Abstinence-only education can work?

February 3, 2010 11:19 AM

A new study of middle schoolers found for the first time that abstinence-only education can help to delay [the students'] sexual initiation. While the evaluation adds important new information to the question of “what works” in sex education, the evaluated program was not a rigid "abstinence-only-until-marriage" program of the type that, until this year, received significant federal funding, leaving intact existing evidence that abstinence-only-until-marriage programming that met previous federal guidelines is ineffective.
posted by lunit (55 comments total)

I don't think this is news. I've heard/read a number of times that abstinence-only sex ed does tend to make kids wait longer until they have sex, but then they are less likely to use protection when they do have it.
posted by orange swan at 11:25 AM on February 3, 2010 [1 favorite]

The statistics I'm more interested in are the unintended pregnancy and STD contraction rates.
posted by ErWenn at 11:25 AM on February 3, 2010 [11 favorites]

File under "no shit".

The criticism of abstinence-only programs was not that they encouraged abstinence, but rather that they were lacking in supplemental education. "Abstinence-focused" is fine, as long as they still teach the basics, and explain the importance of contraceptives/prophylactics.

Despite what fundamentalists might believe, I find it unlikely that there are any sexual education programs that don't encourage waiting. Even in the most liberal enclaves of America, a teacher would be sacked nearly instantly for encouraging kids to get it on earlier. It's always been, "wait if you can, be careful if you can't."
posted by explosion at 11:28 AM on February 3, 2010 [7 favorites]

I think the "newsy" bit is the fact that the reporting on this study has been terrible. Victory for abstinence-only education! the press release junkies crow.
posted by muddgirl at 11:28 AM on February 3, 2010 [2 favorites]

No one really opposes teaching abstinence. What people oppose is only teaching abstinence. It's not between abstinence-only and condom-only. It's between abstinence-only and comprehensive sexual education. Students should be taught the risks of sexual intercourse and that delaying it is usually a good idea. They should also be taught how to use contraceptives and how to prevent STDs.

Also, from the abstract:

"Fewer abstinence-only intervention participants (20.6%) than control participants (29.0%) reported having coitus in the previous 3 months during the follow-up period."

How long was the follow-up period from the course? If it was more or less than 3 months that seems like an odd question. Furthermore, they only compared to the control. How much more or less effective was abstinence-only than comprehensive sexual education?

"Abstinence-only intervention did not affect condom use."

Compared to what? The control or the comprehensive classes? Did it not affect the rate of condom use or did it not affect its effectiveness? It seems the study did not determine pregnancy rates. I would expect that the students taught how to use a condom would have lower pregnancy rates even if their rate of condom use was the same as the abstinence-only group because they would be more likely to use the condoms properly.
posted by jedicus at 11:31 AM on February 3, 2010 [2 favorites]

In other news, authorities confiscate cigarette lighters in attempt to reduce lung cancer rates.
posted by Mental Wimp at 11:32 AM on February 3, 2010 [1 favorite]

If girls realized the consequences of having sex, nobody would be having sex.
posted by Joe Beese at 11:34 AM on February 3, 2010 [1 favorite]

Note the language and drive of the study: they want to stop sex, not prevent pregnancies.
posted by Malor at 11:34 AM on February 3, 2010 [10 favorites]

I heard this story on NPR yesterday. I'm pretty sure that they said something like, "One of the reasons that some abstinence-only advocates don't like the program is that it teaches about contraception."

In what sense is it "abstinence-only education", if it teaches about contraception?

Maybe I just misheard.
posted by Flunkie at 11:36 AM on February 3, 2010 [2 favorites]

A rigid "Homosexuality-only-until-marriage" program would be 100% effective against unintended pregnancies, why don't they teach that?
posted by wcfields at 11:36 AM on February 3, 2010 [37 favorites]

Participants A total of 662 African American students in grades 6 and 7.

Anyone else see a problem here?
posted by Joe Beese at 11:36 AM on February 3, 2010 [2 favorites]

I am picturing a bunch of abstinence only proponents pointing at pro-sex-education people and saying "Now you're fucked".
posted by srboisvert at 11:37 AM on February 3, 2010

No, I think I heard correctly. Here is the transcript, which includes the following:

Peter Sprigg, a senior fellow with the Family Research Council, which supports abstinence-until-marriage programs, says that he hopes the new study will encourage the Obama administration to reconsider funding abstinence programs. But that doesn't mean he would support discussing contraception with teens, which this program does.

posted by Flunkie at 11:38 AM on February 3, 2010

Sarah Brown: CEO of the National Campaign to Prevent Teen and Unplanned Pregnancy and Meredith Schonfeld-Hicks: State strategies program manager for Advocates for Youth talked about this study on Minnesota Public Radio's Midmorning program yesterday
posted by Think_Long at 11:38 AM on February 3, 2010

If girls realized the consequences of having sex, nobody would be having sex.

Yeah, I mean, who wants to be on the cover of "People" next to Jon and Kate?
posted by King Bee at 11:39 AM on February 3, 2010 [2 favorites]

I had no trouble delaying my sexual initiation. No trouble at all. AT ALL. Thanks for asking.
posted by unSane at 11:41 AM on February 3, 2010 [9 favorites]

acne + christianity + membership of computer club pretty much does it
posted by unSane at 11:42 AM on February 3, 2010 [2 favorites]

>The criticism of abstinence-only programs was not that they encouraged abstinence, but rather that they were lacking in supplemental education.

The abstinence-only program involved did not provide information about condoms and contraceptives, but neither did it demonize them, which I think is key.

Frankly, as a sex educator, my interest lies in using programs that are effective, and then replicating the evidence-based curricula faithfully. Faithfully means no tweaking this or that to make it fit your version of what's OK.

This sounds like it is decently evaluated and works, at least for African American 6th and 7th graders in an urban setting who are willing to participate. If that means we get communities who are opposed to comprehensive sex ed to implement this curriculum (again, faithfully!), then I'm all for it.

>But that doesn't mean he would support discussing contraception with teens, which this program does.

See, now that's just weird and contradicts what I said above. I wonder what they mean by that? By definition I'd think that discussion about contraception would make the program fall into the comprehensive sex ed category. Maybe they talked about contraceptives as something the participants would need to learn about later?

I want to get my hands on the full study. Does anyone have access to it that would be willing to help me out?
posted by Stewriffic at 11:46 AM on February 3, 2010 [1 favorite]

Things that bother me about the study:

1) I would expect kids who are in the class where they actually talk about sex to be more willing to admit that they had sex. I would expect kids who are in the class where they don't really talk about sex to be less willing to admit that they had sex.

2) The abstinence only classes were one weekend, and (presumably) the more comprehensive classes were longer than a weekend. Not a fair comparison.

3) All it takes is one sex partner to know about condoms for a condom to be used. How many of the abstinence-only kids had sex with another abstinence-only kid? If they all ended up hooking up with kids who got schooled about condoms, then no shit condom usage didn't go down.
posted by 23skidoo at 11:49 AM on February 3, 2010 [3 favorites]

I'm taking an Applied Quantitative Reasoning class and I've already learned enough about research experiment design to spot non-rigorous results or ineffective communication of those results. Now I can actually explain why I don't trust the statistics!
posted by sciurus at 12:07 PM on February 3, 2010

I don't have full-text access here, so I can't comment on the raw data, but I see these two bits cited in the abstract:

Participants A total of 662 African American students in grades 6 and 7.
Results ... The model-estimated probability of ever having sexual intercourse by the 24-month follow-up was 33.5% in the abstinence-only intervention and 48.5% in the control group.

Um. OK. So you've concluded that your (extremely non-representative) sample (whose 85% enrollment rate at the end of the study period is low enough that you should be seriously doubting the validity of your data) of eighth and ninth graders are less likely to have engaged in any sexual shenaninganry between the ages of 12 and 14. And now you've decided to publish with a title that's indicative of having proved the global efficacy of abstinence-only sexual education? Where's the reporting of the actual adverse outcomes? The STDs, the unwanted pregnancies? Hell, where's your reporting on the rates of sexual activity from your group once they're in the age range where most kids start having sex?
posted by Mayor West at 12:08 PM on February 3, 2010 [6 favorites]

But this still doesn't solve the problem. We still need to do a better job of getting people to use birth control WHEN (not if) they start having sex. The fact that abstinence-only and more comprehensive classes both have identical condom use rates isn't anything to be happy about. They're both too low. And one thing is for sure - the abstinence-only approach is certainly not going to be MORE successful in getting people to use birth control.
posted by yarly at 12:09 PM on February 3, 2010

No one really opposes teaching abstinence. What people oppose is only teaching abstinence.

It's the same kind of thinking that suggests that if you aren't anti-choice, you must be pro-abortion. (not to bring that subject into the thread...)

It's a way of thinking where there are no shades of gray, and everything to one side of their ideal moral purity is wrong.

It continues to frustrate me because it is still allowing them to frame the debate. Yes abstinence can be a good thing, but how about we let the kids know all the stuff they should do when they are ready to not be abstinent. That way they are ready.
posted by quin at 12:11 PM on February 3, 2010

From the abstract Results: No other differences between interventions and controls were significant.

What I assume this means, and I realize that this is an assumption, is that the rate of STDs and pregnancy was not higher in the abstinence-only group. This has always been the weak point for abstinence-only, that the participants overall have higher rates of disease and pregnancy.

But what the abstract does not say is what the comparison is to. The rates of sexual intercourse are obviously calculated based on the number of participants in each arm, but is that also the denominator for things like STD and pregnancy rates? The denominator should be the number of people who reported sexual activity, which is a very different number for the two groups. At least as written the overall rates could be the same, while the rates specific to those who had sex could be significantly different. I wish the abstract gave more information.
posted by OmieWise at 12:18 PM on February 3, 2010

Participants A total of 662 African American students in grades 6 and 7.

Anyone else see a problem here?

Could you be a leeeetle more specific?
posted by desuetude at 12:22 PM on February 3, 2010

Participants A total of 662 African American students in grades 6 and 7.

Anyone else see a problem here?

There should be a colon after Participants.
posted by jimmythefish at 12:23 PM on February 3, 2010 [7 favorites]

Yeah, that's the ticket...
Tell school kids and teenagers to NOT do something...

What do you think they'll do first??
posted by Drasher at 12:27 PM on February 3, 2010

They can't touch the effectiveness of Blizzard's abstinence program.
posted by mullingitover at 12:27 PM on February 3, 2010 [4 favorites]

"Fewer abstinence-only intervention participants (20.6%) than control participants (29.0%) reported having coitus in the previous 3 months during the follow-up period."

Eesh, I don't think I've ever reported coitus.

Okay, presumably the follow-up questionnaire was less clinical in word choice. But use of the world "coitus" in the study seems to indicate to me that they mean only PIV sex. Most kids don't start out with coitus, they start with other sex acts, so it seems strange to trumpet the delay of sexual initiation.

Does the full text say more?
posted by desuetude at 12:31 PM on February 3, 2010

Shoddy reporting like this is why newspapers crying poverty get no sympathy from me. Thanks, I can read press releases for free.
posted by empath at 12:32 PM on February 3, 2010 [1 favorite]

A friend/colleague of mine who is a pediatrician specializing in adolescent medicine has offered to get me a full text version. Can't wait to read it.
posted by Stewriffic at 12:34 PM on February 3, 2010

Participants A total of 662 African American students in grades 6 and 7.

Anyone else see a problem here?

Not _A_ problem, no.

The study also indicated that it broke that sample set up into 5 groups. So they were comparing groups of about 130 folks, 110 after the 15% attrition.

So when they say "Fewer abstinence-only intervention participants (20.6%) than control participants (29.0%) reported having coitus in the previous 3 months during the follow-up period" it's hard for me to get worked up about a 9 person difference. I've had 3 month celibacy periods in my life that had nothing to do with MY decision to go without sex. Maybe those 9 folks have the same quality skin I did as a teen.
posted by phearlez at 12:37 PM on February 3, 2010 [1 favorite]

The Young Turks had an interesting breakdown on this study yesterday, how the pro-abstinence movement will twist what this study means, and how bad the reporting has been.

Re: 6th & 7th graders in the study, this means starting with 11-13 year olds. Two years later when the study gets back to them, those kids are 13-15.

But if the average age that young people lose their virginity in the US is around 16+... how much effectiveness are they measuring?
posted by Pirate-Bartender-Zombie-Monkey at 12:48 PM on February 3, 2010 [5 favorites]

Oh, and trust the survey results you get from grade-schoolers and high-schoolers, too.

Implicitly!
posted by Drasher at 12:56 PM on February 3, 2010

I say we teach kids outright lies about sex that will confuse them so greatly they'll be put off it for years. Like that the wang goes into her belly button and that having her poop in his mouth is an effective contraception technique.
posted by GuyZero at 12:58 PM on February 3, 2010

It's often very difficult to do true experimental studies in education because of the legal, ethical, and financial hoops involved. So usually what you end up with is pseudo-experiments done with convenient populations with lots of caveats that the conclusions may not generalize.
posted by KirkJobSluder at 1:10 PM on February 3, 2010

I say we teach kids outright lies about sex that will confuse them so greatly they'll be put off it for years. Like that the wang goes into her belly button and that having her poop in his mouth is an effective contraception technique.

"I want to poop back and forth."
posted by aught at 1:13 PM on February 3, 2010

I heard a report about this on my local NPR station and one question I had was "how will they ensure the quality of educator?" which is a huge chunk of the success of a program. A mediocre educational package with a rockstar educator is more likely to be successful than a mediocre educational package with a mediocre educator.
posted by plinth at 1:22 PM on February 3, 2010

It's important to note that this study only looks at benefits of a modified abstinence project and not any costs. Not only does it ignore STI rates, but it's worth asking what kinds of sex these kids end up having when they do have sex. Do they feel happy with their bodies and the choices they make? Do they have enough information to say yes when they want to say yes and no when they don't? Are they ashamed of themselves and does that shame lead to negative outcomes?

I'd rather have someone have positive feelings about sex that is ultimately a healthy expression of their sexuality 23 months after a comprehensive sex ed class than have someone wait exactly 24 months and one day before beginning a lifetime of unhappy and unhealthy sexual encounters.

Just as importantly: comprehensive sexual education is a universal human right. The ability to know and understand one's body is something we cannot deny to people for no reason other than the fact that we don't like what they may do with the information. While there may be many avenues of discovery and information, using the public schools to mislead and lie to children about something so important is an affront to any decent society. The program in question did not shame and did not lie to individuals, which already puts it way outside the average ab-only program. I'm not surprised there are limited findings that show a few positive outcomes. But there are costs that come with not educating our children about their bodies. It may not need to happy in 6th grade but there must be a comprehensive program developed that does address these things over time.
posted by allen.spaulding at 1:27 PM on February 3, 2010 [3 favorites]

happen, not happy.
posted by allen.spaulding at 1:28 PM on February 3, 2010

...abstinence-only education can help to delay [the students'] sexual initiation...

Yeah, maybe until last bell.
posted by chillmost at 1:53 PM on February 3, 2010

The statistics I saw from CNN were 33% had sex after abstinence only, and 54% of safe-sex only education. So the question is how much unprotected sex was had in reality. It's certainly inaccurate to say that the 33% were all having unprotected sex, while all of the 54% were protected. It also says nothing of the attitudes, subjectively healthy or unhealthy, that were fostered by these different programs in regards to sex in general.
posted by karmiolz at 2:19 PM on February 3, 2010 [1 favorite]

NPR's coverage of this made me irate. They buried the actual nugget of useful information: that this program was more successful than no education at all.

NO FUCKING SHIT. I thought it was pretty obvious that the debate was not between abstinence education or NO education, but rather between abstinence education and comprehensive education.
posted by odinsdream at 2:48 PM on February 3, 2010

This study — and all the media hype built upon it — is built of pure fail.

Which wouldn't be that big a deal, if it weren't that idiots are going to make failing policy decisions based on it. Talk about setting up society to fail.
posted by five fresh fish at 3:59 PM on February 3, 2010

Small sample size, hardcore response bias, borderline statistical significance, no external generalizability what-so-ever. Meh, meh, meh. That is all. Also, meh.
posted by drpynchon at 6:12 PM on February 3, 2010

With any randomized experimental design, you don't have external generalizability, do you?
posted by scunning at 7:12 PM on February 3, 2010

You taught kids that they should answer 'no' on a survey question, and lo they answered 'no' on a survey. Come back with hard outcomes.
posted by a robot made out of meat at 7:22 PM on February 3, 2010

With any randomized experimental design, you don't have external generalizability, do you?

There's no such thing as perfect generalizability. That said, relative generalizability depends on how closely the study cohort mirrors the population of interest (ie this study might apply to black 6th and 7th graders best, non-black 6th and 7th graders next, and so on), but more importantly in this case, how the intervention reflects what might go on in the real world. While you could perhaps use the same curriculum, the findings in this study have no generalizability with respect to other abstinence-only curricula. In other words, this study is about one particular intervention. Not about abstinence-only curricula in general, which is what the press might run with.

Also, as scunning noted, self reporting of sexual activity is highly prone to response bias. If you teach kids that abstinence is the way to go, then when you come back in 2 years they're far more likely to tell you they've been abstinent than if this wasn't a hallmark of your intervention.
posted by drpynchon at 7:34 PM on February 3, 2010

I mean as robot made of meat noted.. Carry on..
posted by drpynchon at 7:35 PM on February 3, 2010

Thanks drpynchon. The generalizability issue is real, but randomized trials like this is the best we have for studying the efficacy of a curriculum. We now have hard evidence of a causal effect. Assuming away the other problems (which I agree is not a good idea, but I'll take that in a second), if this result held up in other trials, I would feel comfortable saying that we have evidence that this particular kind of abstinence program reduced sexual debut among middle schoolers. Would it work with Whites like I did with Blacks and/or would it work with all Black children in all parts of the country? Til we see more studies like it, we won't know, but this result makes me want to know.

The real threat to this study is not, in my opinion anyway, the generalizability (or lack thereof) but of what you call the response bias. If children are systematically less likely to respond truthfully because of taking the treatment, then that could explain the results.

My research is on sexual behavior (here), and response bias is basically unavoidable. I sometimes work with aggregate proxies for sexual behavior, like sexually transmitted infection incidence, and sometimes with survey data. But what I mainly like about this study is the careful treatment evaluation that it appears to have. The one thing about it that concerned me was that the effects struck me as too large, though. That seemed fishy, and I wonder if it would hold up under scrutiny.
posted by scunning at 6:47 AM on February 4, 2010

I agree with you almost entirely scunning. The biggest issue is the response bias potential which is very hard to try to measure, and off the top of my head, I'm not sure what methodological approaches to this question could correct it.

In this particular case though, I personally wouldn't be shocked to find that about an additional 10-20% of sexually active 6-7th graders might be less than truthful about their sexual history after this type of intervention. Again that's just my guestimation, but if it were true, it would account for just about the entire effect that the researchers found. So in a way, the effects actually aren't large enough to phase me because of my strong suspicion for bias.

Other food for thought: why go to the trouble of conducting a randomized trial, and splitting your subjects into 5 groups, when you're essentially powered only to make one valid statististical comparison between two groups? The p-values and such that they report should be corrected for multiple comparisons (which they weren't) even for the primary outcome. Had this been done, it would have crushed the authors' stats.
posted by drpynchon at 7:49 AM on February 4, 2010

They claim that the social desirability bias is minimal, because the abstinence only group does not report that abstinence is more desirable, which makes no real sense. Plus they made the kids PROMISE to tell the truth.

You think that their effect sizes are too large? The RR on past 3 months is 0.94 (0.90-0.99), which just barely touches nominal significance. The RR for initiation is smaller, but the upper CI limit is still 0.96. Small studies + publication bias means that you get exaggerated effect sizes.
posted by a robot made out of meat at 7:50 AM on February 4, 2010

I heart statistics nerds!
posted by five fresh fish at 8:14 AM on February 4, 2010

Other food for thought: why go to the trouble of conducting a randomized trial, and splitting your subjects into 5 groups, when you're essentially powered only to make one valid statististical comparison between two groups? The p-values and such that they report should be corrected for multiple comparisons (which they weren't) even for the primary outcome. Had this been done, it would have crushed the authors' stats.

The bottom line is that this produced no statistically significant result. That it was on an endpoint of suspect utility only emphasizes the uselessness of the result, as it wasn't even faux statistically significant on any useful endpoints (actual sexual activity, condom use, STI, pregnancy). I second the "full of fail" evaluation. Nonetheless, you watch, the right-wing slut haters will trumpet this result as an "I told you so" for years to come.
posted by Mental Wimp at 10:01 AM on February 4, 2010

I tried an absinthe-only education for a while but I red tulips green fairies madeleine.
posted by kirkaracha at 11:09 AM on February 4, 2010 [3 favorites]