From a loyal MR reader

I read through the Heckman debate. He does what he always does. One response is terrible (quality of early intervention doesn’t matter, just do a lot of it) but most of them make decent points. Carol Dweck hints at the problem of writing people off.

No one considers neurodiversity. No one considers that many successful people take big risks, follow their impulses, fail to comply, have bad habits, and otherwise misbehave. Heckman himself may be an example.


Can we all get our banal observations blogged, or do we have to mood affiliate with Tyler's hobby horses?

ahhh....more econo-junk speak: "neurodiversity."

Apparently it's PC for autistic.

al, neurodiversity is not econ-junk speak. It is actually at odds with the common econ framework of a representative agent (a more one-size fits all actor). I believe the idea is that there is considerable heterogeneity in ways of thinking across people and what's productive in one setting may be unproductive in another and vice versa. I don't quite get the context here, but the reader may have an issue with the standardized interventions.

otherwise misbehave. Heckman himself may be an example.

Because some of your readers may not know the context, Heckman is both brilliant and a notorious difficult personality, so much so that places have been reluctant to hire him despite his astounding productivity. (He was the "menace" cited in this hilariously transparent non-named reference here.

But I don't get the reader's comment. It's not like teaching kids to misbehave (whatever that means here) would make them more successful (however they measure that). A lot of super smart people 'get away' with eccentric behavior because they are smart and successful. The annoying ones are people who act eccentric to signal that they are smarter than they are, but those are easy to spot. It's easy to teach eccentricity or bad behavior, but that's not going to make them smarter or guarantee good life outcomes.

I don't think he is making a policy statement, just noting that the "non-cognitive skills" Heckman describes may not have as strong a predictive effect as we might like. This doesn't seem unreasonable to me. High tolerance to risk, impulsiveness, and disrespect for authority (among other "problems") could actually have a positive influence when combined with cognitive and social skills -- I'm reminded of the psychology literature connecting leadership ability with psychopathy. This is, of course, much different from smart people getting away with bad behavior.

I don't think the reader would argue that we should teach kids to misbehave, just that the existence of misbehavior does not necessarily represent a harmful lack of non-cognitive skills.

But it is a (subtle) policy statement. Heckman is arguing for interventions that support certain non-cognitive skills and the reader is saying well some of the counter skills are actually good for people (including by assertion Heckman himself). Of course, I don't think this person is suggesting lessons in bad behavior, just more of a live and let live mentality. It is my understanding that our cognitive and non-cognitive skills have many dimensions and we can often use strengths (if we have them) in one to offset weaknesses in another. I am just not sure that these low income kids are well served by neuro-diversity concepts, since we (and their parents) are already saying that their point in the socio-economic-diversity spectrum is not desirable. But I agree with Heckman (and Murray) there's a lot to be learned from studying the non-cognitive / environmental aspects.

So, top 10 economics departments are just like primary and secondary schools who tolerate bullying to keep the whipper snappers in line. I wonder if the bullies tell younger ones "hey, I gotta be me, this is a great place to be unique!" (hahaha)

Andrew' how did you ever get that out of my comments? I disagree (of course). There are bullies (or pushy people) everywhere though their methods may differ. Did you know some people even bully with humor?

The joke being made is that Heckman's ornery personality might well be exactly what these kind of government interventions in personality development advocated by Heckman would try to hammer out of little boys in their clutches. Fortunately, nobody managed to pound down Heckman's personality because it, combined with his stratospheric IQ, has made this Nobel winner one of the great social scientists of our time.

An interesting part of the exchange is the way Heckman unmasks Murray's total disingenuousness, distortion of the research record.

Nonsense. Murray is correct that Heckman touts the results of small, unreplicated studies. The history of intervention studies is one of failure.

Heckman's response is that, sure, most interventions have failed, just like most proposed medicines fail. That doesn't mean that all of them have failed in the past or will all fail in the future, just as it doesn't mean medicines never work.

I appreciate Heckman's presentation and dissecting of the evidence very much. It is plausible that the educational system is not the cause of all the problems mentioned [but perhaps that the educational system has responded endogenously to those problems]. Nevertheless, I do not see the linked article getting to the root of anything.

Oh, CSF, I wouldn't call Murray disingenuous or distorting, merely naive, perhaps uneducated.

"merely naive, perhaps uneducated"


I wonder how familiar these commenters are with Murray's work... Yes, in The Bell Curve, Murray focused largely on IQ and genetics, but for decades Murray has been talking about how much culture matters and about how much the lower classes have been hurt by cultural shifts. What do you think Coming Apart was about? It focused on how changes in lower class society have led to worse outcomes - changes like the disappearance of fathers, babies out of wedlock, loss of work ethic, etc. Murray understands as well as anyone else that culture and childhood environment matter in determining life outcomes.

Here is a quick run-down of Murray's position:

Regarding cognitive ability:
- Intervention programs like those Heckman supports have little lasting impact on cognitive ability, and cognitive ability is largely heritable

Regarding the socio-emotional factors Heckman discusses:
- The lower classes can benefit enormously from positive changes socio-emotional abilties/lifestyle/culture, just like they have suffered greatly from negative changes in their socio-emotional abilities/lifestyle/culture over the past several decades (e.g. what he wrote about in Coming Apart)
- However, the sorts of social intervention programs Heckman advocates have been ineffective at achieving this goal. Decades ago the lower classes had better values and came from better homes, and were better off for it. But, the type of programs Heckman supports haven't been able to replicate those positive effects in the long-term.

Murray understands the value of socio-emotional skills as much as the next guy - he just is familiar with studies that have shown improvement in neither cognitive ability NOR socio-emotional ability as a result of the kind of programs Heckman supports.

Heckman cites the Abecedarian and Perry projects as successes. There is no shortage of criticisms of these projects. Most notably, the control group in the Abecedarian project already had a mean IQ 5 points lower than the experimental group BEFORE the intervention program even started.

Now, check out the results of the Infant Health and Development Program:

There were some initial gains, but after five, eight, eighteen years there were virtually no differences between the control and experimental groups.

Anyway, it's easy to call a leading researcher in the field "naive" or uneducated when your knowledge of the matter consists in having read a few op-eds. If you're really interested go ahead and read the literature, not just someone's summary. And by "the literature" I mean the entire literature, not just a few initial studies whose results have not been replicated by later studies. Again, if you're interested, you can start by following the link I provided and investigate whether there was any lasting impact of the Infant Health and Development Program.

It seems that you have ignored Heckman's response to the IHDP criticism. The commenters on Murray's work have largely been criticizing him because it seems, if you believe Heckman's response to Murray in the Boston Review, that Murray misrepresented the IHDP findings implications for the Abecedarian. They don't claim that Murray believes culture matters. I think the bias referred to is Murray's bias against government programs to tackle inequality of opportunity.

Your point about the 5 point short fall in the Abecedarian is interesting. It will not have slipped Heckman's notice. I do not know how he would have gone about correcting for it: he won't have tried to cover it up.

Murray anticipated Heckman's response to his IHDP criticism by arguing that the claimed benefits for the heavier babies represent just "after-the-fact slicing and dicing of the data." The claim that the heavier babies benefitted was questioned by Bacharach and Baumeister in their critique of the program. They argued that the effect reflected biased sampling:

The one small remaining effect they attempted to salvage (as reported by Brooks-Gunn et al., 1994) was a 3.7 mean IQ difference (p = .03) favoring heavier infants in the treatment group. But when we examined their summary data deposited with the National Auxiliary Publication Service (NAPS), the mean difference between heavier birthweight babies was 3.0 points (p = .09).

We (Baumeister & Bacharach, 1997) maintained that the NAPS data were superior in that they represented the same children tested at both points in time. Simple arithmetic reveals that Brooks-Gunn et al. (1994) included children at 5 years who had not been tested when they were 3 years. Furthermore, at the 8-year follow-up (McCarton et al., 1997), where the reported difference (for the heavier infants) was still about 4 IQ points, still other numbers of children were tested. We concluded that there is a bias in these data simply too blatant to be ignored and that could easily account for any small difference, statistically significant or not.

More generally, Bacharach and Baumeister note that controlling for maternal IQ generally renders intervention results trivial or non-existent.

Thank you very much for pointing me towards the article.

I don't wish to repeat what Heckman has already said. However, I think it is pertinent to your argument to point out:

1) Heckman does not claim that the Abecedarian and Perry programs have long run effects on IQ. He claims that these programs have long run effects on non-cognitive skills and later positive outcomes in life. The comparable treatment group had lower scores on the youth risk behaviour scale.

2) Murray's slicing argument would be relevant if IHDP were a true replication of Abecedarian, it is not. It is the Abecedarian applied to poor health as opposed to disadvantaged from an SES perspective. The slice that Heckman chooses is a slice that selects a group with similar characteristics to those targeted by the Abecedarian, the SES disadvantaged.

3) Heckman also points out that the IHDP does not account for substitution bias.

This said, the IHDP results are still a little worrying. They should not incline one to ignore the results from the Abecedarian or the Perry Preschool programs as mere statistical anomalies. If it is the case that their success is due to the enthusiasm and skill of those involved then this has implications for how we attempt to scale them.

In general they give reason for optimism about the potential for designing successful interventions, as do many other programs worldwide. The methods and mechanisms have not been fully understood. The returns in the case of success are sufficiently high that it warrants much more research. Particularly well run replication of the Abecedarian and Perry Preschool programs. I find it strange that there are so many people opposed to the idea that we can develop a scientific understanding of child development that allows us to improve child outcomes, particularly among the poor.

"Most notably, the control group in the Abecedarian project already had a mean IQ 5 points lower than the experimental group BEFORE the intervention program even started."

Didn't the Abecedarian program start from birth (in fact with prenatal visits aas well). This would make it difficult to have an IQ test before the intervention program started. I would like to know your source on this. Or did you just hear it off a bloke in a pub.

The project started when the participants were a few months old. Cognitive tests were first administered when they were six months old, and the gap between the treatment and control groups at that age was of similar magnitude as the gap in adolescence. From the paper I linked to above:

At age 15, the 4.6 point WISC-R IQ difference in the Abecedarian Project was not statistically significant (Farran, in press). The mean ability test score of the intervention group was somewhat higher than the control group's at 6 months, shortly after they entered the Project. Although their score remained in the average range throughout, by 18 months, it was appreciably higher than the control group's only because the mean score of the control group had declined until it began, by 48 months, a steady recovery. In general, the experimental group never increased in IQ, but remained in the average range. Nor did the control group decline into mental retardation. The final IQ difference, not incidentally, was about the same as the difference at 6 months; a difference that Ramey, Yeates, and Short (1984) admit cannot be attributed to the intervention.

In regard to this conspicuous lack of enduring effect in the Abecedarian Project, Spitz (1999, p. 283) raised a question and then proceeded to answer it: "What happened during those first 1.6 months at the day care centre to produce an effect worth 6 points, whereas an additional 4 1/2 years of massive intervention ended with virtually no effect? It seems to me that it is not unreasonable to infer that nothing happened, but rather, some initial difference in the control and intervention groups had (by chance) escaped randomisation, and revealed itself at six months of age." We found similar problems with the IHDP.

So either the first 1.6 months of the program had a huge effect on IQ and the remaining 4.5 years had no effect at all, or there was a problem with randomization.

Mr Nah, I have studied Coming Apart. The description is very good. I see bits of it around me every day.
The point is bits: Murray, not unlike Heckman, does not understand how competition works [admittedly in the face of policies that limit it].
"It's not over 'til it's over", and it'll never be over.

The inside story on the James Heckman - Charles Murray relationship, as I've pieced it together over the years, appears to be this: Murray and Herrnstein got a fair amount of help from Heckman in checking over their draft of The Bell Curve. Thus, Murray was surprised at Heckman's angry review of The Bell Curve in Reason in 1995. Then, Heckman set out to prove The Bell Curve wrong about IQ. After a number of years, Heckman mostly gave up on IQ, privately admitting Murray had been largely right. Heckman switched to looking for ways to improve the character and citizenship of poor children. Murray was pleased by this change in emphasis on Heckman's part toward what Murray has been more or less saying all along (as is obvious from Coming Apart): society does have some influence on the morals of young people.

Do you think "Murray and Hernstein got a fair bit of help" should probably be "Murray and Hernstein got a fair bit of advice/critique"? If his advice/critique was not acted upon you would expect Heckman to criticize the flaws he had previously said, privately, that their work had.

The Reason article ( is not angry. It praises and recommends the book but then discusses how it falls short of what it aims to be. One of the faults, he finds with it, is the time they spend establishing that a single dimension of human skill determines outcomes (g). The multidimensional nature of human skills has been a large part of Heckman's work predating the bellcurve.

Heckman's subsequent work doesn't set out to prove the bellcurve wrong on IQ. Heckman's work set out to show that parenting and various programmes could be effective at addressing various social problems. In some sense this is a critique of Murray's earlier work more than The Bellcurve. But all in all I doubt that Heckman had Murray much in mind when pursuing any of these research agendas. They will have been pursued because of their intrinsic importance and interest.

I'm pretty sure that The Bell Curve debate was what made Heckman realize the importance and relative unmalleability of IQ. I ran Google Scholar queries for Heckman papers containing the word 'IQ' published before and after The Bell Curve (1994). Before 1994 there were just three hits for IQ, all of them false positives. From 1994 on, there are 122 hits. The results are similar for related search terms like 'intelligence' and 'cognitive abilities.'

Fair enough. That seems pretty compelling.

I would add the note that his work focused from the outset on determining the role of factors other than IQ in determining socio-economic outcomes. Which is consistent with his current research agenda.

"Do you think “Murray and Hernstein got a fair bit of help” should probably be “Murray and Hernstein got a fair bit of advice/critique”?"


At least, that's not what I heard. What I heard was that Herrnstein and Murray greatly appreciated Heckman's pre-publication help, and Murray (the survivor of the pair) was surprised by the Reason review.

I don't think the Reason review holds up well, so I'm glad that instead of just re-iterating himself on IQ, Heckman in this century has devoted much effort to investigating whether policy can improve conscientiousness and other valuable character traits. That seems more promising.

I think you describe the article incorrectly with your comment about Heckman reiterating himself on IQ. His criticisms of the book are representative of his current research.

The criticisms of the book that he makes are:

1) Arguing that a single dimension of heterogeneity, the g dimension, is the chief determinant of outcomes is unnecessary to the thesis and that there is much evidence suggesting it is untrue. Heckman claims that "motivation and attitude are as important--and possibly more important--for success than is raw IQ." .
2) That unless it can be established that skills are determined entirely by nature the correct approach to evaluating policy that targets skill development is to use cost-benefit analysis. He claims that they do not use cost benefit analysis. He also points out that they mischaracterise the returns to education as being entirely driven by g.
3) That there are a series of measurement problems in their evaluation of the relative effects of SES and g on socioeconomic outcomes.
4) Their visions of the future do not follow directly from their empirical claims, conditional on those claims being true.

Criticisms 1), 2) and 3) are completely in line with his current work, as well as his prior work. His empirical concern in the article is to reject that some entirely genetically determined g is all that matters in determining life outcomes. This is exactly what his research on development and early childhood intervention have been on in this century.

"His empirical concern in the article is to reject that some entirely genetically determined g is all that matters in determining life outcomes."

And Heckman's "Reason" review of The Bell Curve did a wonderful job of demolishing that straw man.

Wonderful, Gossip! The sociology of science is frequently more interesting than the methodology of science, to say nothing of the results.



The Abcedarian Project is 40 years old and the Perry Project is 50 years old. Considering the vast sums poured into educational reform and research by philanthropists and government, shouldn't they have been replicated by now?

On the other hand, there is some evidence that society can get kids to behave better or worse. Young people behaved worse in 1977 than in 1962, and they behave better in 2012 than in 1992.

Education reform will go on forever. Many resources have and will be spent on it and that is fine. That is sort of the point.

2nd paragraph:?

"2nd paragraph:?"

E.g., juvenile crime rates were lower in 1962 and 2012 than in 1972 or 1992. Something caused that.

Of course something caused that. You can say it is societal, or maybe it is primarily a few specific factors that drove that. We could set up a regression model with 1,000 factors and perhaps say that 900 of those are societal. Is everything that is not genetic considered societal?

sure, it will go on forever, but I'd expect to either have nailed down or disproven the results of studies that old in a field that important. There have been two entire generations since the first one they're arguing about. That's absurd, those kids were born around the same time as my mother.


Spot on.

Comments for this post are closed