statistics is not so much “Are these statistics true?” but rather “What is the nature of the truth these statistics are telling?”
When we are muddled in our answer to this last question, we end up with lots of different statistics giving us a wide variety of what appears to be “the truth.” Such is quite often the case with the social science research regarding sexuality. That is why we are currently stuck with a “my researcher can beat up your researcher” situation, where one side pits its favorite studies against the other. As long as you don’t look too closely at the full breadth of the research on homosexuality, you can make just about any claim you want. (This is somewhat of an exaggeration, of course, but it often feels to be the case.)
I fully realize that to write what I write below is to take part in the sort of dynamic I discuss above. But I do so in order to illustrate the point that Glesne’s promiscuity statistics do not support his conclusions adequately enough for us to formulate, on their basis, any coherent social or ecclesiastical policy.
2) The truth about so-called “homosexual promiscuity” is that we don’t know the truth. Sure, there seems to be a subset of gay men who seem to be quite promiscuous. But how big is this subset really? We don’t know with sufficient certainty. Nor do we know, with adequate precision, the size of the subset of heterosexuals who are promiscuous.
3) It seems to be bad logic from the get-go to link homosexuality and promiscuity on the basis of group statistics. In a church-based context, “promiscuity” is obviously a moral category. As with all moral categories, we measure sin not by noting the actions of a certain percentage within a group but rather by the actions of the individual. It doesn’t matter how many Samaritans are “bad” or “unclean” or “idolatrous.” What matters—to the bible, at any rate—is opening our eyes and hearts enough to see and appreciate the good Samaritan. He is who God holds up as the model for our behavior.
I find it illogical to argue the relative morality of homosexuals (and their associated homosexual behavior) on the basis of group statistics. All that these statistics can tell us is how prevalent a particular activity is within the group. Because some—even if it’s many, even if it’s most (which I don’t believe it is)—homosexual men are promiscuous by our standards (the church’s, this discussion’s) does not in any way support the idea that each individual homosexual is promiscuous. Nor does it support the idea that homosexuality inherently leads to promiscuity. Correlation does not imply causation. To pretend otherwise is to practice bad science.
Consider this parallel example: According to the US Department of Justice, an estimated 28% of black males will enter State or Federal prison during their lifetime, compared to 16% of Hispanic males and 4.4% of white males. What does this tell us about the criminal propensity of the black male compared to the white male? Should we conclude that being black has some causal correlation to committing crime? Should we restrict black men from certain activities and freedoms simply because other black men have committed crimes in greater proportion than white men? Of course not. We realize that an African heritage does not in any way cause criminality. We realize this now (most of us), but there was a time not too long ago when studies were done, experiments were made, that “proved” the moral inferiority of black people. Those researchers made the same type of mistake then that Glesne makes now: they noted correlations of skin color and behavior, and they then proceeded to pursue social policy (including church policy) as if the one caused the other. Correlation does not imply causation.
4) Please understand that I do not intend for my presentation of statistics below to somehow “prove” that gay men are not more promiscuous than straight men. As I have said above, there is simply too little good data to support such a claim. Instead, I present these statistics as a way of demonstrating that different researchers will obtain very different results, depending on where they point their microscopes.
One of the big difficulties with the social-science literature is that it is extremely difficult to obtain representative samples of a hidden—and stigmatized—population such as homosexuals. Most researchers must resort to what is called a “convenience” sample, which is what it sounds like: the sample is gathered in a particular way because it is convenient to do so. This method potentially skews the results. For instance, many sexuality researchers have solicited study volunteers in such places as gay bars, bath houses, AIDS health centers, and pornographic magazines.
Obviously, if researchers look in places that are likely to be frequented by highly promiscuous gay men, no one should be surprised if what they find is a lot of promiscuity. Most often, it is impossible to conclude that a convenience sample is representative—that is, that it applies to the general population of gay men.
Many studies have used convenience samples for the expressed purpose of finding gay men who are at high risk for being HIV-positive. The idea here is not to find samples representative of the general population of gay men, but rather samples representative of men who are most likely to contract AIDS. Having many sexual partners is associated with higher risks of contracting AIDS. What these studies intended to do was find gay men who might have AIDS. Since the researchers had reason to suspect that AIDS correlated with multiple sex partners, and they wished to test their hypothesis, the researchers looked for gay men in places where they knew relatively promiscuous gay men interacted.
In many cases, these are not “bad” studies. But as I discuss above, we have to be careful when discerning the nature of the truth these studies tell. The researchers found what THEY were looking for—not what Glesne is looking for.
III. STATISTICS SUPPORTING THAT GAY MEN ARE NOT MORE PROMISCUOUS
Laumann, Edward O., John H. Gagnon, Robert Michael, and Stuart Michaels, 1994. The Social Organization of Sexuality: Sexual Practices in the United States. Chicago: University of Chicago Press.
This study is one of the few (it might be the only one done to date, actually) reasonably comprehensive and methodologically sound surveys of America’s practices and beliefs. It does not rely on what researchers would call a convenience sample. But now let’s look at what Laumann et al themselves concluded about their statistics. Here are some quotes (emphasis added):
“There is a clear overall pattern in this [study]. In all cases, when we dichotomize our sample, the group of people with same-gender partners (or who define themselves as homosexual or bisexual) have higher average numbers of partners than the rest of the sexually active people in the sample. In many, if not most, of the cases for the men, _these differences are not statistically significant_. (314)
“The higher mean numbers of partners for respondents reporting same-gender sex corresponds to a stereotype of male homosexuals that is widespread in our society. . . . While some evidence in our data supports this general tendency, _the differences do not appear very large_ in view of the higher variability in our measures that results from the small sample size of homosexually active men.” (316)
This last finding, that the “differences do not appear very large,” is given some interesting, important attention by Jeramy Townsley, who writes
“[The] mean, the statistic presented by Laumann, may not be the best measure to report. A further analysis of the GSS data (on which Laumann based his results) indicates that the median (50th percentile) number of sexual partners for heterosexuals is five and for homosexuals is six. The discrepancy between the mean and median is indicative of a small sub-population of gay males who tend towards high rates of sexual partners, skewing the mean, while the majority of gay men tend to have rates about the same as heterosexual males.”
(Please see Townsley’s website, especially (www.jeramyt.org/gss/partners.html ), for a more detailed look at the raw numbers.)
It seems that the Laumann, et al study, as far as it goes, actually supports the idea that gay men are not much more promiscuous than straight men, and strongly argues against Glesne’s point. Although the Laumann authors admit that the sampling size of identified “active homosexuals” (my term) is relatively small, they believe their “data, limited in some respects though they may be, represent the most varied and comprehensive measures of different aspects of homosexuality to be collected on a representative sample of U.S. adults” (p. 320). When proper methods are used to gather samples (e.g. random sampling, statistical controls, etc…), even small samples can be statistically meaningful. Conversely, even a relatively large sample, if not collected appropriately, can result in seriously misleading findings.
“Prevalence and trends in homosexual behavior in Norway”
Scandinavian Journal of Social Medicine. 1997. Vo. 25(1):33-38.
This study was a big one and studied the degree of change in sexual practices among homosexual men during a 5-year period. The data comes from two questionnaire surveys of sexual behavior in the general population of Norway. The study found that the for men with current homosexual experience, the number of male partners decreased from a yearly median of 1.0 in 1987 to 0.3 in 1992. This is not what I would call promiscuity.
As we might expect, the numbers of sexual partners among gay men who have AIDS was found to be higher. Among HIV-positives, the number of male partners decreased from a yearly median of 4.3 before to 1.6 after awareness of HIV-seropositivity.
“Demographic Characterizes of Heterosexuals with Multiple Partners: The National AIDS Behavioral Surveys”
Family Planning Perspectives. 1993. Vol. 25 (5): 203-214
This study sampled from the general population of heterosexual men in the US. What is interesting to note in this study is its findings about heterosexual malesexual behavior.
Percentage of heterosexual men, by number of partners in the year before the survey:
0 partners 10.5 percent
1 partner 77.9
2 or more 11.2
Compared to Vierod et al, it seems heterosexuals have more sexual partners per year. (Of course, the Dolcini sample comes from the U.S., while the other studied Norway. Are these two cultures similar enough to compare? Also, we note that the Norwegian study listed the “median number of partners,” while the US listed percentages of men by number of partners they’d had sex with. Nevertheless, it seems quite probable that the median for the heterosexuals studied in the US study, with about 89 percent having 1 or more partners, would calculate to be higher than, or perhaps equal to, that of the Norwegian study, with a median of 0.3 in 1992.)
Another interesting stat from Dolcini:
Percentage of heterosexual men whose marital status is “separated,” by number of partners in the year before the survey:
0 partners 17.9 percent
1 partner 53.1.9
2 or more 29.1
The above statistic seems to suggest that homosexual men have no peculiar talent for promiscuity. In fact, the study found that in certain “high risk cities,” (cities thought to have high levels of AIDS), previously-married black men were over 12 times more likely to engage in multiple relationships than were men in general. Never-married white men were found to be over 10 times more likely then men overall. Never-married white women were over 37 times more likely than women overall.
“Partners National Survey of Lesbian & Gay Couples”
Journal of Gay and Lesbian Social Services (Vol. 1, #2, 1994)
This study has generated statistics looking much different than the ones Glesne presents. Researchers Bryant and Demian report that 63 percent of study participants (gay, male) reported being involved in monogamous relationships; 78 percent reported a frequency in extra-relationship sex of zero.
For this study, “survey forms were circulated through gay churches and community centers, but most couples requested the forms after reading notices in the gay press.”
Compare the above statistic with the Laumann study, which reports that between 21.7 and 26.4 percent of the men studied said they had had sex with someone other than their wife while they were married (p. 590). While this percentage—roughly one quarter—may be better than what some might have previously thought, is it really low enough for us to declare that heterosexuality is a “healthy lifestyle?” What should our measure of health be for the heterosexual lifestyle?
So, can we conclude that this study is satisfactorily representative of the general population of gay couples? Maybe not. But at the very least, what this study illustrates is how various the scientific research is. To list merely those studies which support one’s position, and leave out others, does not seem to be an adequately thorough approach.
The following text is by Jeramy Townsley. (See http://www.jeramyt.org/gay/gayhealth.html#prom). I have read the studies myself and have checked Townsley’s work. He reports the data correctly.
“In a study of sexual behavior in homosexuals and heterosexuals, the researchers found that of gay and bisexual men, 24% had one male partner in their lifetime, 45% had 2-4 male partners, 13% had 5-9 male partners, and 18% had 10 or more sexual partners, which produces a mean of less than 6 partners. (The statistics I did by myself using the data presented, which is presented as a percentage of total males interviewed, both gay and straight (p. 345)—they can be verified yourself by looking at the numbers given in the paper)(Fay; n=97 gay males of 1450 males total). In a parallel study, a random sample of primarily straight men (n=3111 males who had had vaginal intercourse; of the total sample of n=3224 males, only 2.3% had indicated having had sex with both men and women), the mean number of sexual partners was 7.3, with 28.2% having 1-3 partners, and 23.3% having greater than 19 partners (Billy). This data indicates that gay men may have fewer number of sexual partners than heterosexuals.”
The Billy study states:
“[M]ore than one-fifth of men have had 20 or more vaginal sex partners in their lifetime, and a similar proportion of never-married and formerly married men had had four or more partners over an 18-month period.”
(p. 58)
Keep in mind that the Billy study was of men of the relatively young age bracket of 20–39 years.
What I present here is data taken from the General Social Survey (GSS), collected by researchers at the University of Chicago. The GSS compiles information from surveys taken since 1972. The information is considered by almost everyone in the social sciences to be reliable. These are high-quality surveys. More specifically, as noted in the Laumann discussion above, the information from the GSS as it relates to sexual behavior is considered to be one of the very few high-quality surveys of its kind ever done, since the GSS data was collected in a truly random fashion with all the appropriate controls and balances. In fact, I believe the GSS is the only appropriately controlled statistical survey that’s been done regarding the sexual behavior of homosexual people. That is, it’s the only one that I know of that can, to any significant degree, be generalized to the population of all homosexuals in the United States.
The GSS website has a wonderful statistical interface that allows one to interact with a vast amount of information and produce accurate reports of one’s own design. (See http://webapp.icpsr.umich.edu/GSS/). I did some work with this interface; here are some results.
To simply matters, what I present in the table below looks only at males. It also doesn’t take into account bisexuals; since the GSS has “found” so few bisexuals, the data is not statistically significant. Across the top we see the “number of partners in the last five years”; along the left we see the “sex of those partners.” The values indicated are percentages. So for example, reading the table, we see that, of the men who had sex with men, 12 percent had only one partner. Of the straight men, 17.8 percent had only one partner. And so on . . . (By the way, the category at the far right, labeled “1+ DK#,” stands for “one or more partners, didn’t know the number.”)
Number of partners in last five years (expressed in % of N)
|
sex of partners |
0 |
1 |
2 |
3 |
4 |
5–10 |
11–20 |
20–100 |
100+ |
1+ DK # |
|
|
Exclusively Male |
0 |
12.0 |
8.0 |
12.8 |
12.8 |
19.2 |
14.4 |
14.4 |
4.0 |
2.4 |
N= 125 |
|
Exclusively Female |
0 |
17.8 |
14.2 |
14.4 |
14.4 |
23.3 |
9.3 |
5.1 |
0.4 |
0.7 |
N= 1,903 |
Comments:
As you can see, the gay men’s statistics track fairly closely with the straight men’s—up to a point. Where we see a significant discrepancy is at the very high levels of promiscuity (i.e. 20–100, and over 100 partners). This finding corroborates other surveys I’ve seen.
You will note that I have selected the data so that we are comparing gay men with straight men who are not currently married. I separated out the married straight men. I did this because it is important to compare, as it were, apples to apples. So, in comparing apples to apples, I compare groups that are in a similar situation—that is, I compare gay relationships taking place without the benefit of marriage with straight relationships of the same.
My conclusion, then, is that (according to the GSS data) there is a segment of gay men (roughly 13% of the total of gay men) who are promiscuous to a degree that straight men are not. That is, 87% of gay men display levels of promiscuity that are parallel to those of unmarried straight men.
In my view, the fact of this 13% is of grave concern. Our society needs to address this problem. But I don’t feel this 13% should persuade the church that a gay man (or, for that matter, a lesbian woman) who enters into a committed, loving relationship should be excluded from ordination and blessings on account of that relationship.
To make a racial comparison, I note (according to the GSS) that among black, never-married, straight men, 31.6% have had 11 or more partners in the last five years. Of white, never-married, straight men, the statistic is 15.5 %. The difference between black and white men on this measure is almost identical to the difference between gay men and unmarried straight men (of all races). So, should the church hesitate to ordain black, unmarried men?
If we are going to apply statistics to the question (i.e. “Should the church ordain/bless gay people in committed relationships?”), wouldn’t it be merely honest to apply the same sort of statistics when ordaining/blessing black men and their relationships? Yet the church does not apply statistics in this way, and for good reasons.
----------------------------------------------------------------------
Tim Fisher
Minneapolis, MN
timrfisher@christian.net
November 1, 2006