Point By Point: A Look At Paul Cameron’s Medical Consequences of What Homosexuals Do
Part 2: “Oral Sex”
In Which Dr. Cameron’s Famous Survey Makes Its Appearance
January 31, 2006; revised February 27, 2006
In the previous section, Dr. Cameron provided us with a broad overview of “what homosexuals do.” Now he’s ready to begin discussing each sexual activity in more detail, starting with oral sex. This seems a good place to start since oral sex is very popular among many heterosexuals, and the average reader is likely to be very familiar with this topic. By starting here, Dr. Cameron gently introduces his readers to more explicit subject matter while lowering the risk of alienating his audience so early in the pamphlet.
After a brief statement in which he claims (without attribution) that nearly all homosexuals fellate their partners, Dr. Cameron quickly jumps into the purported medical significance of fellatio. Unfortunately, he offers no explanation as to why heterosexuals who engage in this practice would not be exposed to the same risks:
… Semen contains many of the germs carried in the blood. Because of this, gays who practice oral sex verge on consuming raw human blood, with all its medical risks… Since many contacts occur between strangers (70% of gays estimated that they had had sex only once with over half of their partners17, 27, and gays average somewhere between 106 and 1105 (sic. – ed.) different partners/year, the potential for infection is considerable.
17. Bell Alan P.; Weinberg, Martin S. Homosexualities: A Study of Diversity Among Men and Women (New York: Simon & Schuster, 1978).
27. Beral, Valerie; Bull, Diana; Darby, Sarah; Weller, Ian; Carne, Chris; Beecham, Mary; Jaffe, Harold. “Risk of Kaposi’s sarcoma and sexual practices associated with faecal contact in homosexual or bisexual men with AIDS.” Lancet 339, no. 8793 (March 14, 1992): 632-635.
Due to a typographical error in the online version of this pamphlet, it appears that the following references were intended to be cited in this paragraph:
5. Corey Lawrence; Holmes, King K. “Sexual transmission of hepatitis A in homosexual men.” New England Journal of Medicine 302, no. 8 (February 21, 1980): 435-438.
6. Cameron Paul; Proctor, Kay; Coburn, William, Jr.; Forde, Nels. “Sexual orientation and sexually transmitted disease.” Nebraska Medical Journal 70, no. 8 (August 1985): 292-99.
Cameron, Paul; Cameron, Kirk; Proctor, Kay. “Effect of homosexuality upon public health and social order.” Psychological Reports 64, no. 3 (June 1989): 1167-1179.
Dr. Cameron is a little ambiguous in his terminology. In referring to “oral sex,” he refers specifically to fellatio, not cunnilingus. By extension, he discusses activities exclusive to gay men and continues to ignore lesbians.
“Semen contains many of the germs carried in the blood. Because of this, gays who practice oral sex verge on consuming raw human blood, with all its medical risks.” This is quite a stretch. Saliva also contains many of the germs carried in the blood — think of a common cold sore or the small scratches on the surfaces of the gums which can occur during tooth brushing and flossing. Wouldn’t the simple act of kissing “verge on consuming raw human blood”?
Nevertheless, in discussing oral sex among gay men Dr. Cameron ignores heterosexual couples who do the same thing, presumably with all of the same risks of “consuming raw human blood.” He is silent when one of his sources (Gebhard & Johnson, #16) report that the Kinsey surveys of 1938-1963 found that about a third of all heterosexual married women performed fellatio in those pre-“sexual revolution” times.T He also neglects to mention that even his own flawed study (“Effect of Homosexuality Upon Public Health and Social Order”, #6), which we will discuss in the next section, reported that 75% of heterosexual women fellate their partners.U So why are the risks of oral sex unique to gay men?
Dr. Cameron appears to supply this answer with a discussion on promiscuity, which is what this section is really all about.
“70% of gays estimated that they had had sex only once with over half of their partners.” For this claim, Dr. Cameron cites two sources.
The first source is Bell & Weinberg’s (#17) Homosexualities. But when you turn to Table 7 on page 309 of that book, you will find this statistics for White gay males only. When Black gay males are included, it drops to 64%. While 64% is a pretty large number, it’s apparently not large enough to suit Dr. Cameron’s taste. Again, this is part of a common pattern in Dr. Cameron’s work: the selective manipulation of data to make his claims more outrageous than the original authors intended — even when an honest rendering of the data can be bad enough.
But not only does Dr. Cameron manipulate data without revealing what he’s doing, he turned to a book which he elsewhere claims to be rife with inaccuracies, saying it is not worthy of more than a “D-”.V Yet despite his low opinion, this book remains one of his favorite resources, and he cites it in nearly everything he publishes for general audiences.
But it turns out that Dr. Cameron is correct to criticize the book’s inaccuracies. He should heed his own criticism, especially considering where the study got its gay male sample:W
|Where Recruited||% of Respondents|
|Bars & Nightclubs:||29%|
|Parks & Other Public Cruising Areas:||4%|
|Private Bars (Sex clubs):||6%|
|Homophile Organizations, Mailing Lists:||12%|
Clearly they didn’t use any of the common methods of obtaining a truly random sample. There was no random-digit telephone dialing, no selection according to census data, nor were any other random probability sampling techniques used. Nearly half of their sample consists of people they found hanging out in bars, bathhouses, sex clubs and public cruising areas. And all of this was in San Francisco, 1969-1970 — in the city of love not long removed from the summer of love, at a time and place where many young heterosexuals were also looking for their Mr. (and Ms.) Goodbars.
Unfortunately, the authors do not provide data from a similarly-obtained local heterosexual population to compare with the homosexual population, so we cannot know how this non-random selection of gays match up with a similarly non-random selection of heterosexuals. And even if we did, the comparison would not be apples to oranges. Bell & Weinberg included in their study everyone who rated anywhere from a “2” to a “6” on the Kinsey scale — everyone who was mostly heterosexual to exclusively homosexual (where a “1” denotes exclusive heterosexuality and a “6” denotes exclusive homosexuality). We cannot know if the mostly-heterosexuals (the “2’s” and “3’s”) affected the results while in the epicenter of the free-love revolution.
This study cannot be used to describe gay men as a whole. At best, it can only describe the behaviors of those who were found in those locations at that time, and who agreed to participate. The authors themselves agree, explaining:
It should be pointed out that reaching any consensus about the number of homosexual men or women exhibiting this or that characteristic is not the aim of the present study. The non-representative nature of other investigators’ sample as well as our own precludes any generalization about the incidence of a particular phenomenon even to persons living in the locale where the interviews were conducted, much less to homosexuals in general. … We cannot stress too much that ours is not a representative sample.”X
And when Bell & Weinberg learned how Paul Cameron was using their work, they reacted this way:
“For him to use our figures to estimate differences between homosexuals and heterosexuals across the board in the general population is ludicrous.”Y
The details of the Beral, et al. study were first discussed here.
But Dr. Cameron presses on. For additional support, he again cites Beral, et al. (#27) which consisted of a convenience sample of 65 men who were selected for their exceptionally high-risk behavior, all of whom contracted AIDS as a result. Contending that the promiscuity of this high-risk convenience sample demonstrates that all gay men are promiscuous is simply wrong.
“Gays average somewhere between 106 and 1105 different partners/year.” In the online version of Medical Consequences of What Homosexuals Do, Dr. Cameron doesn’t appear to offer any support for this claim. But it turns out that this claim is exaggerated due to a typographical error.
As I mentioned before, much of Medical Consequences was reprinted in Dr. Cameron’s book The Gay Nineties, including this entire paragraph.Z And by comparing this pamphlet with those passages in the book, it becomes obvious that the characters “5” and “6” are supposed to be reference citations, which means the passage should read “between 106 and 1105 different partners/year”. With this error still present in the online version of Medical Consequences after more than 6 years, it reinforces the charge that Dr. Cameron’s isn’t much concerned about the accuracy of his work. His exaggerated claim of “between 106 and 1105” is making the rounds throughout the internet, where it is often accepted without question.
For the corrected claim of 110 partners per year, Dr. Cameron is citing Corey & Holmes (#5) — the “daily sexual diaries.” But again, this study was yet another convenience sample of 96 gay men, each and every one of whom was a patient at an STD clinic.
But that’s not the only problem with how Dr. Cameron uses this study. The study’s authors state that one group of gay men (67 subjects) had an average of 2.3 partners per month, while the second group (29 subjects) had an average of 1.2 partners per month. The combined figure is just under 2 partners per month. This isn’t even close to 110 partners per year as Dr. Cameron claims. The study that Dr. Cameron cites simply doesn’t offer any support that the gay men in their STD clinic have 110 partners per year. But since the average reader is unlikely to follow up on Dr. Cameron’s footnotes, this fabrication nearly always goes undetected.
Typographical errors and outright fabrications aside, it’ true that twenty-four partners per year is a lot. That’s promiscuous by any reasonable standard. But we must remember that these are patients from an STD clinic, which tends to be a promiscuous group of people. It is well known among epidemiologists and social scientists that patients at STD clinics are there precisely because they are engaging in higher risk behaviors than those who are not.AA Consequently, when you go trolling for data from STD clinics to describe “what homosexuals do”, you only discover what promiscuous homosexuals do. And the same holds true for heterosexuals as well. Unfortunately, this particular study has no control group of heterosexuals for making such a comparison, but other studies consisting of heterosexual men attending STD clinics suggest that they have a similarly high numbers of sexual partners.BB
As for the more conservative estimate that gay men may have on average 10 different partners per year, Dr. Cameron relies on his own discredited studies, “Sexual Orientation and Sexually Transmitted Disease”, and “Effect of Homosexuality on Public Health and Social Order”, (#6) both of which are based on a flawed survey he conducted between 1983 and 1984. Dr. Cameron often returns to his flawed survey, directly or indirectly, in much of what he publishes.
Paul Cameron founded and chaired the Institute for the Scientific Investigation of Sexuality (ISIS) in the early eighties. Soon after, he conducted the seven-city ISIS Survey in 1983, (expanded in 1984 to an eight-city survey with data from Dallas.) In these seven cities (Bennett (NE), Denver, Los Angeles, Louisville, Omaha, Rochester (NY), and Washington, DC.), volunteers went door-to-door and handed the questionnaire to the first adult who answered the door, with an adult male given preference if one was in the house.
Twenty years later, Dr. Cameron continues to refer to this survey in many of his publications, which he often claims to be the largest random sex survey in the U.S. But many of the problems with this survey were noted immediately following its first appearance in the Nebraska Medical Journal. Professors Robert Brown and James Cole, of the University of Nebraska were among the first to compile a long list deficiencies in this survey.CC Later Dr. Gregory Herek, of the University of California at Davis, expanded on these criticisms.DD
The survey was not based on a random sample. Dr. Cameron says he used a “conventional one-wave systematic area cluster sample”, claiming that it results in a random sample. But if his method was systematic, he doesn’t describe how he designed his cluster samples. There is no evidence of using census data or any other common demographic tools to select neighborhoods or households. If his methods were “systematic” he does not describe them anywhere, not even to answer his critics.
There are no apparent criteria for deciding which cities to select for the survey. He offers no demographic evidence that these cities offer a representation to match different regions of the country or the nation as a whole. He ignores rural areas altogether, as well as the entire Deep South, the Pacific Northwest, and the especially the highly influential urban centers of the Northeast. It’s hard to imagine this survey representing such diverse gay populations as Boston, New York, Philadelphia, Chicago, Miami, Seattle or Portland when none of these cities were selected for the survey. Surprisingly, he even omits the entire San Francisco bay area.
Perhaps in response to this criticism, Dr Cameron began in 1989 to describe his selection of cities according to “high (Los Angeles, Washington, DC), intermediate (Denver, Louisville) and low (Omaha) levels of homosexual activity as indexed by published homosexual guides to sexual opportunities.” (Notice that he dropped Bennett and Rochester in his list of cities, although he continued to use their data in his tabulations.) But he doesn’t describe which “homosexual guides” he used or any statistical measures used for selecting the cities.EE
In 1995, apparently in an effort to correct this problem, he begins to mention ”levels of homosexual activity as indexed by homosexual guides to sexual opportunities (e.g., Spartacus),”FF yet he is still unable to explain how each city is defined as high, intermediate, or low. And he still doesn’t mention Bennett or Rochester, even though their data is still included in the data set.
The data from the individual cities are not adjusted for population sizes. The data from all of the cities were simply added together without regard to the proportional sizes of the cities themselves. Thus, the 979 respondents from Omaha (metro area pop. 313,939) were simply lumped in with the 934 respondents from much larger Los Angeles (metro area pop. 7,477,239). Denver (metro area pop. 1,428,836) was represented by 970 responses, while similarly sized Rochester, N.Y. (metro area pop. 1,030,630) was represented by a miniscule twelve respondents.
There was virtually no rural representation in the survey. In the only rural contribution to this survey, tiny Bennett, Nebraska, (pop. 505) contributed 43 responses, which were completely swamped by those from the larger cities, giving rural America virtually no representation at all (assuming all of rural America is exactly like Bennett).
This is important because there is considerable evidence that community-based surveys are likely to give very different results from broader probability sample surveys. A recent study in London examined the results of the Gay Men’s Sexual Health Survey 2000 (GMSHS), which consisted of men recruited from London metropolitan gay bars and other businesses, as well as from STD clinics. These results were compared to those from the National Survey of Sexual Attitudes and Lifestyles 2000 (Natsal), which was a national probability sample survey which included small towns and rural areas from all across the U.K. When comparing the results from London’s GMSHS, which was not a probability sampled survey, to the results of Natsal, which was probability-sampled, several things jumped out. The GMSHS reported that men were more than three times more likely to attend an STD clinic than the men from Natsal, were more than fourteen times more likely to have had an STD in the past year, and were more than 5½ times more likely to have had at least one sexual partner in the past twelve months.GG
And this wasn’t just because gay men in London were more active than those in rural areas. Even when only London gay men in the probability-sampled Natsal survey were examined and compared to the GMSHS, many of these differences held roughly the same. But by not including small towns and rural areas in a statistically significant and valid way, Dr. Cameron’s survey is almost guaranteed to greatly overstate the activity of gay men and women overall.
The survey appears to suffer from a very low response rate. He ignores the standard measure of response rate, offering instead something he calls a “compliance rate,” which, due to a math error, he erroneously calculates to be 43.5% (it’s actually 47.5%, which he finally corrects in 1988HH, although he slips and cites the 43.5% figure again in 2002II). He didn’t include in his calculations the count of households in which contact was not successful (the “not-at-homes”). If he had, he could have provided a true response rate, which, based on the data he provided, appears to have been only 23.6%.
This is important for surveys which are intended to be probabilistically representative, including this one for which he claims to have conducted a systematic area cluster sample. On such surveys, the sample would be constructed ahead of time to match the general population, and “not-at-homes” or other refusals can alter the survey’s outcome in unknowable ways. With a “compliance rate” of 47.5%, the response was low; with an even lower response rate of 23.6%, we cannot have much confidence in the ability of the data to reflect the general population.
While we don’t know for certain the effect of the low response rate on the outcome of the survey, Dr. Cameron offers his speculation on what it may be. He acknowledged that the average age of the non-respondents was older than respondents, and the respondents were more educated and more Caucasian in proportion to the general population.
This leads us to one possible explanation for the low response rate, which is a common one with surveys like this: people are reluctant to answer highly personal questions to total strangers who knock on their door out of the blue. Many others are likely to be reluctant to take a personal survey left by strangers for them to mail in later. They may not trust the stranger’s assurance of anonymity. Dr. Cameron noted that during the survey process, “the police were called at least once in every city in which we interviewed, and it is our impression that sexually conservative people were more apt to refuse to cooperate.”
Dr. Cameron’s impressions are probably correct. But if the sexually conservative are less likely to participate, then the flip side of that observation is also true, that those who participated are likely to be more adventurous. This has the effect of exaggerating the reported “bad” or “kinky” behaviors. Yet Dr. Cameron is eager use that exaggerated data to paint an ugly picture of all gay men and women.
The questionnaire was excessively long and complex, consisting of a whopping 550 questions, which may have contributed to the low response rate. It is common for survey respondents to experience “survey fatigue” with long questionnaires. The fatigue problem in this survey is compounded by the complexity of some of the questions. One question had a daunting forty-four possible answers to choose from. This places an unreasonable burden on respondents to consistently read and consider each possible response before making a selection.
Many questions were ambiguous. For example, one question asked the respondent’s age at which various people “made serious sexual advances to me” without any further definition of the term “serious”. Some may consider a verbal invitation to be “serious”; others may consider nothing less than an aggressive grope to meet this definition.
Some questions may have been impossible to answer accurately. For example, one question asked, “With how many homosexual virgins have you had homosexual relations?” People typically don’t ask about the virginity of casual sex partners, making it impossible to answer this question accurately. Another particularly bizarre question asked how they felt about sharing toilet facilities with a homosexual, with one possible answer being “very positive, I’d enjoy it greatly”. Possible answers like this may keep some respondents from taking the questionnaire seriously.
There are many ways in which researchers can ensure that the subjects are answering the questionnaire truthfully. One way is to ask similar “check” questions in different parts of the questionnaire. This allows the surveyors to check the respondents for consistent answers. This was not done in this survey.
Dr. Cameron’s impartiality was in question while the survey was being conducted. As the survey was underway in Omaha, the local newspaper published a front-page article with the headline, “Lincoln Man: Poll Will Help Oppose Gays.”JJ In the article, Dr, Cameron predicted that the survey’s results would provide ammunition for those who are fighting gay-rights legislation — before the data was compiled and analyzed. Furthermore, he had no qualms about his bias, saying to the reporter, “all research is biased, and mine is no exception.”
Ironically, fifteen years earlier Dr. Cameron performed a study in which he determined that answers to questionnaires can be affected by interviewers who appear to endorse a particular outcome. In his study, he noted that “substantial changes in opinion (on the order of 13%) can be effected.”KK In other words, he knew that his actions would affect the survey results because he himself published an article which describes this very effect.
With all of this, it’s no surprise that the survey results are seriously out of whack. Based on this survey, Dr. Cameron claims:
- Gay men have a median of 10 annual partners and 100 lifetime partners.
- 85% have had an STD at some point in their lifetime.
- 41% have had a traffic accident in the past 5 years.
- 10% have committed murder or attempted murder.
While his survey first appeared in 1985 in the Nebraska Medical Journal ( “Sexual Orientation and Sexually Transmitted Disease”), we get the clearest glimpse into his most common statistics from his 1989 report, “Effect of Homosexuality Upon Public Health and Social Order” in Psychological Reports. According to his numbers, he only managed to get 42 gay men to respond, along with 25 lesbians. He sometimes combines the bisexuals with the homosexual population when its advantageous to do so, yielding higher numbers which vary between 77-89 men and 66-79 women, depending on the definition he chooses to use at the time. But regardless of his definition, no reputable researcher would attempt to characterize a broader population on such a paltry sample size.
If Dr. Cameron’s survey really was representative of the general population, then he should be able to offer calculated margins of error for his survey results. Using the highest number of gay/bisexual men in his articles, the best margin of error would still be 10.5% (at a 95% confidence level — the typical confidence level used for public opinion surveys). This makes most of the statistics from Dr. Cameron’s survey largely meaningless.
But the problem isn’t just with the small size of the gay population. Any single one of the methodological problems cited can skew the results wildly. Combined, and the survey becomes completely unreliable. In 2003, Dr. Cameron finally noted that “we did not and do not claim that our findings came from ‘representative sample of the American population’ nor … “can legitimately be generalized” to the U.S. adult population.”LL But he continues to downplay the many deficiencies in the survey, and cites it extensively again as recently as 2005MM without these caveats.
Paul Cameron and the Nebraska Medical Journal
You can read more about Paul Cameron’s hysterical reaction to criticisms of his ISIS survey in “Paul Cameron vs. Professional Ethics.”
But no matter. As a final indication of how worthless this survey is, I’ll let you look over just a few of the statistics from the much larger heterosexual population (which, if representative, would have a margin of error of only 1.5%):NN
- 52% of males have shoplifted. The figure is 36% for females.
- 38% of males have had a traffic accident in the past 5 years.
- 34% of males have committed a crime without being caught. The figure is 15% for females.
- 27% of males have contemplated suicide. The figure is 34% for females.
- 24% of males have had sex in front of others.
- 20% of females have obtained an abortion.
- 22% of males have been arrested for a crime.
- 17% of males have had sex in public.
- 16% of males have been in a physical fight in the last year.
- 13% of males were jailed for a crime.
- 12% have committed murder or attempted murder.
I will leave the question of this survey’s credibility up to you: how many of these statistics do you believe?
Please continue with:
Part 3: “Rectal Sex”. It’s not for the squeamish, but Dr. Cameron offers a bit of comic relief by passing on an amusing urban legend.
AA. Fenton, Kevin A.; Johnson, Anne M.; McManus, Sally; Erens, Bob. “Measuring sexual behaviour: Methodological challenges in survey research.” Sexually Transmitted Infections 77, no. 2 (April 2001): 84-92. Full text is available online for free at http://sti.bmjjournals.com/cgi/content/full/77/2/84.
Ellen, Jonathan; Gaydos, Charlotte; Chung, Shang-En; Willard, Nancy; Lloyd, Laura; Reitmeijer, Cornelius A. “Sex partner selection, social networks and repeat sexually transmitted infections in young men: A preliminary report.” Sexually Transmitted Diseases 33, no. 1 (January 2006): 18-21. This study drew participants from STD clinics in Baltimore and Denver. These men averaged 2.0 sex partners in the last two months.
Gorbach, Pamina M.; Stoner, Bradley P.; Aral, Sevgi O.; Whittington, William L.H.; Holmes, King K. “It takes a village: Understanding concurrent sexual partnerships in Seattle, Washington” Sexually Transmitted Diseases 29, no. 8 (August 2002): 453-462. In this study, male STD patients averaged 2.9 sex partners in the past three months; female STD patients averaged 1.7 partners in the past three months.
Mehta, S.D.; Erbelding, E.J.; Zenilman, J.M.; Rompalo, A.M. “Gonorrhoea reinfection in heterosexual ATD clinic attendees: Longitudinal analysis of risks for first reinfection.” Sexually Transmitted Infections 79, no. 2 (April 2003):124-128. In this study, 42.2% of STD patients reported having more than two partners in the past month.
Schwebke, Jane R.; Renee, Desmond. “Risk factors for bacterial vaginosis in women at high risk for sexually transmitted diseases” Sexually Transmitted Diseases 32, no. 11 (November, 2005): 654-658. In this Birmingham, Alabama study of women attending an STD clinic, the participants averaged 2.0 sex partners in the past thirty days.
DD. Gregory, Herek M. “Bad science in the service of stigma: A critique of the Cameron group’s survey studies.” In Gregory M. Herek (ed.) Stigma and Sexual Orientation: Understanding Prejudice Against Lesbians, Gay Men, and Bisexuals (Thousand Oaks, CA: Sage, 1998): 223-235.
GG. Dodds, Julie P.; Mercer, Catherine H.; Mercey,Danielle E.; Copas, Andrew J.; Johnson, Anne M. “Men who have sex with men: A comparison of a probability sample survey and a community based study.” Sexually Transmitted Infections 82, no. 1 (February 2006): 86-87. Abstract available online at http://sti.bmjjournals.com/cgi/content/abstract/82/1/86.
JJ. Flanery, J.A. “Lincoln Man: Poll Will Help Oppose Gays.” Omaha World-Herald (May 23, 1983): 1. A photocopy of this article is available at http://psychology.ucdavis.edu/rainbow/html/WH052383.html.