US Virgin teens 'have same STD rate'

[emad]

From BBC News:

Young Americans who pledge to remain virgins until they marry have the same rates of sexually transmitted diseases as those who do not, a new study says. Teenagers who take a public vow to abstain from sex have fewer partners and get married earlier. But they are much less likely to use condoms, the research found. "It's difficult to simultaneously prepare for sex and say you're not going to have sex," the study's author Peter Bearman told the AP news agency.

The research, which is being presented at the National STD Conference in Philadelphia, examined the sex lives of 12,000 adolescents. The data was gathered from young people aged 12 to 18 who were questioned again six years later. According to the study, the STD rates were:
Whites who pledged virginity 2.8% - did not 3.5% Blacks: pledgers 18.1% - non-pledgers 20.3% Asians: pledgers 10.5% - non-pledgers 5.6% Hispanics: pledgers 6.7% - non-pledgers 8.6%

The study's authors say that from a statistical point of view, the rates are the same for both groups. "The message is really simple: 'Just say no' may work in the short-term but doesn't work in the long-term," Peter Bearman of Columbia University's Department of Sociology said. Critics of abstinence-only education said the findings showed that adolescents benefit from sex education. "It's a tragedy if we withhold from these kids information about how not to get STDs or not to get pregnant," Dorothy Mann, of the Family Planning Council, told the Associated Press.

But promoters of abstinence argue that telling young people about condoms and other forms of contraception increases sexual activity - and as a result increases the rate of teenage pregnancy and the transmission of sexual diseases. US President George W Bush has massively increased funding and support for the US sexual abstinence movement during his term in office. Christian abstinence organisation True Love Waits says it has more than one million card-carrying young members. On 13 February, thousands of American teenagers celebrated a "Day of Purity" to promote sexual abstinence.

http://news.bbc.co.uk/1/hi/world/americas/3546007.stm

Just say no to George, kids....stuff the abstinence

[THUNDER HANDS]

1. boy Asians are really getting screwed

Pardon the pun. Their rate is way off the charts. 10.5% among virgins and 5.6% against non-virgins.

Seems kinda odd to me.

[emad]

4. Memory plays strange tricks after 6 years' abstinence....

and kids love bluffing about their sex lives

[Snow]

7. Likely a very small sample and high variance.....

probably if you put a confidence interval around those rates, they'd be the same as the rest.

[oldcoot]

33. The numbers

It would be nice to know how many members of each racial category participated in the study. It also would be nice to know more about the backgrounds of the participants in this study. Did all of the kids who participated in this study live urban areas? How many of these kids were immigrants and or had parents who were immigrants? I would like to know what, if any, role culture played in this study.

[Beetwasher]

2. Abstinence Only Makes Teens Ignorant

IOW, the teens state publically they will not have sex, but do anyway and are uneducated about protecting themselves since they ONLY learned abstinence...That's why the numbers are the way they are...

[salin]

15. I think it goes beyond ignorance.

The message of abstinence also include heavy doses of shame and guilt. Thus there is the psychological barrier... that even thinking about protection (and they would have to live in a vacuum not to know that there is such a thing as protection)... is to mean one is thinking about having sex which produces huge levels of guilt. Thus - it is easier to disassociate the thoughts in the lead up (that is - attempt to pretend one isn't having those feelings and that one isn't going to act on them... ) and then *boom* not have protection in the moment, than it is to recognize that they are about to break the "vow" and to prepare for that event.

This was alive and well for girls for eons. Good girls don't have sex. Good girls don't bring up things like asking about protection, because that sendss the message that they are not good girls. In the end many of those 'good girls' ended up being sent a way to 'homes for unwed mothers' or in shotgun early marriages.

[Beetwasher]

18. Exactly and Also

The same institutions that are unrealistically teaching ONLY abstinence and instilling shame and guilt about perfectly natural sexual feelings are also ignoring teaching young women the proper coping and interpersonal skills for dealing w/ the social pressures to have sex...Teaching a young woman/girl to be abstinent is fine, as long as she is also taught the proper skills and also has access to info about STD prevention and birth control, just in case...I have no problems w/ teaching abstinence as long as it's part of a bigger picture and done properly...Instead, many of these abstinence only programs are essentially telling young women "just say no!" without anything else, or telling them to pray for guidance, and that just doesn't work...

[SemperEadem]

32. exactly!

"But promoters of abstinence argue that telling young people about condoms and other forms of contraception increases sexual activity - and as a result increases the rate of teenage pregnancy and the transmission of sexual diseases. "

Telling teens about contraception does not increase sexual activity... they were already sexually active to begin with, they just didn't admit it. Education only increases their attention to make sure that if they are going to be active, they take the proper precautions.

Sex is the second strongest drive in the human being and ignoring it because of Puritanical FEAR and IGNORANCE does not make it go away. Guiding the teen in how to manage their feelings and whatnot is far more constructive than pretending (and insisting in the moment) that those feelings don't exist; if all you've got to offer is "don't think about it", then teen std and pregnancies amongst those who vowed to abstain are going to skyrocket because they're too ignorant of the facts of their bodies to know the correct way to proceed.

[BlueEyedSon]

3. I'm soooo confused....

How do you get a "sexually transmitted disease" if you're not having sex? Toilet seats?

:crazy:

[enki23]

5. they are having sex

they are "pledged virgins." not all of them, obviously, are *actual* virgins.

[BlueEyedSon]

6. ROTFLMAO! (see link)

http://www.technicalvirgin.com/

[veganwitch]

8. non-penatrative sex,

oral for guys and girls, manual stimulation can spread warts, herpes. not washing hands or toys before and after, not using barriers - gloves, dams (saran wrap if you in a fix) can spread infections.

[emad]

11. Just remembered: the cat bit my son Joe's penis when he was

about 14 and Joe had to have tetanus shot, anti NSU-medication AND a circumcision.....Feline genital herpes is tramsittable apparently....

[kath]

24. whoa -- why did they have to circ him??

That seems a bit much, IMHO (as an MD) -- unless the foreskin was infected, very swollen, etc. (The tetanus shot, etc I can see)

[NickB79]

39. I just have to ask

Why was your son's penis exposed to the cat in the first place where it could be bitten? Damn, I can only imagine how much all that would have hurt though.

[havocmom]

12. saying one thing doing another can also lead to STD

standing up for abstinance in public to get praise then being human in private is a common source evidently.

Sadly there is a big push to deny acurate and realistic sex education in many school districts. Parents in denial that kids are active and kids in denial about the nature of things, all clamoring to be vocal and visiable about chastity is a recipe for disaster.

[0007]

10. Underwear, Toilet Seats, Watching South Park, Listening to Howard Stern

are just a few. But toilet paper is probably the biggest piece of evident, 'don't 'ya think? :hurts:

[slinkerwink]

20. STDs can be transmitted via oral sex

and many teens think oral sex isn't real sex.

[BlueEyedSon]

21. I guess that's the legacy of the Clinton Presidency

:)

[kath]

25. That's right , one more thing that can be blamed on the evil

Clinton Penis of Doom !! (aka The Clenis) :-)

[BlueEyedSon]

9. "Sexually active virgins"

Noyce!

[truthbetold]

13. Oooh boy.

I better not let my mom read this, or she'll be dragging me off for a pap smear!

[donsu]

23. are you nuts? you should be dragging youself off for a pap smear

don't you care about your own health?

[BOHICA06]

14. Agenda Pushing

The study's authors say that from a statistical point of view, the rates are the same for both groups.

Do I see some bias here? - the difference for Whites/African Americans/Hispanics is 25% - 12.1% - 28.3% on the positive side, which is good. Such a difference in any medical study would make the significant threshold. Wonder about that 40% greater in the Asian population .... hmmmmm

SO the message of abstinence doesn't work well among all, it appears to work well for some.

[damnraddem]

16. Just like promises to obey the Golden Rule, ...

promises to 'abstain' have little effect on outcomes. So much for 'faith-based' interventions.

[Muddleoftheroad]

17. Lies and damn lies

So, in three out of four groups studied, the STD rate is substantially lower for pledgers, yet that is not what the "analysis" said.

1984 here we come.

For whites -- A 25% increase in STDs for those who did not pledge.
For Blacks -- A 12% increase in STDs for those who did not pledge.
For Hispanics -- A 28% increase in STDs for those who did not pledge.

Only Asians showed more STDs in the pledged population.

Now, something that is not apparent in the article. What are they counting as STDs? Oral Herpes is prevalent in the U.S. population. Do they count that?

Considering about one in five Blacks in the study have STDs -- a horrifying reality -- I support cutting down that number.

[kath]

27. Dpends on the statistical anaylsis -- depending on sample size, etc

what looks like a difference may not be a "statistically significant" difference

[Muddleoftheroad]

28. The base sample was 12,000

That's a lot more than you ever see for the surveys involved in the presidential race.

A more rational mind than those who did the survey might conclude that more study is warranted to analyze the gaps. But that wouldn't fit the agenda.

[Snow]

43. Still not enough, broken up into 4 groups...

To get statistical signifance, that is, to be able to rule out chance at the 5% level, here's what you need for each group:
Whites: 8,198
Blacks: 6,281
Asians: 430
Hispanics: 2,924

That's from this site:
http://calculators.stat.ucla.edu/powercalc/binomial/case-control/b-case-control-samp.php
And what you do is plug in first the % in the control group, then the ratio of the two groups, put in 5% for the sig level and .8 for the power - those are standard - be generous & call this a 1-sided test, and there you go!

Now if you want to say, for example, that black pledgers have a higher rate than white pledgers, you need only 54 subjects to rule out chance with the observed 6.5 times difference.

[Muddleoftheroad]

47. So what is the point of the survey?

After all, they surveyed 12,000 teens and found differences of up to 28%, but that's not statistically valid?

[Snow]

48. Very good question....

could be they were hoping for higher differences, then are trying to run with what they've got, which isn't much. Shoulda run a pilot, so they could have had some idea of what the differences would be in advance, so they could get a bigger sample. Looks like about double would do it. The other tactic would be not to split by ethnicity, but maybe there's no difference that way, either. This is realy hard to comment much on without the article, though - looks like it was a conference presentation rather than an article.

[Muddleoftheroad]

50. They found differences

They just discounted them, rather than analyzing them.

Where I come from, 28% is a pretty darn big difference.

And if 12,000 isn't a statistically valid sampling, then how are we to accept the political polling data that uses vastly smaller numbers?

Ultimately, it seems they allowed their own views to color the results.

[Snow]

55. I see your point,

and where I come from 28% ain't much, where I come from is epidemiology, where we're looking for association that are more often biological than behavioral - so I honestly don't know whether this difference should be considered small.

Regarding polls, those are usually set up so they have an error, a confidence limit, of +/- 3%. They can do it with much smaller numbers because there're several things that go into a sample size calculation - not only the size of the effect you're looking at, but how common the outcome is in your base population. Look at the difference is sample size requirements for whites as compared to blacks in this study, for example. The whites need the bigger sample because the outcome is less common.

It's hard to tell the degree to which they discounted the results rather than analyzing, especially from a news report. Scientists tend to be conservative, as you know, and if they can't discount chance, then they're going to be very conservative. I would guess that the kind of people who do this sort of study were hoping to find a positive result - that they didn't speaks well for their objectivity if that was the case. You seem to feel they were hoping for a result of no association, in which case I don't envy them because that's logically much more difficult to prove.

[dai]

58. "Where I come from,

28% is a pretty darn big difference."

This is a very foolish comment. A large difference in percentage may or may not be a large difference in actual numbers, especially dealing with rare events. Without more information one cannot conclude whether or not a large difference in percentage reflects a large actual difference.

Consider this hypothetical study:
You have a sample of, say, 100 "pledgers" and 100 "non-pledgers". 2 pledgers have an STD compared to 3 non-pledgers.

Oh my god, 50% difference!

This underscores the importance of statistical significance, which can better illustrate actual differences than reporting percentages alone...no matter where you come from.

[Muddleoftheroad]

59. The sample size was 12,000 people

That's a pretty large sample and several times over what they use for presidential surveys. That's 60 times your example.

So, there are two possibilities. Either the survey was poorly done and didn't have enough pledges to really count. In which case, the whole thing is bogus and we are all wasting our time. Or it did have enough pledges, in which case the difference is significant and bears further investigation.

[dai]

62. You cannot compare to presidential

surveys, because (1) these do not measure events such as disease and (2) these do not compare rates between two populations. It is a very different type of "study" and, as such, requires different sample sizes.

My point was that large percentages do not always mean large differences, please tell me you at least understand this because your original claim that 28% is a big difference is quite absurd.

What you are missing is the effect of similar rates, not sample size. Another example, with a sample size of 12,000:
Out of 12,000 people 2 "pledgers" and 3 "non-pledgers" have STDs...
again, a 50% difference but the sample size is 60 times larger.

But we can even use the actual rates...since we don't know the racial divisions I will use the rates for whites to describe everyone (these are probably the most accurate):

Assuming even divisions:
168 "pledgers" with STDs
210 "non-pledgers" with STDs

Out of 12,000 is this really a big difference? Not enough to discount chance.

You've missed the third and most likely possibility, which reflects a lack of statistical understanding on your part. The most likely conclusion is that there is no difference between the two groups and the difference you see is the result of chance, and has nothing to do with pledging at all.

[Muddleoftheroad]

63. I grasp all the stats bit here and none of us has the needed info

Of course, "large percentages do not always mean large differences." But based on the sample size, the survey was quite large.

Let me ask you a question, if 12,000 is statistically inaccurate, then why was this survey done in the first place?

And, as for your example, one of the subgroups was white teens. I think it is safe to say that is the largest group assuming they kept even vaguely close to American demographics.

Assuming that then, either there weren't enough pledges to make this valid, in which case the whole thing is useless. Or there were, in which case the variance is valid.

Actually, the most likely possibility in my opinion is that the survey was flawed from the beginning.

[dai]

64. response

Of course, "large percentages do not always mean large differences." But based on the sample size, the survey was quite large.

Once again, different types of studies require different sample sizes. I work with clinical trials, where our sample sizes are up to a thousand times smaller, but our subjects are carefully matched to remove the effect of chance and we can achieve statistical significance.

Let me ask you a question, if 12,000 is statistically inaccurate, then why was this survey done in the first place?

Not statistically inaccurate, statistically insignificant. Which means either (1) the difference is very small and they don't have enough people in their sample to show a difference, or (2) there is no difference in the actual population.

The study was done to determine if there was a difference between "pleding" and "non-pledging" populations. Concluding no difference is almost impossible, because one can always demand a larger sample size (see the two possibilities in the above paragraph). Concluding that there is a difference is much easier, but from this study - the authors cannot make this claim. Snow already stated this much better. (O/T question for Snow, did you take your handle from Dr. John Snow?)

Assuming that then, either there weren't enough pledges to make this valid, in which case the whole thing is useless. Or there were, in which case the variance is valid.

It is not useless, from the study we know that we cannot conclude a difference between "pledging" and "non-pledging" populations. There aren't enough subjects to show a difference, but there may be actually be no difference, so the study is not invalid simply for not achieving significance.

Actually, the most likely possibility in my opinion is that the survey was flawed from the beginning.

Well then, by all means design your own study and get a grant...best of luck to you.

The flaw I see is not one of sample size, but rather the effects of confounding. Then again, I don't deal with observational studies - this may be the best feasible study since trials are probably unreasonable.

[Muddleoftheroad]

65. Ongoing

I think this study actually needs to be vastly more complex than they tried to make it. I think there are many factors at work here that would have to be sampled for.

OK, again, if they "don't have enough people in their sample to show a difference," why would they do the survey?

[dai]

66. ongoing

I think this study actually needs to be vastly more complex than they tried to make it. I think there are many factors at work here that would have to be sampled for.

Generally, randomizing your sample groups corrects confounding, especially with large samples. However, this was an observational study and as such the investigators did not have much control over it. While you are correct that there are many factors at work, controlling for them may not be feasible. People are, after all, not lab rats.

OK, again, if they "don't have enough people in their sample to show a difference," why would they do the survey?

To determine if there was a difference in terms of rate of STD between this nation's population of "pledging" youths and "non-pledging" youths, and the degree of that difference. These are obiously unkown, so they sampled for it, as is standard practice.

If the difference in rates for the population was large, this sample size would have been more than adequate. However, the difference in rates is small enough that even with a large sample (although I suspect the minority samples are much smaller) we cannot tell with certainty which population has a greater rate. In other words, taking the samples over and over again, you would see a lesser rate in the "pledging" sample fewer than 95% of the time.

But, we have a fairly good idea of what the rates are for the population...this comes in an interval because every sample will be different. The conclusion is that either population could have a higher rate, because the rates are so close that the intervals overlap.

Colloquially, one could say that pledging makes no difference and that is how I interpret the study, but "no difference" is basically impossible to conclude statistically.

Perhaps this is not the result you want to hear, but this is the way research works.

[Snow]

68. There's also something to be said here about the difference

between a multiplicative risk and an additive risk, but it's probably too late at night for me to say it and make much sense. I'll just throw out an example, a favorite of epidemiologists: smoking (natch!). Smoking raises your risk of lung cancer about 20 times, but raises your risk of cardiovascular disease only about 4 or 5 times. So which is the smoking disease that has the biggest impact on the health of a community, and for that matter, which is the smoking disease you personally should be most concerned about if you're a smoker? It's cardioviscular disease, because it's so much more common that even that lower risk still adds an enormous disease burden. That's probably not the best explanation; I'll take another swipe at it tomorrow if folks are still interested.

[Snow]

67. O/T - yup, John Snow of Broad Street fame.

[cryofan]

37. could there be a biological reason for the high rate of STDs in blacks

WHy could there not be a bioogical reason for behavior? It seems reasonable to me....

[Snow]

Nope. This is well-known to STD control officers,

one of whom I helped with some mapping of rates a few years ago. STD rates come from lab reports. If your doc doesn't want your std reported, he/she treats you without sending in a lab sample. Your doc is also required by law to report, but in fact in most places the docs never report STD's; it's the labs. However, if you're poor, and we can agree that blacks are at higher risk of being poor than whites, then you get your medical care (during the Clinton years, anyway) from public clinics, where the docs don't know you and don't give a whoop about an std on your records - so they send the swab/culture to the labs, and it gets reported. End result, poor neighborhoods have much higher rates of reported STD than other neighborhoods. DO they have higher actual rates? Very difficult to find that out.

Re this study, it's hard to say whether reporting or diagnostic bias is what's going on. How were these kids diagnosed, was it all by the same clinic, are the kids of similar social class and access to care, and so on.

[Snow]

44. Dupe - deleted

[Voltaire99]

19. Doubts rising over faith-based virginity

It's so unfair! All those prayers, and you're still not a virgin!

[kath]

22. Note that this story comes from the BBC

Any bets as to whether PravdaUSA will cover it??

[Muddleoftheroad]

26. If they do

I hope they do a better job and actually point out the reality of the stats, not the spin.

[kath]

29. Are you saying that you support "Abstinence Only" education, which

withholds important information from kids??
Multiple studies have shown that it is not effective.

Yeah, it's good to tell kids to wait before they start having sex. But they also need real information on how to protect themselves from STDs and pregnancy.

[Muddleoftheroad]

30. No, I support both

The only reason for abstinence-only education is because parents don't trust teachers to teach both fairly.

My point about this survey is that it does indeed indicate that abstinence-only education works in three out of four populations.

Assuming the survey generally reflected the American demographic, then literally thousands of white students responded. And the number of those who had pledged getting STDs was a lot lower than those who did not.

[GinaMaria]

35. Where does it say that those surveyed

received abstinence only education? It said they took a pledge to abstain but didn't indicate what they had been taught before or after the pledge. The article starts a new topic when it discusses abstinence only vs. more comprehensive sex ed. which imo, seems to tie the two topics together. On closer inspection, the two subjects look like they should have been two different articles. Do you have another link to the research?

I ask because of your statement

My point about this survey is that it does indeed indicate that abstinence-only education works in three out of four populations.

[Muddleoftheroad]

36. You might be right

I inferred lacking the raw data. So you might be entirely correct.

Any way you slice it, this study calls out for the raw data and follow-up analysis. But clearly, pledgers have a lower chance of STD. So, then the question is whether that is cause or effect.

[GinaMaria]

38. If there's no other source then

I am right. No might about it. There is nothing in this article that says the pledgers received abstinence only ed.

Their rates are higher than most people would hope for from people who pledged abstinence. They are slightly lower than those that didn't pledge to abstain from sex.

Yes, the cause and effect piece is missing. Also missing are definitions of abstinence. What some people mean by the word may not be what other people mean by the word. Missing also are what constitutes a sexually transmitted disease? What are they counting/not counting?

There's not enough information to make assertions like yours.

[Beetwasher]

40. Excellent Points

I'm almost certain that I will be seeing the actual study this article was based on at some point as it pertains to my work. Should make for interesting reading...

[BOHICA06]

41. Don't use adverbs....

in describing the results. It induces bias.

The rates of STDs are higher for those that did not pledged abstinence in 3 of 4 populations. According to their own stats (cutting and pasting credit to Muddle):

For whites -- A 25% increase in STDs for those who did not pledge.
For Blacks -- A 12% increase in STDs for those who did not pledge.
For Hispanics -- A 28% increase in STDs for those who did not pledge.

Is 12% slight? 25%? 28%?

If in a clinical test, a treatment or protocol had this level success over another it would be deem significant.

[Snow]

45. See my little post on sample size up above

These in all likelihood close enough rates that you can't rule out chance as a factor.

When preparing a grant proposal for the feds (NIH) and hypothesizing that a risk factor is related to some outcome, for the feds to be interested you need to be able to demonstrate a doubling in effect. These differences are, as you say,

For whites -- A 1.25 increase in STDs for those who did not pledge.
For Blacks -- A 1.12 increase in STDs for those who did not pledge.
For Hispanics -- A 1.28 increase in STDs for those who did not pledge

In the field relating exposures to diseases, that's weak. For smoking and lung cancer, the association is around 20. For cervical cancer and wart virus, the association is around 105. So these are very weak associations, IMHO.

[daleo]

46. You are correct

Epidemiologists don't generally get too excited by relative risks on this order, particularly from observational data with lots of scope for confounding.

[GinaMaria]

49. Slowly calculatingly accurately

Before handing tips on preventing biased posts you might want to reconsider posting another's math work especially when those numbers have been disputed, no link was provided for the numbers and no method of calculation provided. Just a thought, since you are concerned with bias in posts. False numbers don't make your point.

So significant enough to change treatment? Only if I had a rare disease. As far as changing disease prevention/avoidance no thanks. Not significant enough. Especially not significant given the lack of information in this article.

Were those who pledged vs. not pledged chosen by the study group? ie: You will be in the pledge group, go take your pledge now. Were the pledgers internally motivated to be part of the pledge group? ie:something they chose. Along with a lot of other missing information from this study.

I will repeat, and I don't think I'm alone, how surprised I was that the pledge group had as high a rate of STDs as the non pledge group.

[Muddleoftheroad]

51. But it didn't have the rate

They just said "statistically the same" which is not the same.

[GinaMaria]

52. let me rephrase/I didn't type what I meant to say

I will repeat, and I don't think I'm alone, how surprised I was that the pledge group had as high a percentage of STDs as they did.

I expected their numbers to be a lot lower than they were, which is why I question how they determined the pledge group vs. non pledge group.

[Muddleoftheroad]

53. Fair enough

I get the feeling that if you had done this survey I might actually believe the results.

[GinaMaria]

54. I doubt that

but that's a discussion for another time and another thread.

[BOHICA06]

56. Now that the statisticians

have weighed in - I see that the difference is nothing to get excited about.

But in my defense - I did the math in post 14 - with the same results as Muddle. So the questions now are: Is this study's result repeatable? And, how big a sample will be necessary to show conclusively if there is a correlation between pledge, non-pledge & STDs

[Muddleoftheroad]

57. I would also like to see the study questions

Sexual history before pledging for instance.

Also, what do they count as an STD? Does oral herpes count?

[Snow]

61. Yeah, I would sure hope they asked those questions,

and the way that they ascertained STD is a major question. Did they examine the kids every few months, or what?

But to answer the sample size question, look at the numbers from my post up above. It appears they need to roughly double the sample size to be able to answer the question of whether chance plays a role (ie, achieve 'statistical significance' (I really don't like that term - sounds a lot more important than it really is)).

[dai]

60. Good questions:

These ones:
Were those who pledged vs. not pledged chosen by the study group? ie: You will be in the pledge group, go take your pledge now. Were the pledgers internally motivated to be part of the pledge group? ie:something they chose.

I seriously doubt this study could have been a randomized trial. Randomizing a diet or vitamin treatment is plausible, but a choice of this magnitude doesn't seem realistic for one to accept on chance. Far more likely, the investigators enrolled and observed participants who made the choice on their own (or their parents did).

So even if there were a correlation, I doubt you could conclude any causality as the "pledgers" would differ from the "non-pledgers" on so many other confounding characteristics.

Like you, I was suprised at how close the rates were, and expected the "pledge" group to be significantly lower. Even if this were the case, however, I would not attribute it to the pledge itself but rather other background qualities which would influence one to take such a pledge in the first place.

[Beetwasher]

69. It's Very Difficult From the Information Given in the Article

to really draw many conclusions about the work, unfortunately. I will wait to see the published study before I draw more conclusions. It will be interesting to see the instrument and also how they determined and defined a "pledger" and a "non-pledger".

[R Hickey]

31. Does lying about their sex lives make Asian teenagers horny?

The whole thing is probably just a sampling error, and the study should be re-done. But if this Asian thing is not a statistical annomaly, then we have to look for some other cause.

When Asian teens tell whoppers to inquisitive authority figures about their aspirations to longterm chaseness, does it make them more sexually active than the truthfull Asians?

Perhaps schools should be teaching abstance to certain minority groups, while handing out condoms and instructions to others.'

[MisterP]

34. the headline is rather insufficient, I'd say

[daleo]

42. I analyzed this with some artificial data,

and it does miss statistical significance with a simple chi-square test in SPSS. Since they didn't give actual numbers, I made a few simplifying assumptions:
- equal numbers of pledgers and non-pledgers in each group (6000 pledgers, 6000 non-pledgers)
- make up of the sample approximating U.S. population (70% White, 15% Black, 5% Asian, 10% Hispanic)
- apply pcts of std by group, as in the article.

The p-value for the entire population was .09 (Pearson chi-square).

So, the results don't appear spun to me.

Feel free to check this yourself. I may have made a mistake.