Two Common Statistical Mistakes About Marriage and Sex

In the previous post I said that every survey on the subject found that men had more sexual partners than women. I also implied that everyone bought this. It is, of course, impossible unless you think that men have a lot of sex outside the country, and women don't, or, of course, that a lot of men have homosexual affairs. Because it takes two to tango - so for every man having sex, there has to be a woman. So women have as much sex as men - over a large enough population. So if British women aren't having as much as British men, somewhere there's a country where women must be reporting more sexual partners than men. Except there isn't. The standard comment is that women might under-report and men over-report, but consider the effect of doing that for these options: (1-2, 3-5, 6-9, 10-15, 15+). The effect would be very obvious. You would also assume that the clever statisticians who analyse these things would have corrected for that. My favourite explanation is that there are a small number of women who pass through an intensely promiscuous period - but they are under-represented in the sample, which also don't capture just how many partners they had. You may live in a more prosaic world.

The other one is the divorce rate. There's a figure of 1 in 3 that's been going round forever, which comes from the fact that for a long time there were about 100,000 divorces and 300,000 marriages a year, which looks like 1 in 3. Except that divorces come from the pool of married people, and marriages come from the pool of single people. So you can't just divide one by the the other. National Statistics report that in 2008 the divorce rate was 11.2 per 1000 married people, with the highest cohort rate for 22-29, where 26 per 1000 married women got divorced. The marriage rate (for 2007, 2008 not available at time of writing) was about 20 per 1000 single people. In case you're wondering, in order for everyone to be married at thirty-five, given they can start at sixteen, the marriage rate would have to be a bell-ringing 71 per 1000. In 2006, half the men in the country were single.

I think marriage stats are fascinating. Start here for a comprehensive pdf and here for some directions from National Statistics.