Don Boudreaux’ latest post on CafeHayek is a well written post that is surprisingly not condescending towards a hostile writer (one that doesn’t understand statistics). This instantly reminded me of my time on Huffington Post trying to explain statistics to those who claim that homosexuality is genetic. I had to explain to a legion of hostile commenters the most basic statistical terms that they could have merely googled. If one twin is a homosexual, there is a 20% chance that their identical twin is also homosexual (defined as having at least one same-sex partner). The results seem to be even less favorable to the thesis that “homosexuality is genetic” if one wants to compare the number of same sex partners between twins.

I was first pointed to the most comprehensive twin study to date on the issue via Bryan Caplan’s “Selfish Reasons to Have More Kids” (which can be currently found on Scribd). He does not go into details about twin study results on homosexuality (possibly for ideological reasons):

…identical twins are more alike in their sexual orientation than fraternal twins. Yet genes are far from the whole story—if you’re gay, your identical twin is usually still straight.

But what are the statistics? Those who want to claim over 50% point to studies which suffer from extreme self-selection bias. But Caplan uses, for his main source of many of his claims, a Swedish Study which uses the entire twin registry of Sweden. This study is the largest, most recent, and most comprehensive twin study to date. The results are astounding:

For Male, Monozygotic twins (identical twins with same DNA), probandwise concordance was .18 for “any lifetime same-sex partner”. This would mean if one twin is homosexual, there is about a 20% chance their identical twin is as well. The study calculated the genetic effect at .39 (.00–.59), the environmental effect at .00 (.00–.46) and Unique Environment at .61 (.41–.85).

The results were quite different for females.

For Female, Monozygotic twins (identical twins with same DNA), probandwise concordance was .22 for “any lifetime same-sex partner”. This would mean if one twin is homosexual, there is about a 20% chance their identical twin is as well. The study calculated the genetic effect at .19 (.00–.49), the environmental effect at .17 (.00–.42) and Unique Environment at .64 (.51–.78).

Why is the Female probandwise concordance higher than male yet the genetic effect less? Remember, they are using Dizygotic (fraternal twins) as a control variable. Women had a higher rate than men for probandwise concordance in fraternal twins.

What this tells us is that no, homosexuality is not genetic. People with the exact same DNA only 20% of the time have the same “sexual orientation”. There may be a genetic component, but it seems to affect men more than women.

A leftist might counter that these numbers include individuals who only dabbled in same sex relations who are not actually homosexual, but then they have to merely look downwards on the chart and see the “number of same sex partners” statistics. The probandwise concordance seems to be much less still. The only two excuses remaining for the leftist is to either claim that non-responding twins represent disproportionately homosexual twins, or that Swedish twins are not representative of all twins.

For now, the statistics do not look favorably on those who claim homosexuality is genetic. The evidence says no.

It is also interesting to note that if one runs the numbers, the percent of people who are homosexual (those having at least one same-sex partner) equals about 5% according to this study. This is in liberal Sweden and is including people who do not necessarily self-identify as homosexual but just engaged in experimentation. This result is closer to the conservative claim of 3% of the population identifying as homosexual, as opposed to the leftist claim that a full 10% of the population identifies as such (a statistic derived from a study using a disproportionate amount of people from prisons).