survey result: women want sex more than men (a.k.a., selection bias)

Well, ladies and gentlemen, time for another round of “survey research and its problems,” as illustrated by our tiny little survey.

It turns out that feminine-identified (or female-identified; I put the two in the same category) readers of this blog who responded to the survey want sex a bit more frequently than the masculine-identified (or male-identified) readers of this blog who responded to the survey. (Due to small N and no non-arbitrary way to plonk them into either of the other categories, the genderqueer-identified folks were removed from this analysis. No disrespect intended, I just couldn’t report anything meaningful.) Like this:


Feminine      Masculine
6.5             5.37

(That number is the average number of time per week the person reported wanting sex, by their own personal definition of “sex.”)

This might go some way in explaining why there was some resistance to the idea that men, globally, have higher sexual interest than women.

Actually, I can tell you that the difference is made up entirely by women aged 25-45:


Age     Feminine    Masculine

18-24   5.92        7.33
25-44   7.375       3.75
45+     5.33        5.07

But that’s a lot of qualifiers: feminine and masculine identified. readers of the blog. who responded to the survey.

In excess of 1000 people VIEWED our “survey,” but only around 100 responded. 10%?? That’s a TERRIBLE response rate. There is very likely some difference between those folks who posted a response and the folks who didn’t, making the results ungeneralizable. I don’t know what differences (how can I know, unless they tell me, and by definition they won’t!)

Response bias is a continual problem in sex research. The kinds of people who are willing to respond to a sex survey may be different in important ways from the kinds of people who are not. Kinsey, clever biologist and specimen collector that he was (bless him), tried to overcome this problem by surveying entire populations: every member of Sigma Chi, every inmate at a prison, every student in a class, every man in a gay bar. 100% samples, to avoid sampling bias.

(People don’t do survey research this way now because there are inherent ethical problems of the potential for coercion!)

To make matters worse, there are very likely differences between the kinds of people who read a sex blog (or a blog that linked to a sex blog) and the kinds of people who don’t – precisely what differences I don’t know, but going by your responses to my religion post you’re a bunch of leftwing godless heathens with nothing but social justice and science on the brain, like me.

♥ Weirdos. ♥

I might also suggest that according to these results, women between the ages of 25-44 who responded to the survey want sex more often than women between the ages of 25-44 who didn’t, based on national results.

So when you read survey results, do bear in mind sample size, response rate, representativeness of the sample, etc etc etc.

Oh, and my next discussion of the results will be about the problems with the question itself and the coding of the responses. VERY exciting because it means everything about sampling bias might be canceled out because of the utter uselessness of the question I asked! Stay tuned.