Young Women Against Abortion

I’ve been analysing some more data from the Whose Body Is It Anyway? interactive activity, and I’ve come across a result that is a little… perplexing.

So a quick recap for anybody lucky enough to have missed me going on about this before. Whose Body Is It Anyway? is about abortion; it’s been completed more than 16,000 times; and during it people are asked whether they think abortion is morally justified (and in what circumstances). I’m taking it that people who have stated it is never justified or only justified where a mother’s life is at risk are broadly against abortion (though, of course, it doesn’t necessarily follow they’ll be against abortion rights).

I’ve been looking at how people under-30 have responded, particularly those who are against abortion. This is what I’ve found that I think is interesting.

1. 29.5% (2032 out of 6892) of men under-30 are against abortion. This compares to 36.6% (1392 out of 3800) of women under-30, who are against abortion.

I’ve put together a Z-test calculator, which is here, and if you plug the figures in, you’ll see that this is easily a statistically significant difference. (The Z-Score is 7.582, which is significant at p <.01).

Okay, so I thought, well, the difference is probably because proportionately more of the under-30 women are religious compared to the under-30 men. And actually that is the case. But this is where things get a little odd.

If you look at religious people under-30, you find:

2. 55% (1269 out of 2299) of religious men under-30 are against abortion. This compares to 55% (950 out of 1719) of religious women under-30, who are against abortion.

This result is not statistically significant. Plug the results in yourself, and you’ll see. (The Z-Score is 0.042, which is not significant at p <.05).

Right, now it’s important to recognise that even though the proportion against abortion is the same for men and women in the religious under-30 category, it’s still possible that it is a preponderance of religious women that explains the overall difference between the responses of under-30 men and women. (If you can’t figure out why this is the case, have a go at this Simpson’s Paradox activity, and have a ponder!).

However, this possibility can’t explain the third, and most perplexing, result.

3. 16% (763 out of 4593) of men of no religion who are under-30 are against abortion. This compares to 21% (442 out of 2081) of women of no religion under-30, who are against abortion. This difference is easily statistically significant. (The Z-Score is 4.553, which is significant at p <.01).

So why are young women of no religion more likely to be against abortion than young men of no religion?

I find that counterintuitive, but I know that other people don’t (because they’ve told me!). It’s possible that I’m missing something here because of my UK background. (Abortion is pretty much a non-issue in the UK.)

Leave a comment ?


  1. Jeremy, I wonder… presumably you’ve lumped everyone who has chosen any “religion” together, I wonder what would happen if you were a little more circumspect and checked for those religions that are theistic as opposed to those that simply suggest a path. Another would be to compare the age of the religion – i.e. Judaism/Christianity/Islam/Sikism…

  2. I’m male, < 30, atheist, in favor of abortion.

    I have to say, these results are exactly what I’d expect. The fact is a lot of men like me are terrified by the prospect of having unwanted children! The classic stereotype on the American sitcom would suggest that “men are afraid of commitment”. Obviously there is a little more to it than that.

    I don’t have any data, but I would expect that women under 30 are more likely to be married or in a long term relationship than men under 30. Can you control for that?

  3. It may be that women under 30 and men under 30 who take your test do not form an especially representative cross-section of the population.

    It would be interesting to find out more information about who takes the test. Perhaps in the future, you should ask for more personal information from the test-takers.

  4. @Timothy – I do have data for different religions. But the trouble is the numbers of non-theistic religious people completing the survey are too small (currently) to be of interest.

    @Jai – Hmmm. Not sure about the married thing. For every married woman, there’s got to be a married man (obviously!). So I guess your hypothesis is that women tend to marry older guys. This is true, but whether they’re older enough to make a large difference to the proportions of married men and women under-30… well I’m not convinced that’s the case.

    @Amos – It’s definitely not a representative sample. But, even so, it’s still interesting that non-religious women are more likely to be opposed to abortion than non-religious men. I’m not sure unrepresentativeness explains that result (though I guess it could if, for example, the link had appeared on a secular anti-abortion site frequented largely by women, but that’s a stretch!).

  5. It could be the way that the questions are framed.

    Women are pro-choice, almost never pro-abortion. The groups which promote choice never confront women with the reality of abortion, of killing a living being or at least of not allowing a living being to grow, as (I recall) appears in your test.

    Males are probably more likely not only to be pro-choice, but also pro-abortion. As Jai points out above, for many young males, a baby means commitment or paying child-support, while for almost all women, a baby means a new life growing within them. That at least is my experience.

  6. How come I’m not surprised? Isn’t it generally true that women, religious or not, want babies more than men, especially irreligious ones?

  7. Also,in the paragraph started “Right now it’s important . . .,” I don’t understand the end of that sentence. I looked at the Simpson Paradox activity and though I didn’t find it a paradox at all, it didn’t help explain the above.

  8. @Ralph – Maybe I wasn’t clear about what I meant with the “Right now it’s important…) bit. I was getting at the following possibility:

    Suppose there are 2620 people under-30. 60% of men are against abortion (360 out of 600). 79.7% of women are against abortion (1610 out of 2020).

    This breaks down as follows:

    160 out of 200 religious men against abortion; 1600 out of 2000 religious women against abortion.

    200 out of 400 non-religious men against abortion; 10 out of 20 non-religious women against abortion.

    Okay, so many more women are against abortion overall than men, but if you break it down you find that in neither the religious nor non-religious category are a higher proportion of women than men against abortion.

    The reason this occurs is because there are many more women in the religious category than men, and both men and women in the religious category are more likely to be against abortion than either men or women in the non-religious category.

    That was my point. The fact that there was no difference in the *proportion* of men and women in the religious category opposed to abortion doesn’t rule out the possibility that it is the *preponderance* of women in that category that explains the overall result.*

    Of course, as it turned out, this doesn’t seem to have been the whole story in this particular case.

    *I could be missing something here, but I don’t think so! 🙂

  9. Jeremy,
    Thanks. I knew you were talking about weighted averages, so should have seen what you were saying in the original. Guess I just wanted to be thick headed.

  10. @JS: ‘3. 16% (763 out of 4593) of men of no religion who are under-30 are against abortion. This compares to 21% (442 out of 2081) of women of no religion under-30, who are against abortion. This difference is easily statistically significant. (The Z-Score is 4.553, which is significant at p <0.1).’

    Shouldn’t your significance statement be ‘significant at p<0.01’? Also, shouldn’t the second significance level in your ‘Z-Test Calculator’ be 0.05 and not 0.5?

  11. Keith – Oops! Yes, thanks! 🙂

  12. My once very elementary cook-book knowledge of statistics is now very rusty I even had to refresh my mind on Z-scores. However one thing occurs to me regarding the nature of the population, which is the fact, that they all have one circumstance in common. This is that they obviously are attracted to philosophical issues (this web site in particular maybe) and most likely, in varying degrees of competence, contribute towards them in some way be it by blogging or otherwise. Perhaps it would be of interest if the test could somehow also be be administered on other web sites dealing with for instance Art, Agriculture, Law, Business studies, Education, and so on.
    I have in mind here the fact that the soundness of behavioural research has in the past been heavily criticised due to its heavy reliance on college students as experimental subjects. Comparisons have been made where students and non students participated under identical conditions and the experimentation results differed in the two samples.

  13. @Don

    This is that they obviously are attracted to philosophical issues

    That actually isn’t true – at least nothing like to the extent that one might think.

    People come into these activities from all over the place. For example, last month about 6000 people arrived via StumbleUpon – and there’s no reason to think they will have any particular interest in philosophy.

    There is a slight distortion because of the fact I mention stuff on here, and because there’s “Philosophy” in the title of the site, but I don’t think it’s a huge issue.

    But, yes, there is a general point about the sorts of people who’ve completed these activities. Not least, their education levels are likely to be much higher than average; plus there are many more unbelievers (in God) than in the population as a whole.

  14. Jivin J’s Life Links 11-9-10 - Jill Stanek - pingback on November 9, 2010 at 2:01 pm
  15. You men are probably less pro-life because they have been raised in a feminized world where it’s important to be sensitive to women. They have been raised in a society that tells them men can’t tell women what to do.

    Secondly, a pregnant girlfriend can “ruin” a man’s life . . . child support for 18 years, etc. From a selfish standpoint abortion is an easy way to get out of any responsibility associated with “accidental” pregnancy.

    Not right, but maybe how they think.

  16. You’re missing that for the man, abortion is purely a hypothetical thing. HE will never be the one climbing on the abortion table and having his living child ripped out of his body in little pieces.

    Women get that in a way men can’t.

  17. Guys – Most people are missing the point here – though I think it’s my fault (I asked the question without filling in all the gaps).

    Why the difference between men and women in the non-religious category when you hold that up against the fact that there isn’t any difference in the religious category?

    Sure, you’re all coming up with reasons to explain why there might be difference in the attitudes that men and women have towards abortion (though, of course, it’s also possible to come up with reasons why one would expect the difference to go in the other direction – e.g., the fact that it’s almost always the woman who is left carrying the baby, so to speak), but why aren’t these operating in the case of religious men and women?

  18. Religious men and women are more likely to be guided by similar ethical principles regarding abortion. Their religion beliefs dictate whether or not they approve of abortion, and both genders have identical religious beliefs.

    Non-religious persons may let their ethical principles regarding abortion be guided by their gender identity or by their perceived gender interests. For example, as a non-religious male, I may be in favor of abortion because it allows my gender to avoid the trap of having to pay child support. I may rationalize that ethical “intuition” and not admit to myself that my intuitions come from my gender interests, but often they do, as other intuitions often come from my class interests.

  19. Their religion beliefs dictate whether or not they approve of abortion, and both genders have identical religious beliefs.

    Thing is, that’s just not plausible. Or rather, I’m sure that’s part of it, but surely the proposition that whatever non-religious factors lead non-religious people to oppose abortion melt away completely in the case of religious people is… well highly counterintuitive (well, I’d say it was). (Mainly because we just aren’t that principled about morality: at least, not normally).

  20. I was pro-choice in all circumstances, even up to 9 months until my 20’s and than only conceded the viability argument. Now with viability being lowered a few weeks a decade and a few days a year, it makes me uneasy about my position. I am beginning to wonder when there will be a procedure as invasive as abortion that might transfer the fetus to an artificial womb or something.

    IMHO, in the United States women have abortion at a much later date than many in the UK and having a child in some states becomes an economic calculus not an ethical one.

  21. “We just aren’t that principled about morality, at least not normally”.

    I agree with that, and it’s interesting that you say it, since that statement breaks a taboo of sorts.

    However, there is a big difference between what people say that they think about morality and what people actually think about morality. I would suspect that that distance is greater in religious people, especially when it comes to sexual ethics.
    What’s more, there is a big distance between what people actually think about morality and what they think that they think and that distance is greater in religious people than in non-religious people.

    The ability to see morality in non-religious terms is a first step towards realizing, as you say above, that we aren’t that principled about morality”.

    I’m not claiming that religious people are “better” or more principled people, but that they use a series of complex defense mechanisms so that the right hand does not see what the left hand does. Not being religious at least leaves you the option of being honest with yourself that you’re not all that principled about morality.

  22. @Amos – Okay, so if you’re right, then we’d expect the same differences to be manifest across other age groups, right?

    If we agree that’s what we’d expect, then we can set that up as our hypothesis, and I’ll mine some data to see whether it’s supported.

  23. My supposition: Younger women didn’t have to fight for feminist rights so they don’t have the traditional mindset that older females have. But many of them still get onboard regarding the importance of motherhood and as a caregiver of her children.

    Being the one that carries the offspring she is much more mindful of her role and the intuition -correctly- that prenatals are just as much her offspring as thise that are born. Not something any non-religious male has to consider.

    BTW you will notice that in the US more yound people overall are Pro-Life & I’d bet that the above and religion has much to do with those numbers.

  24. I’d expect religious people of all ages to have more rigid principles than non-religious people and for those principles to be very similar for both genders in religious people.

    When I speak of the principles that people have, I refer to stated principles or those principles that people think that they have, not to those principles that people act on when faced with real situations.

  25. If my hypothesis is correct and it probably isn’t since it’s mine, then the differences between non-religious men and women would be more marked when it comes to gender-related issues, for example, pornography, the legalization of prostitution,

  26. Testing blog: If my hypothesis is correct and it probably isn’t since it’s mine, then the differences between non-religious men and women would be more marked when it comes to gender-related issues, for example, pornography, the legalization of prostitution,

  27. “Why the difference between men and women in the non-religious category when you hold that up against the fact that there isn’t any difference in the religious category?”

    We are comparing the choice made by young women here. On the one hand young women who have a religion and young women who do not have a religion. So religion seems to be a causal agent. This being the case we are possibly comparing different psychological types of young woman, that is to say not comparing like with like.
    Religion has a profound effect on those embraced by it. They are more likely to be rule followers than relying on their own wisdom and experience to solve moral, and other problems. The atheist does not have the comfort of a holy book which if obeyed will ensure the best possible will happen. We most likely have all been exposed to the lure of religion at some time especially during childhood and many have succumbed, eventually rejecting it, and many hang on, Others like myself, have never embraced a religion. So the young ladies in this enquiry who are atheists and demonstrate a preponderance of agreement over the male atheists that abortion is acceptable. As stated above the atheist has to rely on his/her own wisdom judgement and experience. Free from the shackles of religion the young lady atheist knows exactly how much she has to loose in the event she becomes pregnant far more than the atheist man. She may be a sportswoman making efforts to improve, she may have a promising career ahead of her which pregnancy could destroy. She may be even enjoying life so much that a baby would wreck everything. It seems that the young woman atheist can see the future with the head rather than the heart in contrast to the young religious woman. She is in fact more forthright that the male atheist it seems, when it comes to abortion.
    It appears to me that the young religious woman and the young atheistic woman are in some way different psychological types and perhaps this accounts for the apparent inconsistency in the results of this test. Possibly the Atheist in this case is a bit better at thinking ‘outside the box’. I have absolutely no sources of information to quote in support of all this, and hope that I have in fact addressed exactly what is at issue here. I accordingly await for someone to shoot me down in flames.

  28. That religious people have been forced to interrogate their moral precepts while others are judging by their gut instinct?

    (I don’t actually believe that but – hey! – it’s a thought…)

  29. Since the survey respondents are self-selected rather than randomly selected, I wouldn’t pay any mind at all to this “survey.” See my post on this kind of thing:

  30. @Noelle – Yeah, that’s a nonsense. Sure, one shouldn’t generalize from a self-selecting sample (though it actually isn’t self-selecting in a straightforward way), but the idea that there is nothing interesting in non-representative data is just wrong.

  31. Yes, it’s certainly interesting to think about why people of a certain curious persuasion would be drawn to responding to this survey. (And why others would ignore it.) But I don’t think it says anything about what young women in general think.

  32. But nobody has claimed it tells us anything definitive about what young women in general think.

    However, it’s also the case that it might be revealing something about what young women in general think. This sort of “research” can be a jumping off point for more rigorous surveys, etc. It’s just bad logic to suppose that the only thing this sort of survey reveals is why some people might respond to it, and others, not.

    If you’re really claiming that there is no value in non-representative research, then not only are you doing away with the whole domain of qualitative research, but also with pretty much the whole discipline of social psychology, for example (and indeed, experimental philosophy), since it relies heavily on students for its experimental subjects.

    And finally, your remark about people being drawn to responding to this survey shows you haven’t bothered looking at it properly (I don’t blame you, but there it is).

    It’s not a survey. It’s an interactive activity that has – unbeknown to the people undertaking it – a “survey” aspect. Plus, it’s also the case that very large numbers of people have completed it without knowing in advance what it was they were about to complete (because they were sent to it via StumbleUpon – which functions exactly as it name suggests).

    Noelle, you should think a bit harder before pontificating on this stuff, because you’re doing yourself no favors here. I get that you want to make a point about unwarranted generalization, but… you know, in this instance, your point is very obvious (hence the 3rd and 4th comments in this thread).

  33. I do in fact worry about all the good points you bring up. I think qualitative research is vitally important. But my concern is that the language of statistics that I think you are invoking is not a good way to make sense of all this. (Correct me if I am wrong.) So, what is a good way to make sense of this research? I’m not sure, but I think this is a good question to mull over. And I do think it is important to take into account that in an online setting there are lots of possible reasons that some would participate in an activity and others might not — and that these reasons might have some bearing on the outcomes.

    Have you compared the results of this “activity” to other studies of opinion that was randomly generated? Have you seen any differences?

  34. The language of statistics wouldn’t be the way to make sense of this research if the intent were to claim the results were generalizable, but that isn’t the intent.

    But some sort of statistical significance test is necessary, otherwise what appear to be significant differences within this population, might be nothing of the sort.

    The online setting necessarily makes any sort of rigorous generalization very difficult (regardless of the representativeness of the sample). There are all sorts of variables that it is very hard to control. But, you know, I think about this stuff quite carefully. So, for example, these activities have timers built into them. If people very rapidly click through pages, then their results do not appear in the final dataset. (But then, of course, there might be systematic differences between those who click rapidly and those who don’t).

    But having said all that – if you check out my implementations of the Wason test:

    and Daniel Kahneman’s and Amos Tversky’s decision framing experiment:

    the results I’m getting precisely replicate the results of previous research.

    I checked for survey data on abortion – not very extensively – but it was hard to find anything that drilled down far enough to allow for real comparisons.

    And of course the issue of motivation for participation is important. But that’s important with any research. So, for example, if you check out the psephological research surrounding the 1987 and 1992 general elections in the UK, you’ll find that much of it is flawed because of a phenomenon that was subsequently labelled “shy Tories”. In other words, what polling organisations didn’t control for was the fact that there were systematic differences between those people willing to respond to questions about their voting behaviour and those that weren’t.

  35. Thank you, Jeremy. This is very helpful. I will follow your work with much interest.

  36. No problem. And I apologize for the “pontificating” remark. I was being a little defensive.

  37. Noelle – Here’s an example of a poll that is likely hopelessly flawed because of the self-selecting nature of the sample:

  38. Yes, I agree completely, not just with that mini-poll but with all the ones conducted there.

  39. Michael Lindgren

    I think it is a common mistake for abortion-choicers to interpret the pro-life position as religious in motivation. Oddly, no other fundamentally moral position is characterized as such. Murder, theft, rape, robbery; these are not interpreted as religious in motivation, they are simply crimes that no moral person accepts as moral. I would maintain the acceptance of abortion as moral is the fatal flaw in all moral and/or ethical thinking; the one thing that makes all abortion-choicer thinking suspect. Once the human being is devalued all human beings face the same danger.

  40. I think it is a common mistake for abortion-choicers to interpret the pro-life position as religious in motivation.

    Well, it might be a mistake, but it’s nevertheless the case that the data here suggests that religion is by far the most important factor in determining whether somebody is going to be against abortion or not.

    So, for example:

    37% of the 3983 Christians who have completed the ‘Whose Body Is It Anyway?’ activity support the right of a woman to have an abortion. This is compared to 83% of the 10337 people of no religion who have completed the activity.

  41. Noelle – Not sure if you’re still monitoring this, but I thought it might amuse you to know that Brian Leiter has just called me “an insufferable, sanctimonious arse” after I suggested (on Twitter) that his latest exercize in self-selecting polls was… well just that. 🙂

    He’s probably right about the insufferable, sanctimonious thing, but his poll is still absurd. (It’s absurd even if it turns out that it’s an accurate representation of professional philosophical opinion).

  42. I’m looking forward to reading Leiter’s Autobiography, Ecce Leiter, with exactly the same chapter titles as Nietzsche’s
    Ecce Homo.

  43. My shocking answer… because they lied about their religion.

    Now, I know I sound like a total […] to say it but […] are willing to do a lot of things to achieve their desired result if even on the internet.

    Redo the activity collecting ip address and I feel pretty certain that the results will change significantly.

    Another category that might also change the results would be to add “have a belief in god but not religious” to the list of religions.

Leave a Comment

NOTE - You can use these HTML tags and attributes:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>