Statistics and Sex (Now Do I Have Your Attention?)

My battle with Carol Nichols Turoff will resume next week, but this week it's abstinence-only sex education. Shorter version for the statistically-impaired: If Mongo like result, Mongo claim to like "junk science." But Mongo don't know sex. As one of my law partners pointed out, if parents can't teach abstinence-only to their kids at home, how on earth are teachers supposed to do it at school?

For those of you following the special session on Child Protective Services--or especially, for those of you not following it--you should go to the following link and then contact your legislators using the links at the end.

TO ABSTINENCE-ONLY LEAGUE, IGNORANCE IS BLISS
East Valley Tribune, Nov. 9, 2003

What supporters of so-called “abstinence-only” sex education programs really want -- apart from the traditional adult pleasure of telling kids “No!” -- is for legislators to abstain from any knowledge of statistics.

The 2003 study of the Arizona Abstinence Only Education uses faulty data, invalid experimental design, a continually-changing program, and the absence of any test or “control” group to justify the program. You’d never know the limited, vague, and tentative nature of the study, because proponents’ press releases don’t have to be statistically valid -- and aren’t.

But before considering the wide gulf between what the 2003 study actually said and what abstinence-only supporters claim it said, it’s worth noting that every major sex education program (including comprehensive programs with statistically valid positive results) promotes abstinence as the most reliable method of preventing unintended pregnancy and sexually transmitted diseases.

In other words, abstinence isn’t only for abstinence-only. The real question is whether abstinence-only is effective, making sex the only school subject where less knowledge is somehow better.

The 2003 study has more limitations than my limited space allows, so let’s hit only the high points. First, the program kept changing through its five-year run, so study results may come from program aspects now revised, or jettisoned entirely.

Second, an accurate study would look for fairly long-term effects. The 2003 study surveyed program participants as little as three months later, which simply isn’t enough time to evaluate birth rates accurately.

Third, the study defines “success” as telling a pollster you have a “positive attitude” toward abstinence -- not that you intend to (or actually do) abstain.

Fourth, 45 percent of those surveyed participated in programs sponsored by Catholic Social Services (disclosure: I’m a CSS contributor), so positive results may come from participants’ religious inclinations instead.

Fifth, significant data in the 2003 report comes from prior years, and reports at the time explained that such data may have problems. Seemingly positive news came from "year 2" data, which appears to have overstated the number of program participants and understated birth rates. Years with better data collection methods showed birth rates among program participants basically matching state rates, and little evidence of positive change from the program.

Sixth, the most fascinating recommendation in the 2003 report comes from its observation that participants overwhelmingly thought that the program “talked too much about what was right and wrong.” The report authors assert that “nonjudgmental” abstinence-only programs, which cultivate interpersonal skills “are more likely to be effective than programs that are perceived as saying, ‘Do this because it is right.’” Somehow I doubt that “nonjudgmental abstinence-only” is what program boosters have in mind.

But the most serious flaw in the 2003 report -- the 500-pound gorilla, actually acknowledged in the report -- is the complete absence of any control group against which to measure the program’s results. As the study notes, “[i]n the absence of a comparison sample, or published findings from similar programs serving similar groups of teens, it is difficult to judge the merit of these successes.” Without a control group, it’s impossible to tell if abstinence-only is more or less effective than classes where similar students got lectures about the virtues of Vitamin C, or learned to play the trombone.

Maybe abstinence-only programs can have positive effects, but it definitely hasn’t been proven yet. And it’s odd that people who usually claim that government can’t do anything right, and needs to do and spend less, are suddenly enamored of devoting millions of your tax dollars on their unproven experiment.

If you actually read the studies, you’ll realize that when it comes to abstinence-only, the only statistically-valid thing to say is “Just say ‘we don’t know’.”