Unwed young women who attend or have attended religious schools are more likely to have abortions than their public school peers, according to a new study.
The study also found “no significant link” between abortion and personal religiosity””defined by perception of religion’s importance, frequency of prayer and other religious activities.
Kendall has copied the headline straight from the RNS story, I just checked, but I think it must be an error. “No similar abortion rates”???
Doesn’t make sense.
They either appear to mean
“find similar abortion rates” (between religious and non-religious students)
OR
“find no significant difference”
Thanks, Elves. I was wondering about that.
In response to this news, I have a guess as to why teens in religious schools may have more abortions that those in public schools: They fear condemnation from the “good,” religious students and teachers when it’s discovered they are pregnant. Girls fear being shunned by their peers. Just a guess.
#2, if that’s what the study found. I thought the article lacked detail. Was the study conducted among secondary school students, students in “religious” secondary schools, students from “religious” secondary schools now in “religious” or otherwise college settings, students from a “non-religious” background who are not in “religious” colleges…? What is the definition of “religious schools?” I could not tell from which of many possible cohorts the author or the article’s writer drew conclusions.
[blockquote]While 25 percent of women in the sample reported having an abortion, that is probably an underestimate[…][/blockquote]
25% seems really high, doesn’t it?
Interesting that the sample includes teens who attend or have attended religious schools. My experience of 7 years as a student in a Lutheran school and 3 years teaching in a Lutheran school is that these schools often accept students who have worn out their welcome in the public schools. They attend for a while, perhaps a year or two; some are profoundly changed and some spend their energies trying to profoundly change the rest of the students. For many of them, the religion classes do not produce faith or moral change. In short, I think the study design is flawed.
Here’s a [url=http://www.asanet.org/galleries/default-file/Jun09JHSBFeature.pdf]PDF[/url] that, starting on page 5, contains the full journal article about the study. The study design doesn’t seem flawed, as much as its conclusions might make us wish otherwise.
Started reading the PDF, couldn’t get past the “diety associated with divine wrath and punishment”. No wonder conservative protestant women have fewer abortions: for one thing, they don’t worship a diety that is all about divine wrath and punishment!
Well I would start by assuming a young couple who claim to be affliated with a church which teaches traditional morals would not be having premarital sex in the first place. And if these religious couples are having premarital sex why should we be surprised if they also follow their own morality when it comes to abortion?
And I am not talking about a couple who may err than repent of their behavior and not have sex again until marriage. I mean unmarried couples that are engaging in sex as the norm for their relationship.
And I too could not get past the GOD OF WRATH, but can someone tell me if the survey asked how much influence their partner’s had on the womens’ decision whether to abort?
ember (#6),
Thanks for providing the link to the full study. I scanned it. It made much more sense than the very faulty summary version presented by RNS. Indeed, the first and last paragraphs of the RNS report are contradictory and very confusing.
David Handy+
With all due respect, Kendall may be better at theology than sociology. This article appears to be an example of the kind of “advocacy masquerading as research” that poisons so much of the social sciences and is, with little critical review, published in this field’s journals and duly reported by the uncritical press. (Witness last year’s widely reported “study” that said that young women who took chastity oaths were no more likely to refrain from extramarital sex than those who did not. The actual study, when read critically, proved just the opposite.)
To better explain why this present “result” actually proves the opposite of its headline…if it proves anything at all, consider the basic design. It is a study of a cohort of “unmarried and never-divorced women aged 14-26 who conceived their first pregnancies.” In other words, this is only that subset of women and girls who get pregnant out of wedlock. It is not a study of all women of that age group. It then tries to compare the reported abortion incidence of those who indicated various kinds of religious affiliation to those who did not so indicate and, at least at the headline level, claims that there are no differences. In fact, the study demonstrates highly significant protective effects for conservative Protestant women and even those who just attend schools with large proportions of conservative Protestants. These women and girls reported half the likelihood of undergoing abortions versus the unchurched. Even women and girls from us “other Protestants” showed only 55% the likelihood of abortion. By the way, carefully buried in a couple of the tables but uncommented upon is the finding that availability of abortion clinics in the county of residence increases the likelihood of abortion by as much as 75%. And, the absence of public funding for abortion clinics decreases abortion by 35%.
A more technical criticism of this study concerns the statistical test of significance (the p value) that it alleges shows real relationships. Several variables in the various models that are proposed are reported as significant at the p < .10 level. For the non-statisticians in the group, this significance value would not be recognized in most reputable scientific journals. A value of p < .05 is the usual criterion and is often adjusted downward to p < .01 in studies where this many variables are considered in a model. The p value is an estimate of the probability that a finding could be the result of random chance. I.e., a p value of .10 says that there is a 10% probability that the finding is the result of chance...or conversely, if you did ten such comparisons of completely random variables, at least one of them would appear to be significant just by chance alone. This study's author included as many as 29 variables in some of her analyses so using a p<.10 would give about 2 or 3 "false positives" on random chance alone. Bottom line for this study is that it does not prove that there is no relationship between "personal religiosity" and sexual behavior outside of Holy Matrimony. If anything, it may prove what a waste a PhD in sociology is...for the student and society.
Along I suspect with many I am deeply grateful to Jim K. He confirmed one of my suspicions: if the sample of 1500 included both public and church schools, given the relatively low ratio of church schools to public, how could the sample from church schools be large enough to be significant?
BTW will someone e-mail RNS to ask them (politely) to rescind the article?