Is Roe v. Wade the Main Cause of the 1990's Decrease in Violent Crime?

[Time for change]

It is according to Steven D. Levitt, award winning economist at the University of Chicago, as explained in his recent book, “Freakonomics – A Rogue Economist Explores the Hidden Side of Everything”. Here is a summary of the part of that book that answers the question posed in the title of this thread.


Background

Violent crime trends in the United States
The incidence of violent crime in the United States was fairly steady during the first half of the 20th century. It then began to increase in the 1960s, and continued to rise relentlessly for the next few decades, reaching a peak by about the start of the 1990s. By that time it had risen 80% in the past 15 years. “Experts” were predicting that the situation would get much worse, and saw no end in sight. Then, in the early 1990s crime suddenly and quickly began to fall. The fall would not stop until it had descended to its 1950s level.

Abortion trends in the United States
By the beginning of the 20th century, abortion had become illegal throughout the United States. By the 1960s, abortion laws in some states began to relax, and by 1970 abortion was entirely legal in 5 states (NY, CA, HI, AK, WA). On January 22, 1973, the U.S. Supreme Court’s decision in Roe v. Wade suddenly made abortion legal throughout the country, for the first two trimesters of pregnancy.

In the first year after this landmark decision 750,000 U.S. women had abortions, representing one abortion for every four live births. By 1980 the number of annual abortions reached 1.6 million, where it then leveled off.


Explanation for the role of Roe v. Wade in decreasing violent crime

Some of the strongest predictors for a child to grow into a violent criminal are growing up poor, unwanted, and in a one parent family, and having a mother who is a teenager and has little or no education. These are exactly the kind of circumstances that are likely to result when women are not able to obtain legal abortions. Therefore, Roe v. Wade resulted in a dramatic decrease in children born into such circumstances. And consequently, when the 1990s arrived, millions of children at very high risk of becoming violent criminals, who would have been born after 1973 and would then be entering their late teenage years, had not been born.


Evidence

The book does not contain enough information for one to do a full and independent evaluation of the author’s theory. However, I did find the evidence quite persuasive, including the following:

1. Time sequence
First and foremost is the time sequence of Roe v. Wade in 1973, followed by a precipitous rise in abortions, and then the drop in violent crime in the early 1990s, when the cohort of children who would have otherwise been born in 1973 and soon afterwards would have been entering their late teenage years, as explained above.

2. States where abortion had been legalized prior to 1973
Five states, New York, California, Hawaii, Alaska, and Washington, had legalized abortion prior to Roe v. Wade. In those states, violent crime began to fall earlier than in the other states, and by the early 1990s the decline had been substantially greater than in the other 45 states.

3. Association with abortion rates
States with the highest abortion rates in the 1970s experienced the greatest declines in violent crime by the 1990s. This correlation persists after controlling for several other pertinent variables.

4. Association with age
In states with high abortion rates, the entire drop in crime rate occurred in the age group born after 1973, i.e. the age group that would have included many more high risk teenagers if not for the legalization of abortion.

5. Other countries
Studies in Australia and Canada have established similar links between legalized abortion and crime.


Political implications

With the pending likely appointment of John Roberts to the U.S. Supreme Court, there would be one more certain vote on the Court for overturning Roe v. Wade, which would put the right of women to have a legal abortion in the United States in extreme jeopardy. With numerous Democrats vehemently opposed to the appointment of Roberts, his confirmation probably could not withstand the opposition of Republicans as well. Many Republicans are considerably more concerned about violent crime than they are about allowing women to have a legal abortion. Yet I doubt that any considerable number of them are aware of the link between making abortion illegal and violent crime. Perhaps if this information were widely known the likelihood of John Roberts being appointed to our Supreme Court would suffer a serious set back.


Note about the author

Steven D. Levitt teaches economics at the University of Chicago. He recently received the John Bates Clark Medal, awarded to the best American economist under forty years of age. The book is entirely non-ideological, presenting facts and interpretations that are certain to be highly criticized by partisans on all sides of the political spectrum. He seemed to me to be quite uncomfortable about presenting the information discussed in this thread.

[Zenlitened]

1. Ugh. I've heard this before, and it just leaves me cold.

Eeven if a link could be demonstrated conclusively, "crime control" is hardly a reason to support women's reproductive rights. :(

[Hamlette]

2. no, it is not a reason to support choice, it is simply a fact. Leavitt is

apolitical.

[neebob]

8. The fact is that violent crime went down,

not that it went down because of Roe. I read Freakonomics and, like Zenlightened, was left cold by the assertion that legalized abortion caused a decrease in crime. I have a huge problem with the assumption it's based on - i.e., that all these mothers of potential violent criminals ran right out and got abortions.

I say violent crime has more to do with the number of sociopaths than who their parents are, and I think that's a combination of genetics and environment. Certainly that includes poverty, but then you have all the rich criminals who don't have to resort to violence - George W. Bush, for example.

Anyway, I had an abortion in the early eighties, at a women's clinic, and it cost something like $400 - nearly half a paycheck for me at the time, working as a secretary for $1300 a month. I'm not sure that would qualify me as a poor single mother now; it certainly didn't then. I seriously doubt many poor single mothers were able to scrounge up that kind of money. They're not doing it now; in fact, the younger ones are trying to get pregnant and have babies.

What's the average cost of an abortion now, anyway? I have no idea. The price I paid might have been based on my income. I don't remember.

If it were possible to know who's had the most abortions since 1973, I'd bet on working single women aged 18-34 or teenage girls living with middle- and upper-income parents.

Levitt doesn't even say outright that legalized abortion caused a decrease in violent crime. He just spends a lot of time talking about it. It's the first of several sets of data that he goes on too long about - in a very self-aggrandizing way, I might add. The big message I got was, "I'm a genius!" His point, though, is about looking for patterns and coming up with conclusions that differ from the conventional wisdom.

[Time for change]

9. I don't understand why you don't think that giving women the choice of

whether or not to have an abortion couldn't lead to a decrease in violent crime. You say that you have a huge problem with the assumption it's based on -- i.e., that all these mothers ran right out and got abortions.

That isn't the assumption at all. The assumption is that the mothers who got abortions -- after Roe v. Wade allowed them to do so -- did so because they didn't want to have children. You agree that violent crime is related to the number of sociopaths, and you agree that that in turn is related to environment. Don't you think that being an unwanted child might be one of the primary causes of children becoming sociopaths?

[neebob]

16. I didn't say I don't think giving women the choice

could lead to a decrease in violent crime. I only questioned how much it has to do with it. And it is too the basic assumption, however you choose to say it, that more mothers of potential violent criminals received abortions. Of course they did it because they didn't want to have children.

As for unwantedness, I don't know if that in itself is a cause of sociopathy. It probably doesn't help.

[HereSince1628]

11. Since you've read it...Did the author talk about how he controlled for

the crime rates relationship to other confounding correlations, and how he distinguished the meaningful correlation from among these? I am thinking for example, that a good economy with low unemployment might help reduce crime rates.

[Demoiselle]

12. Yes. (I've read the book.)

[K-W]

13. ignore mispost

[Time for change]

14. Yes, he talks about several other potential causes in some detail

With regard to the economy, he believes that this was not a factor, and he supports this with these two arguments:

First, he says that several studies have shown that the economy has very little effect on violent crime. He says that a poor economy is related to economic related crimes, but not violent crime.

He notes that in the 1960s we had a great upsurge in the economy and in violent crime concurrently.

[HereSince1628]

15. Thanks.

[blondeatlast]

27. He accounted for all of the above. It's really an interesting book, sort

of an "Econpmics for Dummies" work.

Fast and fascinating read; check it out from the library, it's a single session read.

[Hamlette]

22. I didn't get the self-aggrandizing and I thought the arguments

more compelling than you did. Scientific proof is not like mathematical proof. His "proof" is in the scientific mode.

I don't remember poverty being as big a part of the argument as "wanted". It also dovetails into his arguments on what makes "good parents". Lots of us had abortions when we were too young to be rearing children. I wonder who I'd be had I been a mom that young...and who my child would be.

[MrBenchley]

23. It's a ridiculous oversimplification....

Crime decreased throughout the 1990s for a number of reasons....including increased gun control, Clinton's COPS program, an improved economy and the aging of the baby boomers.

Did having fewer unwanted and unloved children contribute? Possibly. But it's preposterous to pretend that that can be meaningfully quantified.

[Time for change]

25. Actually, all those issues you mentioned were discussed in the book

What kind of familiarity do you have with the evidence that leads you to say that it's preposterous to pretend that it can be meaningfully quantitated?

[MrBenchley]

32. Familiarity with reality....

Tell us please, how someone can possibly quantify the societal contribution of someone woh does NOT exist.

[Time for change]

33. It's done all the time in every scientific discipline

It's referred to as using a control group.

For example, it was established several years ago that that aspirin is a cause of Reyes syndrome (a fatal liver disease) in children who take aspirin when they have influenza. The only way to assess this was to compare the incidence of Reyes syndrome in children with influenza who took aspirin during their illness with those who didn't.

Or, sometimes it works the other way around, and it's found that people who lack certain exposures are at higher risk for certain diseases. For example, children who are NOT exposed to fluoride in their water supply are at much higher risk than children who drink fluoridated water to develop dental carries.

It's more complicated than that, because there are several criteria for assessing a causal realtionship between two variables once a correlation has been established. But the point I'm trying to make here is that virtually all scientific studies make use of a control group (i.e., unexposed) to compare with an exposed group, in order to ascertain the difference between them.

Does that answer your question, or would you like me to describe the criteria that are used to establish a causal relationship once a correlation has been established?

[MrBenchley]

38. Who are you trying to kid?

"it was established several years ago that that aspirin is a cause of Reyes syndrome (a fatal liver disease) in children who take aspirin when they have influenza. The only way to assess this was to compare the incidence of Reyes syndrome in children with influenza who took aspirin during their illness with those who didn't. "
Yeah. That's science.

This is the equivalent of comparing children with influenza who took aspirin during their illness with characters in kids' books with influenza who took aspirin during their illness...

The test group in this idiotic exercise in sophistry was never born. It's preposterous and offensive to speculate what they might or might not have done if they had.

You might as well speculate that we don't dominate the Olympics because "athletes" have been aborted. It's both offensive AND idiotic...

[Time for change]

39. Actually, the control group in this analysis WAS born

The control group was children born to mothers who had the opportunity to have legal abortions. This would include the same kinds of children who were born to mothers prior to Roe v. Wade, who did not have the opportunity for legal abortion, minus those who would have been born simply because the pregnant mother could not obtain a legal and safe abortion. This latter group was characterized by a substantially lower rate of violent crime.

I don't speculate that 'we don't dominate the Olympics because athletes have been aborted since Roe v. Wade', as you say, because there is no reason to believe that the increase in abortions occurring since Roe v. Wade would include a disproportionate number of athletes. But there is one category of potential children whom we know WERE aborted in disproportionate numbers since Roe v. Wade: Unwanted children.

I don't understand why you consider this offensive. Is it because you don't believe that a woman should have the right to abort an unwanted fetus, or is it because we are making the assumption that those aborted fetuses would have otherwise been unwanted children, or is it the notion that an unwanted child is likely to be more prone to develop violent tendencies?

[MrBenchley]

40. It's still mindless rubbish...

"The control group was children born to mothers who had the opportunity to have legal abortions."
And the TEST GROUP DOESN'T EXIST. There's no "there" there.

"I don't understand why you consider this offensive."
Guess.

[Time for change]

41. Both groups exist

The main difference between the two groups is that the group born before 1973 is likely to have a much greater proportion of unwanted children in it. How can you say that either group doesn't exist? They're just different, and that's what makes comparison of them useful.

And I don't want to guess why you find this offensive. I don't see why you can't just tell me. I suggested three possibilities, and you didn't respond to any of them, so if you want us to know why you find this offensive you'll have to tell us.

[MrBenchley]

42. Again, not even close to true...

"How can you say that either group doesn't exist?"
Because one group DOESN'T EXIST. They were never born.

"I don't want to guess why you find this offensive. I don't see why you can't just tell me."
I'd love to have a dime for everything you don't see.

"I suggested three possibilities"
Yeah, and Chimpy suggests people opposed to his incompetence hate America.

[izzybeans]

3. It also lends a scary legitimacy to the same Eugenicists who would

have us forget Nuremburg. Someone should tell the good economist that correlation does not equal causation, and innuendo based on tenous linkages are dangerous. Especially when published in a book dumbed down for popular consumption.

[Time for change]

4. He doesn't equate correlation with causation

He discusses in depth the difference between the two.

It doesn't recommend eugenics either. In fact, if anything he recommends the opposite.

The main point is that unwanted children are prone to have very unhappy lives, and that this can have deleterious effects on society as well.

This is one additional reason to for allowing women the legal right to have an abortion if they choose to do so. It poses absolutely no arguement for forcing them to do so. It is also a reason that I think is likely to appeal to Republicans more than some of the other reasons.

[izzybeans]

5. I didn't say he would recommend it. I was referring to the fact

that his argument is the same as their's was at the height of the movement. I read the book. Some of it was good, some of it was crap.

He spoke only briefly about unwanted children, but made the more eugenics minded argument by equating them with the "poor". I'll have to get my hands on the book again to be sure, but that's what I remember. ON edit: the argument that poor working moms should be either sterilized or made to have abortions so that they didn't have families that would contribute to "taxing the system" either through social almsgiving/welfare or the "inherent criminal nature of the poor" as the geneticists believed.

this is precisely the argument undermining the pro-choice movement.

I just did a search on Eugenics and abortion to demonstrate: you will find it's a major focus of anti-choicers justifications for their beliefs. It gives them "moral" capital in the battle of ideas.

http://www.google.com/search?sourceid=navclient&ie=UTF-8&rls=RNWE,RNWE:2004-19,RNWE:en&q=Eugenics+and+abortion

I don't support what's contained in many of the websites, I just thought I'd make mention of it sense it comes up alot during the course of doing research on the subject. Nor do I have an answer other than to say I find eugenics appalling and based on crass reductionist arguments. I just think that we should distance ourselves from such arguments, especially those made on huge logical leaps. It just provides fuel to the lie some of these people are telling in the google links.

[Time for change]

7. Eugenics

I understand your point, and I think I share the same concerns about it that you to. I admit I hadn't even thought of the idea that this issue could be used by eugenicists, before I posted this, and perhaps I should have.

So, it appears that our position occupies a middle ground between two extremist right wing positions, one that would limit a woman's choice to have an abortion, and the other that would limit her right not to have an abortion. They are both right wing positions in the sense that they limit freedom of choice. I would think that right now the former is a much more serious threat, simply because it is so much more prevalent in our country, and because it is probably a major reason why we are stuck with our current "president".

But although I share your concerns, I also think that there are some things that we should consider about this. First is that I have serious qualms about limiting the dissemination of information because we are concerned about the possibility that that information could be misused, especially in this case, where that possibility seems rather remote IMO.

Related to this is the question of the extent to which eugenicists could use this argument. Leavitt spoke in his article briefly of several risk factors that would make unwanted children be at high risk of growing up to be criminals. He mentioned poverty, having a teenage mother, lack of maternal education, a one parent family, and being unwanted. What he didn't go into, and what I probably should have stressed in this thread, is the fact that probably being unwanted is by far the most important, and perhaps the only important one of these risk factors. There are many reasons to believe that, which I won't go into now. But given that, it seems to me that when that point is made, the eugenisists have no argument whatsoever.

And lastly, you speak as if you don't believe the causative link between unwanted pregnancy and violent crime -- but you don't say why you don't believe that, except to say that there is a difference between causation and association, and you mention "reductionist" arguments, which I don't know what you mean by that specifically. I have worked as an epidemiologist for 20 + years, and one of the main requirements of my work is to differentiate between causation and association -- and I can tell you that I found Leavitt's argument on this persuasive (though as I said there was not enough data presented to be able to tell with certainty.) I could go into detail on that if you'd like, but it would take some time to explain it adequately.

[izzybeans]

35. I agree in all counts...and I must say the argument of unwanted

pregnancy and crime is interesting. However, and at this time I can't confirm because my mother has the book at the moment, I had reservations while reading this particular portion of the book. Maybe its because I've collected documents on the Eugenics movement during one of several false starts that I'm overly sensitive when I hear this resurface (in a sanitized rhetorical form no doubt).

Yet, I too must delineate the difference between spurious and non-spurious results in my work. To my recollection I was very unsatisfied with how he reconciled the variable "unwanted pregnancy" with its inter-item correlation with social class measures. I'm not sure that he did. When I mention logical leap, this is what I'm referring to. I fear he is conflating the two. For instance, let's say my wife and I have an unwanted pregnancy, we are middle class folks working in the education field...Which is the causative factor determining the outcome of my son's life? He threw in quite a number of control variables, which is nice...but with a huge sample I can randomly generate regression models and some will spit me back significant predictors, granted I shall not explain much of the variance but you know nobody cares about that, they just want the significance. Games of chance are mistaken for explanation in the epidemiological arena.

If you could remind me of the source of the data I can rest another weary thought of mine. This is basically due to the fact that there were rampant "accounting errors" or fraud (without the quotation marks)in an unknowable number of official crime reports (from the federal level on down). Given these data are from a period coming out of the "War on Crime" where police agencies were under huge pressures from all levels of government to paint the town "green", all sorts of "creative" "policing" took place. If his data are from victimization survey's on the other hand, than I would trust them. I just don't remember. If not, the dip in crime is likely due to an increase in document shredding and turning the other cheek, not the other way around.

That's pretty much what I remember. I even put the book down at that point. But when I get it back I'm going to have to start over from the beginning because of you :;):

[Time for change]

36. Data sources

As I note in my original post, under the "evidence" section, my first sentence makes note of the fact that there was not enough information presented in this chapter to make a full and independent evaluation of the theory. And by the same token, there was little information on the sources of the data.

I would have liked to see more actual numbers and information on sources than I did, and that is certainly the biggest weakness of the chapter. However, I keep in mind the fact that this book was written for a popular audience, not for scientists. That would not excuse sloppy inferences, of course. But I don't feel that Levitt was guilty of sloppy inferences -- he actually presented a great deal of information on how his inferences were made -- but rather only of failing to present enough information on sources and detailed numbers to allow a scientific reader to come to his/her own independent conclusions. That failure of course is the standard for books that are written for popular audiences, as this was. Authors often feel (and remember that this was co-authored by a journalist) that in order to make their book readable for a lay audience they have to lighten up a bit on the amount of data and sources that they present. So I would never discount a theory based on this fact alone, although it does cause me to have at least some doubts about what Levitt wrote about this.

Nevertheless, I still found the chapter to be quite convincing, for several reasons, many of which I have discussed, but some which I haven't.

You mention the conflation between unwanted pregancy and being poor. As noted previously, Levitt presents a number of risk factors for children growing up to become criminals, in addition to being unwanted, including living in a poor family, one parent family, and having a teenage mother with little or no education. I know enough about public health, and I have read enough on this particular subject, that I don't need Levitt to present his sources to know that all of these factors present both a serious risk to the child's future and increase the likelihood that a pregnant mother will want an abortion and will obtain one if it is accessible. "Crime & Human Nature" by James Q. Wilson, is one good source with a wealth of information on the subject.

I don't look at this as being conflation so much, but rather the fact that all these factors are related. There is no question that a preponderence of abortions (not all of course) are performed on women who have these characteristics.

This of course should in no way be taken as an offending remark to poor and single women. Women who are single, poor, young, and uneducated often have very good reasons for not having a child. The fact that they choose to have an abortion in such circumstances, IMO, is often a wise decision, based on their intuitive knowledge that they do not have the material and emotional resources at that time to be an effective parent.

With regard to the question of whether or not the crime statistics that Levitt used were questionable and could have been based on "creative policing" -- I doubt that very much. Remember, the discussion had to do with violent crime, especially murder. Almost certainly these statistics come from official federal records. I doubt very much that the precipitous and large decline in murder during the 1990s was an artifact of "creative policing".

[izzybeans]

37. There is some discussion of the data sources in this article.

It also provides a different route to getting to my concern.

http://www.va.gov/msg/Joyce.pdf

here is Levitt's coauthored original research on the subject (I'll be reading this today)


http://www.bepress.com/cgi/viewpdf.cgi?article=1028&context=blewp&preview_mode=

Here is their response to the first article (abstract only):

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=385070

Here is a response to there response-an indirect one:

http://www.nber.org/confer/2004/hes04/joyce.pdf


All this should be interesting. I'll probably dig into it when I find the time. A good academic fight over the "logical leap" is always contentious and kind of funny actually. I'll bet both sides will go on for years on this. I'm sure you and I could as well. We seem to be coming at this from very different positions and perspectives.

On what you said: I guess I'm suspicious of government statistics because they are very politically loaded and easily manipulated (in a very non-statistical way), even federal statistics.

I had forgotten that Levitt used the words crime and murder interchangeably at times. That's interesting and I don't know what to make of it. :toast:

[Time for change]

43. Wow, that's a lot of reading material

Thank you for digging that up.

I may get around to this sometime too, but right now 200 pages of dense scientific argument is more than I have time for.
:toast:

[K-W]

10. You are mistaken, this is not eugenics.

Nor does it have anything to do with eugenics.

[izzybeans]

34. No...actually it does. A tenous correlation at best popularized like this

provides those still holding on desperately to the earlier "movement" that sought to sterilize the working poor and the "genetically" inferior. It's probably the most illegitimate justification for abortion I've ever heard. You can switch the rhetoric around all he wants, but those who take the numbers nakedly can spin it however they like. He may not have said the word but others are.

http://www.eugenicsarchive.org/eugenics/

[Time for change]

6. What about the prevention of the birth of an unwanted child?

That is also a major theme of this article. Isn't that the primary reason why most women choose to have abortions?

And what's so bad if a reduction in violent crime is an unintended effect of supporting women's reproductive rights?

Political coalitions of people with different motives is essential to the process of passing legislation. If we said that we didn't want the votes of people who had different motives than us we would rarely get any of our legislation passed.

What would be so bad if crime prevention turned out to be a wedge issue that prevented the confirmation of a U.S. Supreme Court Justice who will likely result in the loss of women's reproductive rights, and who will probably be in his position until the day we die (at least that applies to me).

[MissMillie]

17. Correlation is NOT cause

every statitician knows that.

[Time for change]

19. I don't understand your point

Nobody is claiming here that correlation is cause -- but it is a first step in demonstrating cause.

There are several criteria that scientists use to ascertain when causal relationships exist, beyond showing a correlation. These are discussed in some depth in Levitt's chapter, and summarized in this thread.

[Taxloss]

18. So the decrease in the number of pirates causes global warming?

This is the Post Hoc fallacy. "It wasn't raining - we prayed for rain - it rained - therefore our prayers caused the rain."

[Time for change]

20. Your point seems to be the same as # 17 above

Please see my response in post # 19.

[K-W]

21. To everyone making the correlation comments.

In social sciences it is thoroughly understood that real life doesn't often give us controlled experiments. There is nothing in the slightest bit wrong with writing up an argument based on correlation as long as one doesn't claim to have scientifically pr oven the causality.

So yes, correlation does not equal causation, but no, that does not debunk this piece.

[stanwyck]

24. C'mon, people, common sense

will tell you that children who are wanted and loved and cared for are going to be less apt to commit violent crimes.
Visit prisons, schools, hospital emergency rooms, centers for battered women and children, mental health centers, and drug treatment centers. Talk to the people who have been either the victims of crime or who have commited crimes.
Sure, it's cold to think about...but ask yourself, was this person raised by someone who truly wanted this child? Who took care of the child, emotionally and financially?
The correlation should be your own life experience.
Children who are not wanted, not loved, and not given care are more likely to be the victims of crime and to commit crime.
How difficult is that to understand?
Where abortion is accessible, crime is less. Where abortion is restricted, crime is higher.
It's not eugenics.
It's what we all know. All children deserve to be wanted. The world would lose over 90% of its problems if every child was planned for -- emotionally and financially. No starving children. No children tortured by their parents.
Think of the difference that would make.

[Joan of Arc]

26. "children who are wanted and loved and cared for "

No. 24 -- I couldn't agree with you more. It doesn't take a PhD to connect these dots. My question is, if we truly believe all children should rpt should be "loved and wanted", should it be incumbent upon us to become better wanters?

[stanwyck]

28. That's the core problem

There are those reasons people have children we don't like to discuss. But we all know they exist. We know the parents and their children. These are the parents who caved in to social pressure to have children. Or to parental pressure. Or who thought the marriage would improve -- with children. And there are people so selfish that they want children as a way to complete themselves, as though children are an accessory, not the most demanding responsibility any person has.
There are women who have children to try and hold onto a man, or to get money from the father. As a feminist, I loathe the idea, but we have to admit it happens.
And there are men who impregnate women as a source of misplaced machismo. They don't plan to support the children, but they don't take the precautions necessary to prevent pregnancy.
And there are people whose religion prevents them from using birth control, though they lack the means, financially and emotionally, to give children the care they deserve.
So. You're 100% correct. It's us. We have to "want" better.

[Proud2BAmurkin]

29. That's like saying unwanted kids become criminals

oh wait they do

[Beaverhausen]

30. I think it has more to do with economic conditions

and the economy was doing very well in the 90s.

[Time for change]

31. This possibility was discussed in the book in some detail

See post # 14.

[Time for change]

44. An explanation of the assessment of causal connection

There have been a number of comments on this thread to the effect that although Levitt's analysis of the relationship between the passage of Roe v. Wade and violent crime rate showed a correlation, there was no demonstration of a causal link. This is a misconception because there was much material in Levitt's analysis (much of which I summarized in my original post) that did in fact address a causal link. I'm not saying that the evidence contained in the book proved a causal link beyond a reasonable doubt, rather I'm just saying that I found it persuasive.

I've worked as an epidemiologist in the public health field since 1982, and the main focus of my job has been to assess evidence of causal links on issues such as this one. Here are some of the main criteria that epidemiologists use to assess causal links, and how they pertain to this subject:


Biological and social plausibility
If a correlation between two variables can also be shown to be biologically or socially plausible, based on known information, that increases the likelihood of a causal link between the two. In this case we know that the passage of Roe v. Wade prevented the birth of many millions of children who would have been characterized by risk factors for violent crime, including growing up in poverty and having a mother who was single, very young, and uneducated, and probably most important of all, being unwanted and poorly cared for.

Time sequence
Timing is another very important issue to consider in an assessment of causation. Most important of all is demonstrating that the presumed cause came before the effect. That is obvious in this case, but in addition we have the fact that the decrease in crime began exactly at about the time you would have expected it if the author's hypothesis was correct -- i.e., at the time that the children who were not born as a result of Roe v. Wade would have been reaching their late teens.

Strength of the correlation
The stronger the correlation, the greater the likelihood that it represents cause. In this case the decline in violent crime was not quantified exactly in the book, but from the numbers provided it can be ascertained that the decline was at least approximately 50% or so. That demonstrates a strong correlation.

Lack of plausible alternative explanations
If plausible alternative explanations cannot explain the correlation, then the proposed theory becomes more likely. In his chapter on the effect of Roe v. Wade on the drop in violent crime, Levitt goes into a good amount of detail to discuss other potential explanations for the drop, and after explaining the limitations of these other potential explanations, comes to the conclusion that the increase in abortions following Roe v. Wade was the most important.

Consistency with other studies
Levitt briefly shows how his theory is consistent with other available literature on the subject by noting that the same phenomenon has been demonstrated in Australia and Canada, and that in the 5 states in the U.S. in which legal abortion was available prior to Roe v. Wade, the decline in violent crime occurred sooner than in the other 45 states.

Other evidence of consistency and plausibility
The fact that states with the highest abortion rates demonstrated the largest declines in violent crime, and that this correlation persisted when controlling for other pertinent variables, lends substantial credence to the main hypothesis of a causal connection between Roe v. Wade and the decline in violent crime. This is strengthened still further by the fact that the whole decrease in U.S. violent crime in the 1990s was attributable to the age cohort affected by the legalization of abortion in the U.S.