Yesterday, I posted a letter from the EPA president, Phil Hineline, and then a response from Paul Cameron regarding a March presentation by Paul and Kirk Cameron purporting to show that homosexuals comprise a smaller than assumed percentage of the population because older gays are not included in prevalence estimates. Further, the Camerons assert that the reason older gays are not figured in to the prevalence numbers is because most of them die young. That post did not discuss in detail the claims of the Camerons, but rather some aspects of the presentation and the subsequent dissemination of the study by the Camerons.
However, to address the actual claims of early demise, I asked Morten Frisch, Danish epidemiologist, to review the Cameron’s paper “Federal Distortion Of Homosexual Footprint (Ignoring Early Gay Death?). Morten is the lead author of a recent report on environmental influences on marriages decisions among heterosexuals and homosexuals. I wrote about this study here and blogged about it here. He very kindly agreed to do so and replied earlier today. As I suspected, he did not find their arguments compelling, or use of data appropriate. Here is his brief analysis:
Cameron and Cameron’s report on ’life expectancy’ in homosexuals vs heterosexuals is severely methodologically flawed
It is no wonder why this pseudo-scientific report claiming a drastically shorter life expectancy in homosexuals compared with heterosexuals has been published on the internet without preceding scientific peer-review (http://www.earnedmedia.org/frireport.htm). The authors should know, and as PhD’s they presumably do, that this report has little to do with science. It is hard to escape the idea that non-scientific motifs have driven the authors to make this report public. The methodological flaws are of such a grave nature that no decent peer-reviewed scientific journal should let it pass for publication.
As a measure of gay individuals’ average ‘life expectancy at birth’, Cameron and Cameron gathered information about age at death from obituaries for homosexual people in the U.S., and they obtained Scandinavian data regarding the average age at death among homosexually partnered persons who died within a period of up to 14 years after the introduction of laws on homosexual partnerships.
Due in part to reports like the present homosexual persons remain subject to stigmatization. The majority of homosexual people, even in comparatively liberal countries like Denmark, are not open about their sexuality in public. Particularly older homosexuals who grew up in periods when their sexuality was either a crime or a psychiatric diagnosis tend to remain silent about their homosexuality in public. Therefore, the higher prevalence of self-reported homo/bisexual experiences and feelings in younger than older age groups most likely reflects that young gays and bisexuals are less hesitant than older ones to provide honest answers in sex surveys.
The majority of homosexual individuals in the report by Cameron and Cameron were presumably open about their same-sex preferences. The groups studied comprised homosexuals who had entered registered partnerships in Denmark or Norway, and homosexuals in the U.S. whose relatives considered homosexuality to be such an integrated part of their deceased loved ones’ personalities that they felt it natural to mention in the publicly available obituary. Since, as noted, age is a strong determinant of openness about homosexuality, the study groups of deceased homosexuals in Cameron and Cameron’s report were severely skewed towards younger people. Consequently, the much younger average age at death of these openly homosexual people as compared with the average age at death in the unselected general population tells nothing about possible differences between life expectancies in gays and non-gays in general. All it reflects is the skewed age distribution towards younger people among those who are openly homosexual.
To further illustrate Cameron and Cameron’s methodological blunder, imagine a country that sets up a new register to record all cases of sexual harassment against women. After 14 years of operation the register is contacted by an advocacy group who gets access to the data to examine how sexual harassment influences women’s life expectancy. Among those women who died during the maximum of 14 years of follow-up, few women will have died after the age of 50, simply because most sexual harassment cases occurred among young women. Using the same logic and methods as Cameron and Cameron, this advocacy group could arrive at the conclusion that sexual harassment reduces women’s ‘life expectancy’ by 30 years or more. Needless to say, this would be as pure nonsense as the conclusion reached by Cameron and Cameron that heterosexuals outlive gays by 22-25 years.
In theory, despite their possession of academic degrees, the authors may have been unaware of the flawed methodology they used and, therefore, they may have been in good faith when writing their report. If so, they should promptly retract it to avoid further stigmatization of homosexual persons. However, expectations that this will happen are slim. Results simply fit too well with the views they have previously expressed.
Morten Frisch, MD, PhD, DSc(Med)
Senior epidemiologist
Copenhagen, Denmark
Thanks again to Morten for his comments. I hope to summarize my thoughts about this episode by early next week.