On Sun, 12 Dec 2004, Yoshie Furuhashi wrote:
> I looked into the journal Personality and Individual Differences, and
> the issue in which Lee Ellis and Jill Hellberg published their
> article is available online.
Thanks to Yoshie for pointing this out; I got a chance to read the article. After statistically controlling for maternal education, maternal age, and self-ratings of maternal recall of drug use, only one drug--Prednisone--was related to male offspring's sexual orientation. This was due to the fact that two women in the sample reported taking this drug, and in this tiny sample, one of the men was gay, the other het. The authors correctly point out that the sample size here is much too small "for the results to be taken seriously".
For women, prescription diet pills and synthetic thyroid meds intake by mothers was related to offspring's sexual orientation. 37 mothers in the study reported taking the thyroid meds; 10 reported taking diet pills. (Again, subsamples are too dinky to draw any strong conclusions.)
In the discussion, the researchers emphasize that this is not a controlled experiment; the relatively small relationships could be due to other, uncontrolled factors (confounds). They conclude that "more evidence is needed before considering this a well- supported hypothesis". I agree that's a reasonable analysis of the data.
As Yoshie points out, the problem here is typical: magazine/ newspaper/web articles do not accurately summarize scientific research findings, and people then circulate these popularized misrepresentations of the research as "scientific proof" for their own political agendas. All hail Lysenkoism!
Miles