Jay Teachman, an academic, investigated this matter further. The study, which is available online, makes for interesting reading. Teachman's genius was to look stratify the cohabitors risk of divorce by the by the number of sexual partners/cohabiting history.
The study was based on data from the National Survey of Family Growth 1995 cycle and involved over 6500 women.
It was controlled for a host of variables.
The study was in no way sponsored or funded by any conservative organisation.
Teachman's conclusion:
The results presented in this article replicate findings from previous research: Women who cohabit prior to marriage or who have premarital sex have an increased likelihood of marital disruption. Considering the joint effects of premarital cohabitation and premarital sex, as well as histories of premarital relationships, extends previous research. The most salient finding from this analysis is that women whose intimate premarital relationships are limited to their husbands—either premarital sex alone or premarital cohabitation—do not experience an increased risk of divorce. It is only women who have more than one intimate premarital relationship who have an elevated risk of marital disruption. This effect is strongest for women who have multiple premarital coresidental unions. These findings are consistent with the notion that premarital sex and cohabitation have become part of the normal courtship pattern in the United States. They do not indicate selectivity on characteristics linked to the risk of divorce and do not provide couples with experiences that lessen the stability of marriage.Executive summary: It's not the liberal values, it's the number of partners that matter.
And,
This limitation notwithstanding, the results presented here should shift attention away from research that focuses on the selection of individuals into cohabitation and premarital sex to a focus on the selection of individuals who do not marry the individuals with whom they first cohabit or initiate first sex. It may well be the case that, irrespective of the legal status of the relationship, the relevant distinction to make is between people who form multiple relationships and people who form a single, longer lasting relationship.The paper data and methods can be found here. Premarital Sex, Premarital Cohabitation, and the Risk of Subsequent. Marital Dissolution Among Women
(My highlighting)
Oh, in table 4 of the study, Teachman gives probabilities of divorce, which for fun, we will map onto the Heritage Foundation's study.
10 year divorce rate of the Teachman study group was 34%. (I didn't use the 5 year divorce risk in Teachman's paper since the Heritage study made its calculations on the people being married more than 5 years)Teachman didn't plot the risk by the number of sexual partners, merely that more than one and in different relationship contexts, so I have simply marked the range of his findings. Note, the really disturbing one still holds. A soon as a woman has had more than one partner her long term marital stability risk drops to near 50%.
Note: Statements reproduced from the Teachman paper are not believed to violate copyright under the fair use clause. Any violation is unintentional and offending material shall be removed immediately if it violates any repespective laws.
40 comments:
SocPath, you took the wind out of my sails! No joke. I was about to download the NSFG data myself and write the SAS code for a hazard model like Teachman's.
One issue I'd like to see further explored here though is the role of religion. It's been argued elsewhere that it's not lack of premarital sex per se that's important but religiosity. Once religiosity is controlled for, premarital sex might fall out (though I don't think it would). While Teachman includes the woman's religion as a covariate, I'd prefer to see the dummy baseline coding be "no religion" or similar (assuming that was even a response option). Or, even better, I'd like to see a variable included for how "important" religion is in one's life for the wife--just as Teachman included for the husband. (The questionnaire shows this question was indeed asked of the wife.) I see that after controlling for the husband's religiosity, pre-marital sex on the part of the wife is still a significant risk factor.
...And while you're at it, add the socioeconomic variables to the model.
Second, it would be interesting to see this same equation built for men. There was indeed an NSFG section for men, but I'm even less familiar with it. I wonder whether the effect of pre-marital sex on divorce differs for men vs. women.
One more comment on this type of survey: it can always be faulted that the NSFG study is retrospective. If one finds that, say, religiosity, socioeconomic status, or some other variable is a significant risk factor for divorce, in a retrospective study it might be that the risk factor developed AFTER the divorce. For example, a woman's socioeconomic status quite often falls following a divorce; or perhaps the experience of divorce can make one disillusioned with organized religion. ...Though a look at the exact question wording (and the application of common sense) might resolve some of this.
Hi nml,
I've got to rush out to work but briefly, The 2002 NSFG does have data on men, and yes, the more promiscuous the man, the higher the risk of divorce.
Though, it appears that the each partner a man has increases his risk of divorce to a lesser degree than that of a woman. The double standard appears to be real.
I'll reply some more much later in the day.
lzozlozlzolzozzlzozozozl
what if she was butthexed and deosuled in a fed assckong sessions it was secrtely taped by a no0eocn butthexula hero who was had his lies repeated in the weekly standardth about his highteh ? lzozlzlzlzl
http://greatbooksformen.wordpress.com/2010/07/28/lozlzozlzozlzozlzozlozlzlozozo-i-am-starring-in-a-movie-just-like-leonardo-decrapyyho/
What is the 52%--90% band?...is it a confidence interval? If so, why wouldn't there also be confidence intervals rather than just points for the "zero" category?
@David,
What Teachman basically did in this study was calculate the divorce risk for
1) Women who did not have premarital sex and who only had sex with one partner.
2) Women who had premarital sex with one partner only and cohabited with him.
3) Women who had premarital sex with more than one partner under a variety of circumstances. The data of relevance is in table 4 or his study.
The third group of women would have included 2,3,4,...... and so on women. Teachman did not work out the divorce risk each additional extramarital partner above 2, so what in fact his risk coefficients represent are the risk factor for women with 2+ or more partners under a variety of circumstance.
So for example, the category
Premarital sex with husband and other no cohabitation, group would include women with lifetime partner count of greater than or equal to two. There will be women in this category who will have 2, 7 , 12 etc number of partners, he has estimated the risk across those partner cohorts.
I don't know what the risk is for a woman say with 6 partners, who had premarital sex with her husband and other and who did not cohabit, I've simply drawn the line across the cohort.
The top line represents the smallest calculated risk across the cohort and the bottom line the greatest risk calculated. What I was attempting to show is that the Heritage Study's observations fall within Teachmans calculated risks and that they weren't so aberrant, as some people claimed.
GBFM.
What your seeing is the bernankification of women; after being butthexed all they can offer is fiat love.
Tucker max is bernanke
mnl.
Is SAS code difficult to write?
Because it would be really worthwhile to trawl through that data. Quite frankly, there is very little scientific study on the effect of promiscuity on bonding. My own impression upon looking at the data and other papers that I now have is that promiscuity seems to cause some effect with bonding and sexual exclusivity, especially if some of the promiscuity occurred in adolescence.
Practicing religion, as opposed to stated religion, is correlated with a decreased risk of divorce, but the interesting study would be looking at divorce rates amongst the religious by partner count: A sort of reverse Teachman. I imagine that even there we will observe an increase rate of divorce associated with increasing numbers. I suspect that partner count exerts an independent effect on marital stability and the capacity for sexual exclusivity.
Teachman basically controlled for both cohabitation and "liberal values", the fault of his study was that he didn't push it far enough to see what sort of correlation there was between of the number of sexual partners and divorce risk, I imagine that a lot of the variation in Teachman's situational models could be explained by partner count.
As for faults with the data, what data is ever perfect? A quick perusal through the CDC data convinced me that educational and economic factors were not real issues. I think people have to interpret the data with a degree of common sense. Expecting milimetric precision from centrimetric methods is just silly and really just sour grapes. Saying that results are invalid because you can't achieve milimetric precision in subjects that you don't need such precision says more about the objector than the study.
Would you be willing to trawl through the data? I'd do it myself but I'd have to get SAS training which I'm afraid I can't do now. You would actually be in the forefront of relationship science.
@SocPath...
Working in SAS, like almost anything else, isn't hard--once you pass the point of it being hard! In the case of the NSFG data, the hard part probably has little to do with SAS per se (though the data are published as SAS data files). One could convert the data to SPSS or some other "easier" package. The hardest part is probably coming up to speed on the data, ensuring the variable you've got is indeed what you *think* you've got and is the "right measure" for your purposes. Pulling the SAS files now, I see there are over 3,000 variables in the 2002 female data file alone.
Next, like Teachman found, there are likely some hiccups such as missing data (which he describes solving using some data imputation and PROC MISSING). I'm not sure why that should be the case; it makes me nervous.
Finally, there's choice of the right analysis approach. Teachman uses a proportional hazard model. But are you better off with some simpler log-linear models?
It all makes it hard to manage a day job!
I'd be happy to talk more outside the public comments. But I don't see a "contact me" link on your blog. ...Or is that by design?
So the green bars then come from the Heritage Foundation study? Is that available anywhere on line?
Now that's an interesting find! I plan to read this paper very carefully as soon as I find some time.
It's funny how the accompanying press release reads like a piece of unintentional humor. (It's always fun to see liberal academics faced with such a nice juicy reactionary piece that can't be brushed off easily.) First the opening paragraph gives the maximum liberal spin on the results imaginable. Then a professor of sociology informs us about how amazed and startled he is by Teachman's conclusions. Better not tell him that the sky is blue, or it might be too much startlement for the poor professor.
Doug1, the Heritage Study link is...
http://s3.amazonaws.com/thf_media/2003/pdf/Bookofcharts.pdf
Thanks, yeah I found it.
I'm gonna post here a comment I made at Roissy's (Citizen Renegade), who had a post on today's chart etc. here. That's because I'd like to get feedback from the folks here.
--
Actually, it does appear that the denominator of the green bar chart from the Heritage study is all sexually active women over 30. Hence this does overstate the divorce risk. Divorce risk isn’t what it’s actually purporting to stand for and it shouldn’t be read exactly that way. That is if the woman is 30 or more and has been married for only three years she isn’t raising her green bar higher. So what this is really telling us is that those with very few sexual partners before their husband are more likely to 1) still be married for five years at the time of the survey and 2) have gotten married and earlier.
However the first will have much more weight for these reasons. The sample used for the green bar chart goes up to age 45 so the average age of the women included is around 37.5 – by which time most women are married if they’re gonna, and most will have been married at least five years by then. Only 7.8% of American women 30 and over were never married. So this affect while inflating the negative of the green bars which has been thought of as divorce risk, is attenuated.
Cutting the other way is only requiring the marriage to have lasted five years. That hardly means there’s no divorce risk left. In considering all marriages there's a large number of divorces in the under two years period, typically of younger marriages. But in marriages with children they typically do last longer than 5 years before divorce.
A more ideal study for divorce risk would be to look only at women say aged 50 (assuming few divorces occur after this age) who have been in a marriage at some point in their past that produced children or lasted longer than two years, and then ask 1) if they’ve ever been divorced in such a longer term marriage and 2) how many sexual partners did they have before such a longer lasting marriage began. (The breakup within two years of a childless marriage isn’t usually anywhere near the kind of big deal that divorce often otherwise is, financially as well as emotionally and on kids.) It would also be good to 3) ask the educational level of each partner and maybe 4) household income.
This study may not really measure divorce risk numerically but it does a good job of showing the much higher risk of increasing female partner counts.
I've got to head off to work now and will reply later.
@ Doug1
What I think Heritage did was divide all sexually active women by sexual partner number cohorts and then determine the stable marriage rate. Any dilution will be within the cohort.
Also any woman married less than 5 years was censored by Heritage. This information is found at the back of the book of charts, therefore women married five years all less were not included in the study. There is no overstatement effect.
My ideal way of doing this study would be to run it four ways, looking at divorce risk:
1) Religion vs partner count.
2) Income vs partner count.
3) Income and eduction vs partner count.
4) Comparing the differences between 5 year marriage partner numbers, 10 year marriage partner numbers and 15 year marriage partner numbers.
@Vladimir.
Glad to see that you're still around. Do you have a blog or website?
Yep the press release was rather awkward for the liberal position, but then again, a lot of the liberal world view is only logical by a strong dose of reality denial.
mnl.
I usually try to keep contact details to a minimum. This sort of stuff offends people, especially the "tolerant" liberal types. I also work in a public professional capacity so a bit of anonymity is necessary.
I'm more interested in mining the data rather than forming statistical models which explain the effect. Stuff like what's the partner breakdown amongst women who are highly religious. What is the partner breakdown on the 5, 10, 15 year marriage cohorts and so on. I'm not interested in complex regression models because:
a) They are difficult.
b) We all have lives to live
c) This is not a scientific journal and so I don't think we need perfect data, but maybe some rough and raw findings may spur some academic to do a bit of research on the subject.
If you're interested in pursuing the matter further I've set up a temporary Gmail account where we can correspond. Personally, I'd like to give it a go.
rosglitazone@gmail.com.
I've got a couple of papers here that you might be interested.
@Thursday,
I wouldn't mind getting in contact with you privately.
@The Social Pathologist,
No, I don't have a blog. I considered starting one several times, but it always seems like it would develop into way too much of a time sink.
In the meantime, I read the paper carefully, and on the whole, it seems very solid. It should be noted that Teachman overturns the conclusion that was reached by a widely cited study published in 1991, titled "Premarital Sex and the Risk of Divorce", by J.R. Kahn and K.A. London.
Kahn and London's study tried to find the reason why women's virginity at marriage predicts a lower chance of divorce. They concluded that the reason is selection for liberal vs. traditionalist values and attitudes, rather than the causal effects of premarital sex. However, that paper was based on some fishy looking complicated statistics, and in a subsequent issue of the same journal (Vol. 55, p. 241), I found a letter to the editor arguing against its validity -- followed by a rather unconvincing response by Kahn and London, basically saying, yes, our methodology sucks, but that's the best we have.
On the whole, it's crazy how little research in this area actually exists. I have to conclude that this is yet another area of immense practical interest where little to no work is being done because of its ideological sensitivity and propensity for politically incorrect conclusions.
I see the the end notes for chart 15 reference the definition of marital stablility listed for chart 7. You're right that definition does appear to throw out of both the numerator and the denominator women married less than 5 years. But it doesn't throw out unmarried women. Some of these may marry soon. So the green bars which your taking as divorce risk (by subtracting the bar percentage from 100%) are inflated by women 30 to 45 never married at the time of the survey.
I agree with you though that considering 5 years enough to indicate a "stable marriage" deflates the divorce risk that the converse of the gen bar percentage suggests. Plenty of marriages break up after 5 years, to the enormous cost of men and children.
"They do not indicate selectivity on characteristics linked to the risk of divorce and do not provide couples with experiences that lessen the stability of marriage."
To make sure I understand this portion correctly, it is saying that it is the mere fact they slept around that is the problem not that the experiences in and of themselves were bad or that the woman learned to choose men with poor commitment characteristics? The last line of it just throws me a bit as to what point he is making.
A soon as a woman has had more than one partner her long term marital stability risk drops to near 50%.
Actually, it shows no such thing. The study distinguishes between women who were virgins before marriage, women who have only slept with their future husband, and various categories of women with more than one sexual partner.
Now, it could be that having only say 2 partners results in only a small increase in the likelihood of divorce, but the less the women with fewer partners contribute to the divorce rate, then the more the sluts must be contributing to the divorce rate.
On the other hand if there is an immediate increase in the divorce rate from having had sex with someone other than your husband while additional partners do little to raise your risk of divorce well then that implies that a man are taking a big risk any woman who has had premarital sex.
In either case, the data do not look good for self styled sex positive feminism.
On the whole, great work in digging this up though.
@Thursday
A quibble,
The study distinguishes between women who were virgins before marriage, women who have only slept with their future husband, and various categories of women with more than one sexual partner.
You're quite right, Teachman didn't explicitly say that a woman's marital stability risk dropped to 52% after one extra partner, he stated that risk dropped to anywhere between 52%-90% depending upon the relationship category.
Looking at all the data and from other sources, it does seem that each additional partner contributes to an additional risk, but partner no 2 contributes the most.
@Vladimir.
I've actually got the Kahn and London paper here and think that it is reasonably good, except where they introduce the statistical novelites.
It appears that they attempted a novel simultaneous regression equation analysis, which on my analysis appeared pretty dumb. They were trying to quantify "unknown unknowns". However the first two thirds of the paper are quite good and contain quite valuable information.
In the end, their conclusion that when controlled for "unknown unknowns" virgins did not differ from non virgins seems pretty obvious.
However in their defence, they did not claim to know what the qualities of these "unknown unknowns" were, they speculated that it was "traditionalist values" though they freely admitted it was a speculation. It's still a valuable paper.
Women everywhere are different, but what may be a difference is that (typically) with women here you may speak more openly with each other early on in the relationship and be less hesitant to hold back the 'freakier' things. Not that I think Anal and Toys are freaky, but there are plenty who do. One of the things I like about online meeting is the more candid conversations you tend to have early on that don't happen for a LONG time when all conversations are purely face to face.
live call back sex chat
Based on this study you can NOT conclude that "liberal values" are not the real issue.
It could simply be that a woman who marries the first man she has sex with/lives with got lucky finding her Prince Charming right off the bat (excluding the virgin bride).
If a woman who eventually married the first man she had sex with or lived with had not gotten lucky she may have ended up as a woman who eventually had 15 partners before settling down.
Liberal values are what makes the woman take the chance in the first place (along with other factors such as self-esteem, need for companionship or support, hormones, etc.).
Maybe in a society where cohabitation was discouraged and sexuality was saved for very special relationships or marriage only, this woman WOULD have married her first lover because she wouldn't have felt pressured into having sex with every man she dates. She might have been selective and given her relationships time to grow before deciding if the man was worthy of such an intimate part of herself.
The oversexualization of our society is what cheapens sex so much so that it is something many men and even some women expect right off the bat.
Oversexualization also leads people to believe that sexuality is the primary and most important aspect of a marriage when there are other valuable factors such as commitment, integrity, communication, conflict resolution skills, etc.
I could go on and on...
As a woman I'd like to add: The number of sex partners a man has is equally important, regardless of the individual impact on relationships. After all, WITH WHOM DO YOU THINK HETEROSEXUAL MEN ARE GOING TO HAVE SEX WITH IF NOT WITH WOMEN?
Men cannot have 20 partners each while women all remain virgins. Unless you are advocating abusing women via prostitution or prostitution-substitutes (the young women with low self-esteem, no family support and/or no healthy boundaries).
We either encourage both men and women to limit the number of sexual partners or we accept that for each man who "gets lucky" a woman who might one day marry is also having (and maybe even enjoying) sex with a non-husband. This is not a single gender issue.
LOL! 21+ pre-marital co-habitations... WOW.. I wonder how big her sample size for that was.
As anyone with some red-pill understanding realizes, N matters. However, what this study seems to be suggesting is that "Cohabitating-N" matters MORE than straight N. In other words, the number of men a woman has tried a long-term-living-with-relationship is an easy measure of how easily she'll be able to bond to a husband later. I'm certain there's a parallel for men for this, but it's not cohabitative sex.
This is a very interesting read because I wrote on my website "Why stay a Virgin". Some one commented on my article and linked me to this website now I have a better foundation to strengthen why young people or people in general should continue to stay a virgin until marriage. This was an excellent finding. Thank You for this article.
At this point in the West, the staying a virgin until marriage is a fringe cultural phenomenon of devout Christians and non-Western immigrants.
I don't think it is possible to control fully statistically for the cultural differences between those who have sex and cohabit before marriage, and those who do not, therefore.
I better have a great marriage, I'm a thirty year old virgin woman. This brings me hope thank you. I know I will still have to work on my marriage though.
Cassidy
The Heritage Foundation is not a legitimate research foundation or think tank, where studies are peer reviewed. It is a right-wing propaganda organization with a political agenda. How foolish people are to talk about a "study" which clearly has a bias from the outset, and the biased "researchers" twist whatever results to fuel that agenda, in this case some idiotic idea women--never men--need to save themselves for marriage or else they make poor wives. Shame on you for peddling this drivel.
Divorce is an issue between husband and wife who both cannot participate in their role of marriage. The responsibility as a partners sex maybe a minor reason but not total the reason about divorce. colorado divorce law
Modern feminism teaches women that the word slut is a double standard used against women for engaging in the sexual behavior that many men engage in without the same value judgement. It makes logical sense, but from a biological, evolutionary perspective it's not in line with the reality of how women instinctively feel about slutty women. The irony to me is that men who think women who sleep with them lose value as mates for other men after that sexual encounter yet they know they don't want to mate with said woman, still do it. Who is the more morally corrupt person, the women who has been taught the value of her vagina doesn't lose value for every man she has fucked, believes it and so acts accordingly, or the man who KNOWS this is a lie because he feels it about women, yet takes advantage of this cultural shift to devalue women who buy into that belief system? I think men have a responsibility in the spreading of the hook up culture too. If we want to be better than animals, men have to take some responsibility as well in shifting cultural values towards sexual monogamy. But I think we need more women to speak out for men, since men speaking out just get vilified by the feminist mafia. It seems that, biologically, men can't help but feel that way, so feminism vilifying men for that belief (that my vagina is devalued for every man I fuck) is wrong as much as it would be wrong for vilifying women for having PMS. Yeah, it's stupid and irrational, but no one ever said evolution is rational.
I personally bought into the double standard idea for awhile, though in reality, I just had no idea how to actually get a long-term partner as a nerdy chick who didn't develop my feminine side...I have a physics PhD so just by having a vagina I already felt overly feminine! In college, I settled for too long to get what I could get from a few friends with benefits situations. And what I learned is that sex is no good if you don't have a loving relationship. It's just not worth it. I can satisfy myself better than any hook up dude can, so why bother? I'd rather wait for a guy who loves me enough to care about my experience in bed and not just using my body as a way to get off, basically, masturbating on me instead of making love to me. Making love is awesome and I don't want any more sex that doesn't qualify. Note that doesn't mean I don't want rough sex with my long term partner, but if I desire it, he knows it and is doing it because he wants to see me enjoy myself (hopefully he enjoys it too) then it's making love in my book.
I meant to say how MEN instinctively feel about slutty women. Honestly, I've always felt that way about slutty women too, the ones who sleep with lots of lots of guys, but I tried to not judge them because it's not my life.
Hi, thanks for this article.
But please, what does mean exactly the percentage 80.47%, 75% and 69% on first column? Thanks
Slumlord: Here's a very recent report that largely confirms the data in this (now very old) article. I'm posting it here, because this is my GO-TO page for this question.
http://www.dailymail.co.uk/news/article-2003700/Women-lost-virginity-teenagers-likely-divorce-says-new-study.html?printingPage=true
(Hey may wanna nuke some of that spam above here...)
Post a Comment