Steve Levitt has replied to Lott's review of Freakonomics:
Now let's talk about John Lott for a minute. Along with John Whitley, he wrote a paper on abortion and crime. It is so loaded with inaccurate claims, errors and statistical mistakes that I hate to even provide a link to it, but for the sake of completeness you can find it here. Virtually nothing in this paper is correct, and it is no coincidence that four years later it remains unpublished. In a letter to the editor at Wall Street Journal, Lott claims that our results are driven by the particular measure of abortions that we used in the first paper. I guess he never bothered to read our response to Joyce in which we show in Table 1 that the results are nearly identical when we use his preferred data source. It is understandable that he could make this argument five years ago, but why would he persist in making it in 2005 when it has been definitively shown to be false? (I'll let you put on your Freakonomics-thinking-hat and figure out the answer to that last question.) As Lott and Whitley are by now well aware, the statistical results they get in that paper are an artifact of some bizarre choices they made and any reasonable treatment of the data returns our initial results. (Even Ted Joyce, our critic, acknowledges that the basic patterns in the data we report are there, which Lott and Whitley were trying to challenge.)
(Thanks to John Fleck for the tip.)
More like this
This is a long post, so I'll start with two summaries. One sentence summary: It looks as if Lott might have been caught cooking his "more guns, less crime" data.
One paragraph summary: Ian Ayres and John Donohue wrote a paper that found that, if anything, concealed carry laws lead…
In his 6/9/03 posting, Lott claims that Donohue has made a "large number of easily identifiable mistakes". Even if true, such mistakes pale into insignificance compared with the coding errors that Lott made but will not admit to, but let's examine Lott's claims and see how many…
I've discovered another one of John Lott's attempts to rewrite history. Read on.
Lott has written a response to Kevin Drum's summary of Lott's model changing antics. Here's Drum:
1. Lott and two coauthors produced a statistical model ("Model 1") that showed significant crime decreases when…
In most cases, scientific disagreements are resolved in the labs, at conferences, or in exchanges in journals. Sometimes the disagreements are drawn out, sometimes feelings are hurt, but it hardly ever comes to a defamation suit.
Someone forgot to send John Lott the memo.
From the Chicago Tribune…
LOL I love seeing this guy get beat up.
Freakonomics looked interesting enough to order, and if Lott doesn't like it that tends to speak well for it as well. I esp. look forward to the economics of drug gangs; this is a weird phenomenon that I thought Mike Gray covered also covered well from a totally different angle in "Drug Crazy."
too bad for Lott. I wasn't aware that he was taking sides on this issue. No matter how much I personally dislike abortion, even without statistical proof, the thesis that legalization of abortion leads to less crime seems logically irrefutable.
More Abortion, Less Crime. It says something about the nature of the people who have abortions, and for that I'm surprised that the notion is tolerated by the left when they won't tolerate even the conjecture that IQ might correlate with race or sex.
"It says something about the nature of the people who have abortions"
Yes it says they're younger and poorer than the average and they're making a rational decision not to have children when those children are highly likely to grow up in poverty and when their parenting skills are likely to be poor.
No-one on the left with any braisn disputes that IQ correlates with race - they dispute that the correlation is due to genetic factors and not to things like poverty, abuse and drug and alcohol abuse.
Plus IQ measures are imperfect themselves.
I totally agree Meyrick and Ian.
There is at least one specific gene that supposedly correlates to IQ, but only for a few points. I think it should be obvious IQ is, barring physical damage, primarily the result of habit, desire, and access to information. No IQ test can claim to measure innate intelligence because for the most part intelligence is not innate.
Ian and Tall Dave, your comments lead to two ideas. 1) You can't have equal outcomes between different groups when IQ is not equal. 2) This is an undesirable situation that ought to change without infringing on other people's rights.
The situation can, in fact, be changed, and there is a historical precedent showing how IQ can be increased. At the time of the first world war IQ testing (which was a new thing) was used to determine suitability of people for various jobs. Jews, as a group, scored low. But because Jewish culture reveres knowledge and learning, they filled the libraries, created schools, and learned like crazy.
The inner cities' hip hop culture, however, does not prize these things. It doesn't matter where the inner city culture came from, but it's there and it's destructive. If I'm right and culture is the determining factor we have no choice but to change it.
The political right has made mistakes with regard to this, the biggest of which is the war on drugs which has made the drug culture extremely profitable. As much as I hate drugs, we have to look at the fact that this thing has been a horrible failure and created problems for the inner city. It's time to repeal the drug laws. Now, for those on the political left, if they are honestly concerned about their fellow Americans, they have to acknowledge that blaming and punishing white people doesn't help, and that throwing money at a dysfunctional culture makes the problem worse.
And this relates to what Ben said about the correlation between abortion and crime rates. There are plenty of poor people who produce children who become functioning members of society rather than criminals. Second generation children of poor immigrants from places like the West Indies and Asia tend to do very well in America. It's not a matter of poverty or of race -- the only difference between these people and the ones who abort potential criminals is one of culture.
So if you want to improve the outcome, the question becomes how to change the culture.
Stickwick,
Amen to that. People really need to start talking about the culture issue instead of praising everything good or bad about a culture as multiculti heritage and therefore uncriticizable. Similar to your Jewish example, Asian-Americans have a higher avergae income/education than white Americans because their culture values hard work and education.
I also agree the drug war has horrible unintended consequences everywhere from U.S. inner cities to S America. People probably shouldn't use drugs, but then there's a lot of things people probably shouldn't do that have they're free to do as long they don't hurt anyone else. You can't litigate stupid or self-destructive behavior out of existence. You would think the Right would have learned restricting freedom didn't result in greater societal morality with Prohibition, but apparently not.
Sorry, should read:
People probably shouldn't use drugs, but then there's a lot of things people probably shouldn't do that they're free to do as long they don't hurt anyone else.
Yes, but I'm not sure how decriminalization of drugs in itself will change the drug culture.
by removing the profit incentive. And for Ian, you say
Do they dispute the correlations based on science or ideology? Do they dispute the mere conjecture without any logical or factual basis? I do not claim any correlation myself, only that any correlation should be able to be suggested and then tested without cries of injustice from any side. For instance, I could conjecture that white racists, on average, are at the bottom of the IQ scale for all people of all races. Unless of course someone feels my conjecture isn't scientific inquiry, but a strategic attack on some group without concern for the attacks. But how do you distinguish, and is it fair to jump on the person making the claim? In absence of other evidence, benefit of the doubt should be given, I think.
At th risk of moving even away from the initial topic:
Jewish culture has always emphasised the importance of learning. This was reinforced by the bans in most european countries on Jews owning land, farming, entering the military or practicing many trades. The banning of Jews from most areas of public life also meant that Rabbinical studies were often the only avenue open to ambitious young Jews.
What changed in the US, I'd suggest,was an increasing number of American-borm jews who were fluent in English.
On the subject of "hip-hop" culture I'd point out that the economic and social disadvantage of American blacks predates hip-hop by several centuries.
At the same time it's instructive to look at Dominican, Haitian and West African immigrants in the US. These group start out with some major disadvantages compared ot African-Americans - lower education standards; less money and language difficulties for starters.
They are probably subjected to as much or more racial discrimination as Black Americans. (Two of the most notorious incidents of police brutality in New York City in recent years involved black immigrants.)
However in some areas - such as business ownership and home ownership - members of these groups who've been in the US for extended periods outperform black Americans.
So it seems there must be specific factors - probably cultural at owrk within the African-American community. But somehow I doubt they're as simple as "hip-hop culture".
Of course, its possible there's a sort of adverse selection at work here too. Immigrants who are convicted of crimes get deported, American citizens aren't.
Might I point out that the lowest ranking group in intelligence tests for the US army in WWI was Jews from Eastern Europe. African Americans living in northern cities had higher scores.
Eli
Yes - and blacks born in the south who migrated to the north showed higher scores the longer they'd lived in the north.
Yes, but I'm not sure how decriminalization of drugs in itself will change the drug culture.
I'm not sure how criminalization of drugs in itself has changed the drug culture, besides making drug dealers rich. It's very hard to prevent a consensual private transaction between two people, and I think it requires a very strong public safety imperative to even try.
If they spent of tenth of what they spend on incarceration (and that is a HUGE number) on simple education instead, they could probably do a lot more good than they're doing now and without creating horribly destructive market distortions. And the distortions are really frightening when you look at them: cocaine sells for 100 times what it costs to produce in Colombia. You create a profit opportunity like that, and someone is going to seize it one way or another.
Dave
Decriminalisation would probably make the drug dealers poorer - not richer.
Reduce the barriers to market entry (i.e. the risk of incarceration) and competition will increase.
Greater compeittion will lead to lower prices.
Additionally, if you can keep drug users out of jail and out of the criminal justice system their quality of life will probably increase significantly.
Dave
Decriminalisation would probably make the drug dealers poorer - not richer.
Reduce the barriers to market entry (i.e. the risk of incarceration) and competition will increase.
Greater compeittion will lead to lower prices.
Additionally, if you can keep drug users out of jail and out of the criminal justice system their quality of life will probably increase significantly.
I don't doubt that legal abortions has an effect on the crime rate. However, it's a big leap from that to the conclusion that abortions should be legal. If you think that the foetus should be viewed as human (at any point in the time period it is now not viewed as human), then the fact that it reduces crime make abortions no less acceptable than say, locking up all men from age 18 to 25, which would probably reduce crime a lot more.
Also, while it is OT, regarding drug legalizations: Selling drugs is something inherently nasty, because the biggest share of profit by far comes from people who ruin their lives utterly on it. (The same is true for alcohol and gambling, by the way). So my thought is that a legal drug gang, while it might become more business-culture and less street-culture, would be just as ruthless and amoral as an illegal one, and probably more effective at making people suffer on a big scale.
Harald has put his hand on the central point of debate here. For some (including me) information is one input into decisions, sometimes a deciding input, sometimes not, but you don't get to pick your facts. A good example of this is the base discussion here of whether guns freely available in homes increase or decrease crime. The answer appears to be that there is a weak effect, which in a reasoned debate would be informational, but not determinative.
However, for a lot of folks the answer is clear and this drives them to omit, exaggerate, and yes, even lie. We then enter the gladiator's arena, where no one can trust the other and where today's ally is tomorrow's foe.
Harald appears to have opinions on the issues he has raised, but recognizes that facts are facts. One can deal with that
(btw, my impression is that the vast majority of drug users are recreational and that drug use has little effect on their lives c.f. marijuana, same with gambling. The major issue in those cases being the bingers as with alcohol. So my answer would be to try and find interventions to help those who get into trouble, while leaving the vast majority to their small pleasures...but I am open to discussion:).
Harald
So I take it you advocate the prohibition of alcohol, tobacco and all forms of gambling?
I am no fan of Lott, who sometimes brings bad science to what might be good causes. Just because Lott says a theory is true does not imply that it is false.
Is Tim Lambert endorsing Leavitt's "Abortion on Demand allowed pre-natal capital punishment of unwanted African-American children" thesis? I would be wary of going down that track. Leavitt's thesis is has been empirically refuted by Steve Sailer.
Leavitt's thesis is obviously inconsistent with the crimes statistic evidence which shows a massive upsurge in urban minority youth crime during the Crack Wars (1986-1994). This occurred within twenty years of the abortion liberalization (1970) ie precisely at the time when the potential criminal-culling effects of cheap and easy abortion were most likely to be observed.
This is interesting as Sailer is often identified with the bio-conservative (endowment-hereditive) view of urban minority pproblems. Whereas his evidence shows that the strong version of the socio-constructive (environment-acquisitive) view has legs.
Jack, I'm not impressed by Sailer's "refutation". Levitt may be wrong, by Sailer has not given any good reason to believe this. I suggest you read all of Levitt's post that I linked.
Eli, how large a part of the gambling/alcohol/drug lord's income comes from the chronic abusers is a question which can and should be answered scientifically. I know that the norwegian and swedish public health authorities have some research on this - the swedish one claims that 9% of gamblers account for 52% of the profits. This fits your picture that most of the gamblers have good control. However, it also fits my picture that those who profit from this have a huge vested interest in only giving token support to reduce gambling problems. If it wasn't for the abuse, gambling wouldn't be very profitable. I believe the same goes for alcohol, and i'm 100% sure it does for heroin.
Ian, I certainly advocate the _abolition_ of these things, not necessarily prohibition, although that might be sensible in some cases. Btw, I do not think smoking fits this skewed graph I've been trying to describe. With smoking, the damages are more evenly distributed.
ad hominem, Jack Strocchi
trying to discredit Levitt's findings by spinning it with the possibility of a eugenics reading makes no difference whatsoever to their truth value. you are beginning to sound like the PC lefties you so despise
Tim Lambert 24/5/2005 11:46:11
Tim, forgive me, but I am not impressed with your not impressed[ness] with Steve Sailer's arguments. Lott is another matter. He is a partisan of the Right wing of the Republican party whose work is unreliable, as you rightly point out. Steve Sailer, by contrast, has been excommunicated from the Republican party because he prefers to speak truth to power, most notably in his prescient, and validated, criticisms of the recent Gulf War.
Now to issues of substance. Leavitt thesis appears, in some respects, to be blatantly contradicted by facts. Steve Sailer has given very good reasons to disbelieve in it by pointing out facts inconvenient to the "Abortion cuts Crime" thesis. No one in this seminar seems to have bothered to glance at a certain 800 lb gorilla-sized fact that is squatting on the living room of this thesis,
The Levitt thesis can be boiled down to a syllogism:
Major (Theoretical) Premise I: Children who are planned, wanted and cared for tend to be more civil, less feral and therefore commit less crime.
Major (Theoretical) Premise II: Abortion will tend to increase the ratio of cared-for civil, as opposed to uncared for feral, children for any given cohort within a population.
Minor (Empirical) Premise: Legal Abortion, started in the US from 1970+, did in fact reduce the ratio of uncared for feral children that grew to maturity in the post-abortion generation (The pro-Levitts seems to view this development as a kind of pre-natal capital punishment for potentially feral and lethal children.)
Conclusion: Major crime in the US urbs started to fall during the nineties.
There is a huge hole in Levitt's theory: the largely teen-, & twenties-aged, cohort who came to maturity in the mid-eighties to the mid-nineties were the class of persons supposed to be helped by the greater likelihood of being raised by caring, family planning parents. But a largish fraction of this cohort went on one of the most blood-thirsty murder & crime rampages in US history.
Apparently a statistician "driven unflinchingly by the data" has failed to notice this. It sticks out like the proverbials, especially to those persons (like yours truly) who were actually in the US when the mayhem was still happening (1993) and could not help noticing the on-going civil war.
I would not argue with the intuitively plausible Major (Theoretical) Premise I (Planned and cared for kids tend to be more civil). The Major (Theoretical) Premise II (Abortion reduces the feral-to-civil ratio of children) is more controversial. There may be race- and class-differences associated with this process but this is BTP.
Steve Sailer would definitely contradict Levitt's Minor (Empirical) Premise (Abortion did in fact reduce the US's feral-to-civil child ratio) and Conclusion (Crime went down in the post-abortion decade). These propositions are flatly contradicted by the data.
In the post-abortion era the US's ratio of out-of-wedlock births (and therefore unwanted and uncared for children) continued to rise, soaring from from 10% in the early 70s to 30% in the early 90s of all children. No doubt this trend was the result of complex causes, but there is no evidence that, pace Levitt, that the rate of illegitimacy (and hence criminality) slackened off in the first post-abortion liberalization generation-maturity period (1985-1995).
I do not have the figures to hand but my impression is that, over the key (1970-1980) period, overall US birth levels (household X birthrate = volume effect) did not fall so much so as to counteract the increased rate of illegitimacy (two-parented/unparented = ratio effect). The theoretical possibility is that abortion had drasticly reduced the overall size of the (more crime-prone) US youth cohort which might have helped reduced total crime rates. I would welcome empirical correction on this point.
But this possibility looks unlikely given the 800 lb gorila-sized fact that the homicide rate for the key 14-17 year old demographic started to soar during the mid-eighties. This class was the US's first post-abortion cohort to mature and they were the foot soldiers in the Crack Wars. Overall crime rates also soared up during this period, particularly in the high-abortion rate jurisdictions such as WDC, LA and NYC. So like, what happened to the civilizing effect of abortion?
Levitt finally bowed to public pressure and adressed the empirical problem with his theory. But hr did so inadequately using complicated immunizing strategums to protect his headline theory from evident refuation. Sailer has no trouble in making mincemeat out of Levitt's intellectual contortions. Sailer concludes by saying:
Sailer also notes that the burden of proof should be on Levitt to provide a less-leaky and more straight-forward case when making such a large and sensational claim. Levitt obviously has clearly failed to meet it. But many academics, including a few of the (other-wise) worthies in this seminar, seem to be in such awe of this wunderkinds reputation that they have not been game to call him on it.
Sorry Jack, but Levitt is correct when he says that you need to separate out the effects of the crack cocaine epidemic. Sailer doesn't, that's why I don't find his graphs persuasive. It is ridiculuous for you to claim that Levitt failed to notice the upswing in crime in that cohort. Freakonomics cites his 2004 paper which addresses this issue.
Out of wedlock is not equal to unwanted, uncared for.
Tim Lambert 26/5/2005 23:02:14
No need to apologise on my behalf, Tim. I can manage to control my contrition levels on this one without your unsolicited help.
Lets get this straight: we students and law enforcement agents can all relax now so long as we remember to define the crack cocaine gang wars soley as a public health problem? If only all scientific criticism, never mind social problems, could be swept away by such lexicographical sleight of hand!
The US Dept of Justice, might they not know something about crime? Here is one of their graphs depicting recent trends in US crime. Note the massive hump in the crime rates for the 14-17 and 18-24 year olds, which starts to up tick in 1985 and peaks in 1993. Most of the members of these cohorts, many of whom participated in this crime wave, were born after abortion was legalized.
Sailer actually does, here and there, consider the other factors complicating the analysis of US crime rates, such as the introduction of crack and political trends in wealthfare [sic] ("greed is good"), welfare ("ending welfare as we know it...") and lawfare [sic] ("three strikes...") ie the class, culture and caste wars that raged in the US over the same time. He concludes that social science has not got a definitive answer on the relationship b/w abortion and crime. But any inverse relationship b/w them cannot be very strong if it was early and easy swamped by other factors. But what would Sailer know, since he is only a Keep It Simple Stupid (KISS) blogger, unlike the Keep It Complicated Smartie (KICS) academics?
Sailer also rightly points out that in science (as in law) the burden of proof should be on the grand theorist making bold claims, rather than petty skeptic pointing out thorny problems.
I did not claim that Levitt "failed to notice" any problems with his theory. I suggested that, for quite some time, Levitt was tardy in properly addressing what we might politely refer to as "the empirical problem" with his thesi, which Steve Sailer initially flagged way back in 1999. This does not seem such a "ridiculous" assertion to make since it is only in the last year or so - a mere five years later - that he finally pulled his finger out as the book was being marketed.
It would be nice if the academic community subjected Levit thesis to more searching criticism instead of gaggling around him in star-struck awe and ganging up on blogging lightweights who dare to play the satorial-advising child to certain skantily clad academic emperors. Its a dirty job but...
Jack, unfortunately crime is complicated and there are multiple factors that effect it. Ignoring those factors does keep it simple, but is not going to let us what each factor does. That is why Sailer's criticisms have no bite with the folks who are expert in econometrics. The sort of criticism that will requires someone to do their own multivariate analysis. That's Joyce and Lott. And Lott is untrustworthy, so really it's just Joyce. I understand that Joyce has a new paper coming out, so we'll see how that goes.
Journal's have a long lead time so his 2004 paper was probably written in 2002, while the book was not being marketed till 2005.
Tim, all the issues dealt with in your blog reminds me of the old saw that there is no problem so complex that it does not have a simple but wrong answer.
Tim Lambert 27/5/2005 17:41:10 writes:
I couldnt agree more with your formal analysis of the methodological situation. Complex phenomenon require analytic techniques that are capable of complexity reduction - multivariate regression analysis - to tease out the causal power of several factors. In fact this is the very point that the critics, such as Sailer, are making!:
My beefs with Levitt are partly polemical and partly methodological. Many academic critics seem to have cut him alot of slack owing to his status as an intellectual Brahmin, peer-reviewed and public-adored. So much so that the burden of proof seems to have shifted onto the lumpenintellectual untouchables, such as Sailer, who are disreputable, "have axes to grind", have not done econometrics, are mere bloggers etc. This attitude seems overly deferential to Levitt since he has made his name out of putting contrarian theories out into the intellectual marketplace. We all know that contrarianism is the place where cranks thrive.
I am not saying Levitt is a crank, he has a consumate grasp of econometics. But he also has a healthy appetite for applause.
Levitt is the one who appears to be making hay out of the grand simplistic, and perverse, claim here. This leads him to the position where he will try to have his "Occkams razor headline grabbing" cake and eat his "complex problems relegated to footnote" too.
In his headline treatment of this theory he seems to be suggesting that abortion is more effective than anything else in cutting crime.
When criticised he huffily replies that the problem is complex so these people need to do multivariate analysis. He acknowledges that crime is complex and many factors are at work, but this is relegated to the footnotes.
So in marketing "Freakonomics" Levitt says "hey mass abortion - way to cut crime dude!" But in boring old economics he concedes the more ho-hum "abortion may cut crime, or then again it may not" conclusion may have some merit. "Freakonomics" is a catchy title. I dont think Levitt would be happy trying to flog "Ho-hum-onomics".
Levitt said referring to the availability of abortion: "this dampening effect on crime can be outweighed in the short term by factors that stimulate crime" such as the crack epidemic
From which Jack concludes that Lettit says that abortion may cut crime or not. Jack obviously believes in one handed economists, though such a thing has never been found because of the many factors involved in life. What Levitt says is that by itself the availability of abortion cuts crime, however there are other factors which increase crime and which may be stronger so that the net increases. Hey dude was that so hard to understand or you just enjoying a trip in the spin-o-matic?
Eli Rabett 29/5/2005 03:37:07
The Levitt-critics are the two-handed pluralists. My criticism of Levitt is that he is the one who is running the spin-o-matic into overdrive. Why is that so hard to understand, dude?
Cause you have added spin overdrive to the spin-o-matic there fella. So let us look at what Levitt said again for the don't wanna hear crowd:
Levitt said referring to the availability of abortion: "this dampening effect on crime can be outweighed in the short term by factors that stimulate crime" such as the crack epidemic
This says
1. Freely avaiable abortion decreases crime.
2. There are other things that increase crime (like crack for example)
3. The balance between 1 and 2 can be positive or negative.
Now Jack baby, you wanna stick your fingers in your ears, go ahead, but don't expect us to do anything but laugh at you.
Eli Rabett 29/5/2005 11:47:08, perhaps off his medication, swings from denial to projection:
This is good advice, but I suggest Eli take it himself before doling it out to the less needy. Levitt stated that legalised abortion's most powerful impact on social pathology would be to reduce the total quantity of the crime-prone youth cohort and increase the quality of cared-ness amongst the survivors of "pre-natal capital punishment".
No-one, not even Eli, has denied that there was a massive increase in percapita offending rates for the first post-abortion cohorts. Otherwise known as "the crack wars". Now I know that Levitt also says that other factors may have off-set this. My point is that Levitt has never got around to explaining why abortion could turn from such a magic elixir to snake oil once things get complicated. This kind of, more explicit attention to detail, would substantially weaken the explanatory power, and hence sales-grabbing potential, of the crime-cutting-abortion headline thesis. Apparently we will have to await the publication of Joyce's paper for a fuller accounting for this problem.
That still leaves Levitts thesis standing on one other leg: abortion as a way to reduce the size of the crime-prone youth cohort. But it turns out that this leg is a little shaky too. What happened to the massive improvements in family planning, and hence wantedness of children, that legal abortion was supposed to deliver? They did not occur. Sailer points out that legalized abortion did not result in better family planning, especially in the case of underclass females who, for obvious reasons, were more prone to unplanned pregancies:
To put it bluntly, legalized abortion increased bastadry at a much higher rate than it reduced the overall rate of birth. Eli, if he had been listening to the facts instead of laughing at his own jokes, might have heard Levitt himself let slip on that one, relegated to yet another of his blessed footnotes:
Moral hazard? I thought abortion was supposed to reduce hazards! The sorry facts make this bland euphemism sound rather mordant.
Underclass females experienced a very high proportion of these (unplanned) conceptions, legal abortions and (out-wedlocked) children - which is not exactly how things were supposed to work out according to abortion law reform mythology. And of course these broken families were more prone to bring up their fatherless children in a high-crime environment, leading to a higher youth crime rate. Which is exactly what happened when the first underclass youth cohort started to hit the streets, and bite the dust, in the mid-eighties.
The fact is that US out-wedlocked children rates continued to rise up till the nineties which makes it hard to see how this, more meaningful, measure of the size of the crime-prone youth cohort can be used as a predictor of serious crime rates. Since the nineties US abortion and out-wedlocked rates have declined, so is Eli ready to bet that US crime rates will rise in the twenty-teens? And why did CIS crime rates skyrocket in the nineties, a generation after relaxed and cheap abortions became available? Yes, I know, complicating factors, Blah X 3...
These unpleasant facts do not fit the nice little contrarian story that Levitt has concocted. Many people, on both Left and Right, would prefer to hear only happy-talk about the social benefits of legal abortion, curbing population growth and killing off large fractions of potential criminal cohorts. They have been pleased to turn a deaf ear to empirical criticism of Levitts thesis, perhaps because the race-eugenic aspects of his thesis are a little too confronting.
However it is probably unkind to burden Eli with all these weighty matters. So perhaps we will let him keep his fingers firmly plugged into his ears. That way he would be spared the embarassment of having to listen to the derisive peals of my mirth, tapering off in the distance...
"...legalized abortion increased bastadry at a much higher rate than it reduced the overall rate of birth"
Thank you, Jack, for that textbook example of the post hoc ergo proctor hoc logical fallacy.
Until (smart people) come face to face with someone who is tenacious enough to dissect their logic, and resilient enough to endure the thinly veiled intellectual abuse they dish out during debate (e.g. "You don't really think that do you?" or "Well if you knew the <insert obscure reference here> rule/law/corollary you wouldn't say such things"), they're never forced to question their ability to defend bad ideas.
http://www.scottberkun.com/essays/essay40.htm
Hmmm... Sailer has linked to comments here, writing:
This whole debate seems to boil down to whether Levitt properly controlled for the effects of the crack wars (and other factors) on the crime rates. If he overestimated the crime increasing effects of the crack wars and other factors in the 1980s then he also overestimated the crime reducing effects of legalized abortion. Whether or not Levitt properly controlled needs to be studied seriously.
Simple question: How could "wantedness" have risen if the illegitimacy rate skyrocketed?
I mean, if Dad doesn't want to be married to Mom and Dad isn't around then, hey, junior is a lot less wanted, a lot more likely to grow up in poverty, a lot less likely to get medical care, a lot less likely to be disciplined out of criminal activity as a teenager, a lot more likely to be physically abused and sexually abused by other adults in the household, and so on and so forth.
This isn't rocket science kids. Abortion probably helped cause the rise in illegitimacy. So abortion decreased wantedness, not increased it. But regardless of why the illegitimacy rate rose "wantedness" definitely fell after Roe v Wade.
Levitt's argument about "wantedness" is bogus. He doesn't want to consider selective effects on genes that influence criminal behavior. So he goes all P.C. and talks about "wantedness".
As for whether abortion increased or decreased crime rates - regardless of the mechanism - how can he tell? There were too many variables at work. The big increase in crime in the late 80s and early 90s caused a real "Empire Strikes Back" effect on many levels designed to reassert control. Crime was a big deal politically. This helped the Republicans score impressive electoral gains and drove judicial appointments too. The disgust with crime drove a massive prison building program, numerous changes in policing techniques and police technologies (and computer advances did too), sentencing law changes, changes in attitudes and practices of judges and prosecutors, and so on. The welfare reform was passed in part because of the anger about crime and welfare.
So Levitt thinks he can model society with some equations and some subset of social science variables and pull out a signal for the effect of abortion on crime after the big run-up in murder (which parenthetically killed a lot of criminals) and other crime? Ha, how naive.
BTW, where are these claims coming from that Jews in America scored low on IQ tests early in the 20th century? There is no evidence to support this belief. Are you guys still deceived by the misrepresentation of Leon Kamin and Stephen Jay Gould of Henry Goddard's research? Their misrepresentations have been debunked.
Here's what Goddard actually did and found:
Goddard's original paper is somewhere on the web. But I can't find it at the moment.
Levitt's defenders seem to be saying here that the unprecedented criminal rampage in which the first cohort that should have been gloriously eugenised by abortion engaged in when they hit their mid-teens does not make a mockery of his thesis, because it can be explained by an unusual circumstance, namely the crack war. Their criminality was way lowered, you see, but then this other little external factor interfered and it raised their criminality, temporarily cancelling or overwhelming the benefits of abortion. As if the crack phenomenon was some kind of novel contagion that raised criminality but had nothing otherwise to do with criminality itself.
The bulletin is surely this: smuggling, cooking, selling, and using crack are criminal acitivites. When all those kids decided to make their living by supplying it, and to kill people that got in their way, or to waste their lives using it and rob people to make that possible, they were expressing their criminality, the precise sort of thing that should have been minimised by the cull, according to Levitt. It's an obvious nonsense to say that these decisions can neutrally be taken to 'explain' the rise in criminality that followed them, as if by doing so you get to say that the rise in criminality wasn't 'really' about that generation being more criminal than its predecessors, and get your man off the hook that way. That Levitt's admirers think that the mass epidemic of crack-driven crime somehow preserves the thesis rather than aborting it shows that something other than reasoned enquiry is motivating them.
Randall Parker is linking to a site http://www.ziplink.net/~bright/. If you enter that url in your browser you are forwarded to http://www.eugenics.net/ which is, well dedicated to eugenics, in the classical racial sense.
In any case the quote about what Goddard did and did not do are contradicted in
Psychological Testing: History, Principles, and Applications, 4/e
Robert J. Gregory
Wheaton College
http://www.ablongman.com/partnersinpsych/PDFs/Gregory/gregory_ch01.pdf
"What did Goddard find and what did he make of his results? In small samples of immigrants (22 to 50), his assistants found 83 percent of the Jews, 80 percent of the Hungarians, 79 percent of the Italians,
and 87 percent of the Russians to be feebleminded, that is, below age 12 on the Binet-Simon scales (Goddard, 1917). His interpretation of these findings is, by turns, skeptically cautious and then provocatively alarmist. In one place he claims that his study "makes no determination of the actual percentage,
even of these groups, who are feebleminded." Yet, later in the report he states that his figures would only need to be revised by "a relatively small amount" in order to find the actual percentages of
feeblemindedness among immigrant groups.
Further, he concludes that the intelligence of the average immigrant is low, "perhaps of moron grade," but then goes on to cite environmental deprivation as the primary culprit. Simultaneously, Goddard appears to favor deportation for low IQ immigrants but also provides the humanitarian perspective that we might be able to use "moron laborers" if only "we are wise enough to train them properly."
So there appears to be more than a little disagreement here
Mark, two posts above, beat me to the punch, but just let me ask once again: Was the crack epidemic of 80's and early 90's an act of God that totally overwhelmed the soi-disant benefits of abortion? Or is it possible, or dare I say, likely, that legalized abortion, which led to HIGHER rates of illegitimate births in the underclass, contributed to the creation of the crack epidemic and its accompanying massive increase in the murder rate?
I'm an amateur here, and I swear I don't have a dog in the hunt, but I'd appreciate it if the defenders of Levitt's "unwantedness" theory could get me started on one issue.
Granted, (1) Levitt's theory has some intuitive appeal, and (2) the crack crime bulge appears to conflict with Levitt's theory. Given that:
1) What does Levitt do to test his theory in general? What evidence does Levitt find to support his theory?
2) How does Levitt determine what portion of the crack crime wave was caused by external factors, and how does he conclude that similar factors didn't result in the reductions in crime that he attributes to the abortion of unwanted children?
Thanks
Nevermind - I've read the link Tim provided and think I understand the issue.
Tim Lambert Says: May 27th, 2005 at 5:41 pm
Hmmm...looks like "Levitt's Bulldogs" have some explaining to do. It appears that some "folks who are expert in econometrics" have indeed found some "bite" in "Sailer's criticisms". Steve Levitt's "abortion-cuts-crime" theory has finally received the come-uppance it richly deserved. The WSJ reports the findings of a Boston Federal Reserve Bank study which comes up with the thumbs down for Levitt's headline work:
I love the droll reference to "the crack wave that waxed and waned...during the 1980s and 1990s", in direct contradiction to Levitt's theoretical prediction. One does not have to be "Sailer's Bulldog" to notice this. I was in the US during the early nineties and I would have had to be in a coma to miss the drug-fuelled Lord of the Flies being played out on inner-urban US streets.
The only reason that Levitt's equations failed to pick these facts up is because Levitt wanted to be a celebrity more than a scientist and so chose to ignore them. And sooooo many other social scientists were happy to heap praise on the wunderkind, perhaps hoping that some of the grace of god would rub off, with no one willing to spot flaws in the Boy Emperor's attire.
Its way past time that a tenured academic singled out Levitt for a pasting, since we highly un-paid blog carpers are feeling under-gunned. Tim Lambert has provided a splendid public service in debunking the fallacies and fabrications of charlatans like Fumento and Lott. Perhaps he will now turn his critical eye on some of falsities trotted out by Standard Social Science modellers, not to mention the embicilities of the anti-Darwinian Left, and give Levitt the blowtorch treatment.
I hope so, since l'affaire Levitt is an academic scandal. It is unlikely that the celebrity wannabe academics who feted Levitt will turn on their pride and joy. As Sailer points out:
Jack, Foote doesn't mention Sailer at all. It is truly bizarre to suggest that he found some bite in Sailer's criticism.
Levitt's initial response is here.
I'll have a post on this matter when I've worked out what is going on with the models.
Tim Lambert Says: November 29th, 2005 at 4:23 pm
Well we never did find out the name of the boy who called attention to the defects in the Emperor's attire. But the Emperor was nonetheless naked for all of that.
Foote refers to the crack wars ten times during the course of his critique of Levitt. This is the brute fact which provides the "bite in Sailer's criticism". It certainly puts a massive spanner in the works of Levitt's model, which predicts that legalizing abortion in the early seventies should have reduced youth crime in the late eighties/early nineties.
It is "truly bizarre" of Levitt's camp to be in denial of this, no matter how much they may tinker with their models.
Don't be silly Jack. If Sailer had influenced Foote he would have cited him. It's not as if Sailer was the only person to have noticed the crack wars.
It is also silly for you to claim that this is some sort of scandal instead of the normal to-and-fro of papers about a particular theory.
Tim Lambert Says: November 30th, 2005 at 2:19 pm
Don't pick nits in parsing Tim. Intellectual history is full of tales of researchers who did good work but were undiscovered or uncredited by the expert authorities. Sailer's criticisms can be as uncited or unnoticed by as many econometricians as you like and still pass scientific muster ("have...bite") in principle. That is, his analysis has been independently confirmed by subsequent "folks who are expert in econometrics", whether the researchers based their work on him or not.
But I doubt very much that Foote's paper was blind Sailer's criticisms of Levitt. A google search of "abortion+crime+Levitt" returns Sailer's name in five out of the top six most viewed pages. John Donohue's paper was the most significant academic critique of Levitt and lists Sailer in the acknowledgements. So Tim is kidding himself if he thinks Foote was ignorant of Sailer.
But this is not about Sailer's priority. It is about the social science professionals gullibility.
I searched the Crooked Timber shrine, er seminar, on Levitt and could not find a single reference to the Crack Wars as evidence pertinent to his "abortion-cuts-crime" theory. This is not exactly encouraging amongst people who are paid to notice such things.
It is silly of Tim to pretend that the Levitt's abortion theory, and the controversy it generated, was just the "normal to-and-fro of papers about a particular theory". In fact Levitt's theory is the headline work of the most lauded young social scientist of the naughties. It is a major academic intervention into the heart of the Culture Wars, making a contrarian appeal to both the ("pro"-abortion rights) Left and the ("anti"-young minority-group male) Right. The book was distributed and spun in a very un-academic way, using every marketing trick in the book. It was the subject of a major political gaffe. This is not a run of the mill academic debate dealing with a mundane problem in a routine way.
And the vast majority of social scientists, journalists and politicians fell for it uncritically. It does not inspire me with confidence that so many bright people have dropped the professional ball for personal or political reasons. It also makes me wonder what it takes for a scholar/theory to become reputable, or disreputable, in the eyes of the social science community. (No names, no pack drill.)
I should re-emphasise that Tim Lambert is not a "science professional" who I would call "gullible" or was one of those who "uncritically...fell" for Levitt's dodgy work.
The most celebrated nonfiction book of the year is "Freakonomics" by U. of Chicago superstar economist Steven D. Levitt and journalist Stephen J. Dubner. The most admired aspect of the book has been Levitt's theory that legalizing abortion cut the crime rate. Now, it turns out, according to two economists at the Boston Fed who have checked Levitt's calculations in detail, that Levitt's theory is based on two mistakes Levitt made. So far, Levitt admits to making one error, saying it "is personally quite embarrassing."
This fiasco reveals much about what's wrong with public policy discourse in modern America. Fifteen minutes of Googling would have shown that the abortion-cut-crime theory hadn't come close to meeting the burden of proof, but, instead, much of America's intellectual elite fell head over heels for it.
Ever since my 1999 debate with Levitt in Slate.com, Levitt's fans have been telling me that my simpleminded little graphs and ratios of national-level crime trends showing, for example, that the teen homicide rate tripled in the first cohort born after Roe v. Wade couldn't possibly be right because Levitt's state-level analysis was so much more gloriously, glamorously, incomprehensibly complicated than mine, and Occam's Butterknife says that the guy with the most convoluted argument wins. Now, two economists have finally redone Levitt's work and found two fatal mistakes in it.
The WSJ reports:
'Freakonomics' Abortion Research Is Faulted by a Pair of Economists