I planned this post yesterday, but the discussion it spurs might provide an interesting counterpoint to hilzoy’s most recent post.
There is an interesting paper by Jonathan Klick entitled "Mandatory Waiting Periods for Abortions and Female Mental Health".
It suggests a correlation between waiting periods for abortion and a decrease in women’s suicide. The correlation is robust when filtered for many possible variables, and is on the order of a 10% reduction. The paper controls for a possible general-trend reduction in suicides by comparing reductions in the female suicide rate to the male suicide rate in the same state. Since men and women have differences other than the fact that only women can have abortions, I’m not entirely sure if that is an accurate control. The correlation is across many states with many different situations, which should provide some additional verification possibilities.
There are however some possible criticisms:
It is unclear whether or not this correlation is linked to a reduction in abortions, a heightened level of preparedness which decreases regrets when the abortion is later performed, or some other factor. For purposes of the waiting period policy, this distinction is not necessary. For purposes of policy in general it would be interesting to know.
From a statistical point of view it would be better to track the suicide rate among women who actually attempted to have abortions and then encountered the waiting period rather than the whole child-bearing age female population of a state. That kind of study is unfortunately impossible with the privacy restrictions on abortion reporting.
Suicide rates are not a perfect proxy for mental health, though they are closely related to severe depression. The supposition that measuring suicides lends insight to depressions that do not lead to suicide seems defensible–especially considering the fact that a randomized depression study of women who attempted to have abortions would be very difficult to pull off in the current political atmosphere.
In all, the report seems very convincing. Waiting periods are often portrayed as a useless stumbling block. There is now evidence that they can have a significant mental health effect.
It’s a loose correlation, and, as all of us who read How To Lie With Statistics know, correlation does not equate to causation.
Jonathan Klick would appear (judging by a brief scan of his other papers in scholar.google) to be starting from the premise that terminating an unwanted pregnancy is intrinsically a bad thing: hence, no doubt, his picking up on this loose correlation.
Terminating a pregnancy in the first three months is, as is well known, considerably safer than giving birth.
The practical problem with an enforced waiting period is, simply, that it makes it more difficult and more expensive for a woman to get an abortion, when if a woman wants to terminate a pregnancy in the first three months, the faster it’s done the better.
The moral problem with an enforced waiting period is the same problem with all anti-choice measures; it assumes that someone, or some authority, knows better than the pregnant woman herself what her needs are.
(I have no problem with providing counselling services to women who want to discuss their options and find out more, of course: but once a woman’s made up her mind to have an abortion, it’s wrong on several levels for the state to think it deserves one last chance to try and make her change her mind, especially when that “one last chance” further delays the termination and makes getting one more expensive and more inconvenient.)
Pick your poison.
Note that the report finds a statistically significant increase in the rate of suicide when women are denied access to abortions (specifically in the study, denied Medicaid funding for abortions). The author speculates that it may be linked to despair about too many children to care for.
So it appears that laws restricting access to abortion have a negative impact on female mental health — if you find this study’s statistical analysis convincing.
I skimmed this paper — the conclusions are not well supported by the statistical analysis. The problem is eliminating from the data all other possible reasons for the variations in the suicide rate, since there is zero information as to the reason for the suicides. The models used do not do this on any convincing basis. It’s a real stretch to ascribe it to waiting period laws, or anything else.
The example they used in my statistics class:
Both, of course, had a relation to the amount of people and the houses they built.
Here is the link to a Dutch PDF about in metastudy done in 2003 to determine if there is a correlation between abortion and psychosocial complaints. Most of their researchmaterial was from the US and their conclusion was that the majority of women have no problems, a minority has some problems but they go away quickly and a very small part have problems after an abortion. Risc factors are youth, emotional problems in the past and previous abortions. More females experienced positive feelings than negative ones after the abortion.
We actually have a waiting period in the Netherlands (5 days) to make sure the decision is not made whilst put under pressure by the environment. I actually don’t think it is necesary. But abortions are (till 12 weeks) free and available though, so it adds only a bit of inconvenience and I can live with that if it prevents women from making decisions they do not wholeheartedly support.
If you want to prevent abortion problems, you have to aim at preventing people from getting pregnant unexpectedly. I read that in the US in 2000 49% of the pregnancies was unplanned and half of them resulted in abortion. Your abortion figure is 20 per 1000 people, ours is 8.4 per 1000 people (it used to be 6.4 and 0,3 of those under 15, but the rate has increased in the last few years).
There is lots of improvement in that area.
Lots of improvement *possible*
Dutchmarbel, the problem with waiting periods in the US is that in many areas it is not possible for a woman to get a pregnancy terminated near where she lives: she has to travel. In such a case, making her show up for a clinic appointment, then making her go away again for two to five days, is always going to be more inconvenient than useful.
Jesurgislac: from the factsheet about the US I linked too: “The number of abortion providers declined by 11% between 1996 and 2000 (from 2,042 to 1,819). 87% of all U.S. counties lacked an abortion provider in 2000. These counties were home to 34% of all 15-44-year-old women.”
I agree. I actually ment to say that the waiting period in all likelyhood will NOT have an effect on how women feel after they had an abortion, because so many other factors are much more important. We have a waiting periode, but I don’t think it really makes a difference – and we can have on BECAUSE it is only a slight inconvenience (you have to visit twice instead of once, but it is never really far and it does not cost anything).
Seb: I am (perhaps unwisely) responding to this after having read only the abstract — I don’t know how much time I’m going to have this morning. However:
You were right, when I posted about the housing discrimination survey, to point out that it didn’t control for credit ratings, and that this was a big problem, since in that variable, various possible explanations for the study’s result might lurk. For exactly the same reason, only more so, I am inclined to be very, very skeptical of attempts to infer any connection between a state’s waiting period for abortion and its female suicide rate, absent attempts to control for the various alternative possible explanations. And offhand, there are just too many of them for any study of this length to have controlled for.
— Since curiosity just got the better of me, I read secs. 1 and 2, in the latter of which he claims to motivate his methodology. That discussion concerned the claim that suicide among non-elderly people is a (conservative) marker for mental health, and not anything to do with the legitimacy of drawing inferences from a correlation of state suicide rates and state abortion policy to any sort of causal link.
Comparing female to male suicide rates does, I think, take care of one potential source of problems, namely questions about whether suicides might be reported more honestly in states w/o a waiting period. (At one point, early 80s, I checked out the claim that Sweden had a high suicide rate, and came to the conclusion that it might very well be that all that was going on was that Swedes report suicide more accurately than, say, Catholic countries, where suicides can’t be buried in consecrated ground.)
But it doesn’t really begin to address the innumerable possible confounding variables, let alone just plain extraneous factors, that might explain the sort of correlation he claims to find. And offhand, I would think it would be impossible to do so — for one thing, there just aren’t enough states for all the relevant differences to be adequately controlled for.
As I said, he may, for all I know, develop some nifty statistical method of dealing with all my problems later in the paper. But offhand, I think it’s unlikely that he does, or can, overcome the methodological issues.
I’m genuinely curious about the concept of an imposed waiting period.
It would seem that the decision to wait and think about the decision has probably already been made by most women (and their partners, if they are aware of the situation) given simple human qualities such as indecision and procrastination.
This is not to say that counseling should not be sought and provided. I favor counseling for women who seek and undergo abortions and for those who seek and decide not to have abortions. And their partners, if possible.
Further, I favor a waiting period before having sex, but only if society fills the waiting period with all manner of birth control methods, including the morning-after pill, for which there should be no waiting period.
Rapists and abusive partners require a waiting period of roughly forever.
Also, no waiting period for prenatal care, no waiting period for medical services for any post-born child, and no waiting period for equalized school funding.
I hate abortion and I am pro-choice.
I’m all for a waiting period in principle, given easy access to abortion. But as Jes points out, the waiting period in practice is a way to frustrate the ability to get an abortion.
Until the Right quits trying to physically stop women from getting abortions, and turns to addressing the causes that make women want to destroy their unborn babies in the first place, then I have to infer that the real game is controlling women, not saving babies. That’s why, like John Thullen, I’m pro-choice and anti-abortion.
I favor a waiting period before having sex
…How long since you were a young man? 🙂
Nine minutes.
Oh… I don’t mean waiting periods for me. I suffer from premature waiting periods.
8~))
Recent study (PDF) finding no link between depression and abortion. In fact, women who carried an unwanted pregnancy to term seem to be somewhat more likely to be depressed. (Possibly because those who do generally have lower income and education levels.)
I have a couple of comments and questions about this study:
1. Has it been published? The link supplied leads to what appears to be a manuscript page.
2. He’s using a statistical method I don’t entirely understand and am therefore not qualified to comment on in detail, but I do note that he claims signficance for some pretty marginal p-values.
3. Male suicide rate isn’t really the best control. Female suicide rates in states without waiting periods would be a better control. They would need to be normalized against the time before waiting periods were passed, though, to ensure that any effect seen wasn’t due to other factors.
“The supposition that measuring suicides lends insight to depressions that do not lead to suicide seems defensible–especially considering the fact that a randomized depression study of women who attempted to have abortions would be very difficult to pull off in the current political atmosphere.”
Just because good data would be hard (or even impossible) to get does not improve the quality of bad data. To find such a report “very convincing” is merely to reveal one’s own biases IMO.
Wu
Does anyone try to track women who want to have abortions, but cannot, because legally-required waiting periods make it too expensive? (I’m thinking of travel costs, child care, time away from work that may mean losing a job, time away from school that may mean flunking out.) How many of these women commit suicide? How many give birth and THEN become suicidally depressed?