Thursday, July 19, 2007
The New Math of Abortion
CONFUSING THE ISSUE. Sometimes you stumble on some blog entry by chance, read it, go on, and then belatedly it occurs to you that something you read back there doesn't sound quite right.
That happened to me yesterday at Classical Values, where Simon posted this:
Eric at Classical Values is discussing Clayton Cramers's piece on the prevalence of abortion before Roe. His conclusion about abortion is that it may actually be happening at a lower rate since Roe.
A few minutes later, the buzzer went off: "a lower rate since Roe"? I didn't think so. So I followed the link to Eric's post, where he began with a similar statement (all the boldface in the excerpts below was added by me):
Clayton Cramer notes that there were plenty of abortions before Roe v. Wade -- and that there may have been more than there are now.
....even before Roe v. Wade (1973), Oregon theoretically made abortion unlawful except to save the life or health of the mother--and yet still had 199 abortions per 1000 live births in 1970. Does anyone really believe that 1/6th of all pregnancies in Oregon required an abortion for the life or health of the mother? You can pass laws, but if a large fraction of the population strongly disagrees, that law will be disobeyed unless you have a very powerful police presence trying to enforce it. Think back to the national 55 mph speed limit, or most restrictive gun control laws.
I'm reminded of the situation in Pennsylvania, which had similar laws to Oregon's, as did most states. I don't know what the Pennsylvania statistics from the period are, but I do know that as a practical matter it was very easy to obtain an abortion -- provided the individual woman (or her family) had the money to pay for it.
Eric's post goes on to underscore a point that's also relevant to the question of what statistics people are using and how they're using them:
Cramer also links his earlier post on the same topic, in which he concluded that the pre-Roe statistics were "disturbing to the conventional wisdom":
If you believe that Roe v. Wade started a deluge of murdering babies, then why was the abortion rate so high in states that pro-lifers would consider civilized? Does anyone seriously believe that 16% of Oregon pregnancies required an abortion for the life or health of the mother? It should be obvious that a lot of those were elective abortions, disguised as being for "the life or health of the mother."
If you believe that before Roe, America was a barbarous place where women had to get backstreet abortions (except for the five "enlightened" states of Alaska, California, Hawaii, New York, and Washington), then you need to explain why the abortion rates in some of the states with severe abortion restrictions were higher than the abortion rates in the states that allowed abortion on demand.
I find this data fascinating, and disturbing to the conventional wisdom.
I don't think either side in the abortion debate especially wants any of this publicized, and I think it's worth examining why.
The pro-abortion people like to paint the pre-Roe world as a patriarchal hell in which oppressed women were either forced to bear unwanted children or else resort to "coat hanger abortions." The anti-abortion people paint the pre-Roe world as governed by the type of values often portrayed in 1930s movies (aka "traditional values" -- but never mind the less traditional 1920s), in which those few women who might get pregnant out of wedlock would never have had abortions because a Godly America would not allow it.
The truth is unpalatable to both of these "conventional wisdoms."
But what is the truth? Is it at all decipherable? I looked up the data Clayton Cramer was citing. It consists of a table on page 31 of this paper (available as a PDF file):
The states highlighted in yellow are the ones that had abortion on demand prior to Roe v. Wade. The status of D.C. is unclear because the table in the PDF file employed boldface to indicate states with abortion-on-demand, and the D.C. typography was indeterminate.
Now. I don't want to say that Clayton Cramer has cherry-picked his data, but he has cherry-picked his data. The 199 per 1000 figure he cites for Oregon in 1970 is hardly the most interesting or instructive statistic, let alone trend, on the table. It probably is the highest abortion rate in a state without a significantly urban population that year, and in that respect it is actually atypical. The rates for Georgia, New Mexico, North Carolina, South Carolina, and Virginia are strikingly low. The figures for New York, D.C., and Hawaii are strikingly high. And what's the meaning of the crack about "civilized"? The people who might automatically assume that New York and DC are more civilized than Georgia and the Carolinas are probably not pro-lifers. In fact, even this far back in time, the distribution of abortion rates among these states seems to track with the current red state-blue state breakdown, which makes it pretty easy to understand the high Oregon number.
Worse than all this is the failure of the table as a whole to provide any support for Cramer's main point -- the one uncritically repeated by Simon and Eric -- that the data somehow suggest abortion rates might be lower after Roe v. Wade than before. The year-to-year trends on the table are unmistakeable. For the time span shown, the abortion rate increases every year in every single state that has legalized abortion in some measure. Furthermore, we see dramatic increases in the states that passed an abortion-on-demand law after1969, with the exception of Hawaii. For example, Oregon's neighboring state of Washington leaped from a rate of 83 per thousand in 1970 to 265 per thousand the following year, when abortion-on-demand became legal.
Indeed, the most interesting aspect of the data is the huge differentials between the states, which directly undermines Cramer's argument that abortion is somehow analogous to the 55 mph speed limit as a law held in universal contempt and therefore continuously, ubiquitously broken. One doesn't have to be biased pro- or anti-abortion to interpret these differentials as arising from multiple factors that could vary widely from state to state or region to region.
There's a general cultural factor -- the disposition of the population to seek abortion as a personally acceptable alternative to an unwanted pregancy. There's a medical ethical factor -- the disposition of physicians to find abortion an acceptable/unacceptable service to provide their patients and therefore to interpret the existing law strictly or loosely. There's a governmental factor -- whether the state administration regards a restrictive abortion law as meaning what it says or as a wink-wink license for de facto abortion-on-demand. Finally, there's a timing factor -- it may take longer in some states than others for women to realize that a restrictive abortion law is (or isn't) code for abortion-on-demand.
Depending on your viewpoint, for example, New York is a perfect storm of all these factors that drives abortion to insane levels, or Georgia is a perfect storm of the same factors preventing real access to abortion despite its technical legality.
So. Was Clayton Cramer right? Almsot certainly not. Additional searches turned up state-by-state abortions-per-thousand-live-births data for the year 2003. On the upside, this is far enough into the post-Roe era that any early volatility should be smoothed out. On the downside, the data are obviously from a different source and may reflect a different collection and verification approach. Still, the figures are better than no figures.†
*Indicates a state that added abortion-on-demand after 1969.
Everyone is free to draw his own inferences. My own take is that a couple of patterns exist which tend to confirm my interpretation of the 1969-1972 differentials. The states that started out with high abortion rates in the immediate aftermath of limited legalization seem to have plateaued: Oregon, Colorado, Maryland, New Mexico. This suggests that in these states limited legality probably was code for abortion-on-demand to a significant degree. Yet the two locales that suffered the biggest explosion of abortions in the wake of legalization -- New York and D.C. -- show very significant declines from their 1972 highs. Does this mean, as Cramer suggests, that Roe has in fact reduced the incidence of abortion? You're free to think so, I suppose, but there's a more obvious explanation. Note that New York's 2003 rate of 509 is very close to the 534 of its pre-abortion-on-demand year of 1970. It's therefore reasonable to ask how much of the stratospheric 1183 per thousand recorded in 1972 represented an influx from neighboring states (uh, say, New Jersey) where abortion was illegal or severely restricted. D.C. has similarly dropped back from the outlandish 1801 abortions per thousand in 1972 (that's 64 percent of all pregnancies that didn't end in miscarriage or stillbirth!) to a level just under its 1971 rate.
But the real kicker in these statistics is what happened in the states that produced low or relatively low abortion rates in the 1969-1972 timeframe: Delaware, Georgia, North Carolina, South Carolina, and Virginia. They're all way up under Roe. Note, too, that our simplified 2003 chart omits states that added restrictive abortion laws after 1970. (Incidentally, these are not "most states" as Eric declares, but just over half.) If we compare those to the 2003 data, we have quite a few additional examples of† hugely significant increases under Roe:
Of course, none of these figures speak to the question of how many illegal abortions were occurring prior to Roe v. Wade, which is the last refuge for Cramer's argument. Anecdotally, numbers quoted for this category of abortion vary by orders of magnitude depending on whether you are a pro-abortion or pro-life advocate.
Nevertheless, there are are also some data here that one can consult, if not for a definitive answer at least for some strong indications. Consider, for example, this analysis by AbortionFacts.com. It says, in part:
The head of one of the major pro-abortion organizations in the U.S. said: "In 1972 there were 1,000,000 illegal abortions and 5,000 to 10,000 women died from them." True? Or False?...The article also cites a 1981 study "by Dr. T. Hilgers from Creighton University, who estimated the figure probably was at or somewhere near 100,000 abortions annually in the U.S. prior to legalization." Case closed? Not quite. There's a huge qualifying factor in here, which the authors cite:
The United States, since the 1940s, has reported such deaths separately, so we know the number of deaths from illegal abortions. Good! Now if we knew how many illegal abortions it took to cause one death, we could easily calculate the total number of illegal abortions...
The following chart was used on the floor of the US Senate during the tumultuous debate on abortion in 1981. It was compiled from official U.S. statistics and was not challenged by the pro-abortion forces.
The reasons [for the drop after 1960] were new and better antibiotics, better surgery and the establishment of intensive care units in hospitals. This was in the face of a rising population. Between 1967 and 1970 sixteen states legalized abortion. In most it was limited, only for rape, incest and severe fetal handicap (life of mother was legal in all states). There were two big exceptions ó California in 1967, and New York in 1970 allowed abortion on demand...
In these two large states, legalization should have substituted "safe" for unsafe abortions. It should have saved many womenís lives. Actually there was no sharp drop in the number of women dying. Letís look further. By the year before the U.S. Supreme Court decision which allowed legal abortion on demand in all fifty states, the death rate for illegal abortions had fallen to: 1972 = 39 (With 25 additional deaths that year due to legal abortions.) Now abortion was legal in 50 states. Now back alley abortions should have been eliminated with their alleged toll of maternal deaths.
In 1973 there should have been a really sharp drop in women dying. The chart, however, shows that there was no such drop. The line didnít even blip. The previous rate of decline actually slowed, to flatten out in the late 70s and 80s. According to the U.S. vital statistics, as anyone can see, legalization of abortion did not save almost any womenís lives...
letís recap: Pro abortionists claim that in 1972, the year before the Supreme Court legalized abortion, there were 1,000,000 illegal abortions and 5,000 to 10,000 women died.
Actually only 39 women died ó less than one per state per year.
But they canít have it both ways.
- Either there were not many illegal abortions
- Illegal abortions were all extremely safe.
Since we assume that all illegal abortions were not extremely safe, it seems obvious that THERE WERE NOT MANY ILLEGAL ABORTIONS.
Letís look at the late 1950s. Those were the supposed bad old days. All abortions were illegal, and illegal abortionists were alleged to be busy. In the July 1960 edition of The American Journal of Public Health, there was an article by Dr. Mary Calderone, founder of SIECUS and medical director of the Planned Parenthood Federation of America. She stated:
"90% of illegal abortions are being done by physicians. Call them what you will, abortionists or anything else, they are still physicians, trained as such; . . . They must do a pretty good job if the death rate is as low as it is . . . Abortion, whether therapeutic or illegal, is in the main no longer dangerous, because it is being done well by physicians."
This opens to door to the possibility that a lot of abortions were being performed illegally under faiirly safe conditions prior to any legalization. Might this rescue Clayton Cramer's claim? Probably not. Why?
Isn't it reasonable to assume that the first legal abortions were performed by the very doctors who were previously willing to offer that service under the table? Isn't it also reasonable to suppose that there were such doctors in every state, albeit in different numbers? And finally, isn't it reasonable to suppose that women seeking an abortion would always prefer to obtain it legally rather than illegally?
If the answers to these questions are yes, then the figures we have for 1970 and 1971 are almost certainly an approximate upper boundary of the most liberal estimate we might make of the number of illegal abortions performed in the U.S. prior to Roe. In all likelihood, the actual number of illegal abortions performed is lower than the '70-'71 figures, because illegality does deter some percentage of people -- patients and doctors -- from breaking the law. After all, there were solid citizens who did observe the 55 mph speed limit, as even Clayton Cramer might concede.
I am not saying that any of the analysis establishes the facts beyond dispute. Nor am I making any statement whatsoever about what the data suggest with regard to the disposition of the abortion issue itself. What I am saying is that Clayton Cramer's analysis is misleading and insupportable based on data that are available.
I'm politely suggesting that Eric and Simon of Classical Values be more careful in the future about who and what they quote unskeptically.
On all the big social-cultural controversies of our time there are large numbers of what we might call free-floating statistics; i.e., Kinsey's 10-percent gay guesstimate, NARAL's 10,000 women a year dead of botched abortions before Roe, a million missing American children, Michael Moore's 47 million Americans without health insurance, etc. It's hard enough to discuss such matters honestly and productively without the noise created by false but endlessly repeated statistics.
It behooves all of us who write about these topics to be as careful as we can be about not contributing to the miasma of falsehoods the most ardent on any side would have us believe.
Sorry to have been so long about it.
UPDATE. Clayton Cramer writes to say that he never claimed Roe v. Wade reduced the number of abortions. He's said as much in an update to his blog. A close reading of his entries suggests that he is correct, although one can easily see how Eric and Simon of Classical Values made the wrong inference they did. Cramer succeeding in leaving a wrong impression he now regrets. I apologize for transferring a Classical Values misreading to Mr. Cramer. Nevertheless, I stand by my most serious objections to Cramer's entry. He did cherrypick data to make a slight point that was distinctly at odds with the data he had available and did not reproduce for his readers. I will also point out that his update does not acknowledge his material misrepresentation in important ways of the the data he cited† That his selective citation was consistent with his libertarian views I note without malice. My principal point still stands. This is the process by which† fallacious statistics become embedded in the public consciousness. We are all responsible to keep this from happening. That's why I'm being forthright about my own error. I have also asked Clayton Cramer, via email, to be forthright about his. If he is, I'll trust him in future. If he isn't, I won't. Sometimes things can be that simple, and that's as it should be.
For the record, I'll say that I emailed all parties involved about this post. Clayton Cramer responded promptly. I have yet to hear from anyone at Classical Values.
UPDATE 2. Cramer's okay. He meets halfway at least. That's a lot in the egosphere of blogs.
UPDATE 3. Eric Scheie of Classical Values has just checked in. He also defends Clayton Cramer and graciously directs -- in a blog update -- readers here to correct any mistaken impressions. Meanwhile, Mr. Cramer is claiming to have started hacking the Internet in 1972, when I was just a glimmer in my father's eye. Uh, okay. Not true.† 1972, when I was still watching the radicals begin their conquest of American universities. Who, I wonder, earns true veteran status on this battlefield?