Monday, November 17, 2008

CWA and data analysis

EDIT: I really should show my work on this. Some of the numbers I got are ridiculous. That said, I really don't need to use numbers, since supporters of reparative therapy are a minority among psychological professionals. EDIT ENDS.

There is much, I am sure, to critique in this article posted on the CWA website, but I am first and foremost a man of SCIENCE.

The article mentions two studies, one that claims that of some sample of identical twins in which at least one was gay, the other was gay "only" 52% of the time. (Oddly, Shaun Waymire didn't quote the same article that provided the 52% figure on significantly lower numbers... right next to it, and given a higher credence. CWA doesn't have a problem with quote-mining, but it's seriously counterproductive to use it to skew quotes against your argument. The numbers are, in fact, lower than the lowest ones (out of two) that he attributes to the next article. I frankly can't figure out what he's doing.) The other figure, mentioned in the preceding parenthetical rant, is 20%, from the Australian twin registry.

The articles in question all crow that the lack of a 100% perfect correlation refutes the idea that homosexuality has a genetic component. Let's examine this in an extremely condescending fashion for a moment. If there is no genetic component, then identical twins would be just as likely as any pair of people to both be gay. This means that, given one gay person, and, indeed, not given a gay person, any random person has a 52% chance of being gay. In other words, gay people are common enough to form a majority bloc and control all elections ever.

Wow! CWA was right about the gay agenda! I mean, if 52% of the world's population is gay, I guess we really can't be a persecuted minority.

Let's add some rigor. The aforementioned study producing the 52% figure was flawed, but the data still kind of means something.
"The study found that 29 of 56 pairs (52 percent) of the identical twins were both homosexual; 12 of 54 of the fraternal twins were both homosexual and six of 57 of adopted [brother]s were both homosexual." Percentages: 52%. 22%. 11%. (The following figures reflect the genes that vary within the population as a whole.) Identical twins share 100% of genes and an environment. Fraternal: 50% of genes and an environment. Adopted: A theoretically negligible percentage of genes, and an environment.

Comparing just twins, heritability squared is equal to twice (the percent correlation of identical twins minus half the percent correlation of fraternal twins). Crunching all of the numbers gives a heritability of 91%. (This is much higher than the percent concordance because differences in environment between siblings decreased concordance.) Unfortunately, this produces some utterly nonsensical values (44i%) for other statistical measures, so the study doesn't really say much of anything.

Which means I was joking about gay people's being a majority. And Shaun (and the articles he cited) was wrong about what these studies say about homosexuality and genetic basis. Sadly, I'm not dedicated enough to get the Australian concordance data they mention, so I'll mention only that, should the 20% figure hold, there are more gay non-black people in the US than there are straight black people.

That's all for now.

(Later that week, an addendum.)

Another possible explanation for the high concordance rates, aside from heritability, is that the environment produces a tendency towards homosexuality (provided it's not solely our sinful choice that will destroy our lives via vaguely insincere-sounding composite life sketches). In which case, in order to hold the Homosexual Menace™ in check, you'd think they'd be compiling lists of acceptable and unacceptable traits in a family environment. Then, to combat the 30% heritability (I don't have all the data that they used, but that sounds, you know, about right, maybe.) that *shudder* NARTH reports, they could proscribe the proper way to raise a child so as to avoid X% (Where X depends on who's quoting survey data.) of the next generation plunging into depravity and disease. But what about the families, you may ask. Are they to be explicitly told how to raise their child, and make no decisions on their own? If that is what it takes to save the CHEEEEELDREN from the grasp of my pernicious subculture, then so be it, right, CWA? I mean, you totally support the idea of legislating how people can relate to each other, yes? Remember, it's not big government to pass laws that require more governmental effort and experience to enforce.

Not if it's for the children.

No comments: