Reply #29: I'd have to see a lot more on this, with more significant numbers

[jobycom]

He studied less than a thousand people to determine that the odds of being gay would increase by about 2% under the right conditions. That's a swing of about twenty people if his sample group is representative of the population as a while.

First, I'd have to see how he defined "gay." Does he simply accept that a person is gay if they claim to be, or is there some criteria? Next, what percentage of his sample was gay, how did he go about selecting his sample, and how does he arrive at the statistical standard that 3% of men are gay? If his sample isn't representative, then he hasn't proven anything much. If his entire sample is gay, then where does he get a standard non-gay number to compare? Etc.

Seems weak. But interesting. There are so many opportunities for sample bias, and so little statistical difference compared to the whole. (Compared to only gay men, the chances nearly double from 3 to 5, but compared to all men, the statistics barely move--thus his sample method and size are critical.) I'm sure the study figured all of this into the study, but how well they figured it in is still a question. This finding would have to be born out in other studies using other criteria, to mean much.

Seems to me, anyway. But I haven't looked at his study or the facts, and probably wouldn't know how to analyze them if I did.