Thursday, December 21, 2006

That Guttmacher Study

There’s been some noise this week about the study which claims that 95% of Americans have had premarital sex by age 44, and that this percentage has been quite constant for 40 years. The Guttmacher Institute has a clear social/political agenda in all this, but let’s put that aside for the moment. People can have an agenda and still be accurate, after all.

(The immediate joke at Pajamas Media, of course, was that postmarital sex was also unchanged for the last 40 years at 5%.)

Numbers that dramatic should always raise an eyebrow. It pays to look at what is being measured, and how the data was obtained. In this instance, you will look long before you get the answer to some key questions about the data. After the phrase “surprisingly high refusal rates” popped up – in the context of my being unable to locate the actual number – I began to wonder whether this number was being downplayed, or even hidden. Perhaps not. But it’s hard to find. Refusal rates were 17%-27% for the various studies, averaging 21%. That is not an enormous number for survey studies, as many people just don’t want to do surveys about anything, let alone their sexual history, even for the $40 payment. It would be hard on researchers to expect a much better response.

But such high refusal rates, even if they are expected and common, should sound a cautionary note to researchers. If you are going to make even a tentative statement about 95% of Americans, you should be asking yourself “what about the 20% of Americans who wouldn’t answer at all?” If we could magically obtain that data, perhaps it would be similar to the four-fifths of the people who responded. Researchers do often try to equalize the demographic factors as mauch as they can.

But perhaps the numbers from that group are quite different. We don’t know. It would be easy to make up plausible explanations why people at extremes in any direction would avoid being questioned too closely. The study itself, in reviewing abortion data, goes out of its way to note that abortion data is generally underreported by about 50%, and to caution that people should be hesitant about basing conclusions on their data. Another pink flag.

Next we come to the question of what, exactly, is being measured? The study gives out that (round numbers) 75% of Americans have had premarital sex by age 20. From 20-44 years old, apparently, the total rises to 95%. The original 75% number seems high. But perhaps it’s true. If you are counting up everyone that has ever had sex, even only once or twice, maybe the number climbs that high. But let’s look at the second part of that equation. Of the 25% of the population who has not had premarital sex by age 20, (and note that this is quite a different population than the citizenry as a whole) four out of five of them will have premarital sex over the next 24 years? Don’t some of those people, like, get married over the next few years and stay married until age 44? To qualify as having had premarital sex, that previously chaste group of people would either have to be starting sex right away and then marrying, or having affairs. Does anyone think that is going to turn out to be true for 80% of those people?

Unlikely. So the group of people who get married and stay married are going to pull the average down a lot. Some percentage might have sex before their one marriage, or have affairs, but nothing like 80%.

And they are more than half the cohort we’re talking about. Most of those who marry, marry after 20. Those who marry later have lower divorce rates. Thus, this is more than half the remaining group.

This leaves the people who do not marry before age 44, or marry and divorce. Apparently, 110% of these people are having premarital sex. The people in wheelchairs, the developmentally disabled, the germ-phobics, all of ‘em. Seems unlikely.

Well, someone’s numbers are wrong here – perhaps mine. There’s something major missing from the data. Let’s pretend it’s my bad. Let’s give the Guttmacher Institute the benefit of all doubts here – the missing fifth of the population, the unlikely sudden behavioral change of 20 year-old-virgins, the greater-than expected number of marriages where both people are having affairs, the wheelchairs. Give ‘em all of it.

Oh, and make it retroactive for two generations.

Their goal in this, expressly stated in the study and the press releases, is to stop the government from providing abstinence-only sex education. The institute believes this is a terrible and unrealistic approach and wants to squash it. Their interest is in showing that simply everyone has premarital sex, therefore abstinence-only programs are unrealistic and even dangerous. Not so fast. The one does not follow from the other. Remember that we are counting in this population the people who have had unmarried sex even once or a few times over decades. You get counted in their data for any of it.

Unmarried 40-year olds who had sex twice in 1996 and once in 2002 do not consider themselves “sexually active.” And there is nothing to suggest anything about any type of formal sex education after age 20. No one attempts much of that or collects much data. Any data.

Plus, the original 75% that we breezed past early in the discussion – those kids who had premarital sex by age 20 – Guttmacher is counting the ones who had sex once or twice, and even the ones who had sex once or twice with their fiancee. Technically true, of course. But it’s hard to draw sexual safety conclusions at all from people having sex that infrequently. Theoretically, there’s risk every time. But there’s no actual data on what increases the sexual safety of that subgroup. The groups that we study tend to be highest-risk groups, and what works with them. (Which is fine, by the way. That’s who we should be studying. Go where the problem is. But there is nothing to indicate the results for the high-risk group bears on the general population.)

There are several good arguments for some safe-sex, rather than abstinence-only sex education. This study isn’t one of them. But this is exactly the sort of thing that will stick in the mind, and will give the impression of having proved the point.


bs king said...

Um, not to sound Bill Clinton-ish, but what definition of sex are they using? I'd imagine that those results would be much more realistic if they were including oral. Just a thought.

Woody said...

It's too late. With this study out, we'll hear its supposed findings over and over in the media whenever they want to oppose abstinence funding. Such studies take away any feelings of guilt to a group that believes anything is okay, and it throws up a wall against anyone who would make claims of immorality to this group that measures morality by a moving societal standard.

Anonymous said...

I'd also like to know how they define "having sex". If they are including heavy petting then that would be how they are getting their numbers. Otherwise, given my experience in volunteer abstinence education, I'd say they are falling for the nodding-head fallacy.*

I'd also like to point out that typically when we counsel young people against alcholism, we don't suggest that they carry wine coolers with them in case they encounter a situtation where they just can't stop themselves from drinking, and everyone knows that wine coolers are safer than other types of liquor.

Also, when we are counseling young people against violence, we don't usually suggest that they carry a knife just in case they encounter a situation in which they simply must express their violence, since we all know that knives are not quite as deadly as guns.

But any time that I have ever been involved in a debate about abstinence, or read an article about abstinence, there is always the "safe-sex sidebar" advising that we all know that you younguns just can't control yourselves and so you ought to learn how to use and make sure you carry a high-quality condom at all times - just in case, you know.

*The nodding head fallacy - when one person in a group is talking/bragging about her/his sexual adventures, most people in the group will start nodding their heads. It looks like agreement, but if you were to really pin people down, the real nodding-thought is "oh my god, please don't anyone ask me, pass me by, ignore me". I have informally proved this fallacy to many groups, by talking to them about sexual issues while nodding my own head, and describing this fallacy, and then pointing out that nearly everyone is nodding his or her head - does that mean that every one of you has sex frequently? [Horrified looks]

Do you remember the times you've had these conversations? Were you really banging the sports star five times a night, or were you hoping that if you blended in that nobody would ask you directly? Do you really think that everyone else who was nodding their head was nodding because they WERE banging the sports star, or because they were vigorously hoping no one would ask?

More often than not, a substantial portion of the group would grin sheepishly and mutter some kind of agreement with the second option. A wonderful way to counter "everybody's doing it" and reinforce the kids who are behaving well and hoping no one will harass them for it.

Assistant Village Idiot said...

I'm sorry if I didn't make that clear. Their definition was vaginal sex, and the questions were either answered to trained interviewers, on a laptop, or pencil and paper. If anything, that would depress the numbers, as it does for reported abortions. I didn't mention these because I was focussing on problems. This was one of the reasons I thought 95% was suspicious, though.

I don't doubt that the number is high, higher than we suspect or would like. But two things caught my eye and caused me to dig further. Three things, actually, because when I saw the Guttmacher name I suspected immediately that these were going to be advocacy numbers rather than unassailable numbers.

95% of anything is a brow furrower. Thirdly, I looked at the 75%-by-age-20 number and realized that the young people that I know were going to be considerably over-represented in the other 25%. Furthermore, I know the kids I went to school with myself. Kids at Christian highschools - at least in this neck of the woods - may not be in the 0-10% sexually experienced that the schools, parents, and students themselves would like to pretend, but they're not anywhere near that 75% range. I do suspect the percentage may rise pretty quickly after highschool, but nothing like that.

So. I know them, without necessarily making guesses one way or the other about anyone as an individual. People do surprise us. But i thought of that cohort, and considered what happened to them. As a lot of them get married, that sort of writes them out of the premarital sex statistic unless they divorce or are widowed or have affairs.

I think I'm repeating myself.