Politically Incorrect Paper of the Day: Death Penalty Eugenics

Anthropologist Peter Frost and anthropologist and population geneticist Henry Harpending argue that killing murderers pacified the population eugenically.

At the beginning of [1500]… the English homicide rate was about 20 to 40 per year per 100,000 people. At the end [1750, AT], it was about 2 to 4 per 100,000, i.e., a 10-fold reduction (Eisner, 2001).

…Can this leftward shift be explained by the high execution rate between 1500 and 1750? During that period, 0.5 to 1% of all men were removed from each generation through court-ordered executions and a comparable proportion through extrajudicial executions, i.e., deaths of offenders at the scene of the crime or in prison while awaiting trial. The total execution rate was thus somewhere between 1 and 2%. These men were permanently removed from the population, as was the heritable component of their propensity for homicide. If we assume a standard normal distribution in the male population, the most violent 1 to 2% should form a right-hand “tail” that begins 2.33–2.05 SD to the right of the mean propensity for homicide. If we eliminate this right-hand tail and leave only the other 98-99% to survive and reproduce, we have a selection differential of 0.027 to 0.049 SD per generation.

…The reader can see that this selection differential, which we derived from the execution rate, is at most a little over half the selection differential of 0.08 SD per generation that we derived from the historical decline in the homicide rate.

Thus, the authors argue that it is possible that a substantial decline in criminality can be explained by the eugenics of execution. The authors, assume, however, that executed criminals have no offspring which is unlikely, especially if criminals have higher fertility rates.

Hat tip to PseudoErasmus on twitter.


Has there been any research on the impact of high American incarceration rates on the fertility rates of the incarcerated men?

Tyler is right, the biggest weakness of the Frost-Harpending argument is that executed criminals had no offspring. Their proviso that "oversestimation due to the third assumption ... would have likely decreased as the shifting cultural norm began to marginalize violent men on the mate market" seems rather weaselish.

One can hypothesize that in a time of very high infant mortality, many of the children of absent fathers died before growing up. So the premise that the executed men didn't have many offspring is perhaps still a good one.

Well, it is true, per Clark, fertility rates go down with class & income.

In the 1500's??

From 1250 until the early 1800s there is evidence of a pronounced class gradient to fertility rates. But 0 offspring is a strong assumption and they might have calibrated the model by relaxing that assumption.

Sorry, I took you to be saying that fertility rates went down as class and income went up, when of course in the 1500's it was the opposite. I see now we are on the same page.

"Tyler is right..."
This post was written by Alex Tabarrok, not Tyler Cohen :-)

Oops, sorry !

"Cowen", sorry...

"The authors, assume, however, that executed criminals have no offspring which is unlikely,"

I agree that's unlikely, but it seems likely that executed criminals tended to have less off spring than males from the general population. If they were half as likely to have children as a standard group it would tend to drastically reduce their genetic progeny.

For execution to have a eugenics effect, one would need to show that *unexecuted* convicted murderers tended to have children *during and after (if released)* incarceration. The alternative to execution was imprisonment (I assume). Were conjugal visits a thing back then?

Imprisonment as a punishment was not much of "a thing" prior to the 19th century. Hence this argument that such sentences, rather than the death penalty, may violate the original meaning of the "cruel and unusual" prohibition:
"The modern practice of imprisonment is itself a profound innovation. In the eighteenth century, crimes were typically punished with fines, corporal punishment, public humiliation, banishment, or execution. Imprisonment was rarely used and sentences of more than a few years were almost never imposed. Moreover, the modern prison, in which prisoners are completely segregated from society and sometimes subjected to coercive “treatment” for their criminogenic characteristics, was not born until 1790 and did not achieve anything like its current form until the end of the nineteenth century".

The idea that putting violent men into the same small cage for the rest of their lives is compassionate is a comparatively recent innovation.

that is a weakness. another one is that their theory does not account for either 1) why the homicide rates appear to have dropped significantly in england *before* 1500, and 2) why the homicide rates remained much higher in italy even though italy had some very strong states in the medieval period (city-states) -- ones that that *did* execute murderers (although i don't know at what rates):


liberalarts March 9, 2015 at 7:35 am

Has there been any research on the impact of high American incarceration rates on the fertility rates of the incarcerated men?

Yes. Incarcerated men have more children.

But would they have even more kids if they were not incarcerated?

Cool. A little surprised to see this author give an obvious hat-tip to HBDers, although I realized this is by no means the first time it's happened.

It seems you are not familiar with the authors.

Their model implies a justice system that perfectly identifies the top 1-2%. I'd imagine the effect would be much more subtle if you made that a more reasonable assumption

Maybe they were more accurate. Everyone in the village would know the murder victim, know his/her enemies and who had motive.

Most property and almost all violent crimes could lead to death. You don't have to be that discriminitory. Imagine if the US executed almost every repeat offender, even if the system was wrong 50% of the time, you would eventually get most potential offenders in each generation.

I disagree with the eugenic argument though because I believe most human males are inherently violent and are only restrained by societal norms, and that only a small number express this violence in socially unacceptable ways in each generation, even if you eliminate that group every time it will take millenia to genetically remove the violent strain, and in that period the society would rapidly succumb to outsiders that retained that capacity.

Human violence is a protective trait that is also dangerous like an overactive immune system.

All felonies were punishable by death, but there were only six felonies

Maybe it's just you who is violent and constrained by social constraints Roy? Someone warn Roys relatives, he could be dangerous.

And what do you think is different about the "small number" that express themselves in "socially unacceptable ways"?

"I disagree with the eugenic argument though because I believe most human males are inherently violent and are only restrained by societal norms.."

I agree with this statement, but there are degrees of violence. I've been around young males (never a female) who tended to get into violent confrontations, were angry most of the time and tended to drink heavily. They were clearly a small minority of the male population.

JWatts. I agree, there are clearly a minority of troublemakers but how much of this is environment (even if it is an environmental trigger of a latent set of genes). For example it might be a combination of genes for aggression combined with genes for a lack of self control, combined with a failure of discipline because of externally induced family dynamics. While there might be a major genetic component in the sense that they are biologically predisposed to it through inherited genes, I think those genes are so widespread in the population and so useful for reproductivr sucess that it would require great difficulty in eliminating them. Genes for aggression are like sickle cell or fava bean allergy, except they deal with something much more profound than malaria.

I see where you're going but Occam's Razor would tend to imply that the simpler explanation is more likely to be right. And the simpler explanation is that a small subset of the population is more likely to be homicidal.


"...a small subset of the population is more likely to homocidal."

how exactly did I not say this?

“…a small subset of the population is more likely to homocidal.”
"how exactly did I not say this?"

"I think those genes are so widespread in the population and so useful for reproductivr sucess that it would require great difficulty in eliminating them."

Well from this you stated it's widespread genetically, but only expressed in a small subset, but I think the simpler explanation is that it's not widespread genetically.

The 2-R allele of the MAOA gene has been found in a number of studies to be associated with higher levels of violence. It is present in about 5% of US black males and about .1% of US white males.

RoyL: quite right & succinctly expressed.

> I believe most human males are inherently violent and are only restrained by societal norms, and that only a small number express this violence in socially unacceptable ways in each generation, even if you eliminate that group every time it will take millenia to genetically remove the violent strain...

About 1-4 percent of the Adult population in the United States are sociopaths, and sociopathy is highly heritable. 30% of those serving jail terms are sociopaths, and the rate is higher for violent offenders (interestingly, sociopaths are over-represented in positions of authority, such as among pastors and CEO's).

While some differences in criminal behavior are undoubtedly due in part to circumstances (the benefits of crime are greater and the costs smaller for the poor), it is also the case that different individuals facing the same incentives will commit crimes at different rates. Heritable personality traits and personality disorders are one important factor determining criminality.

"About 1-4 percent of the Adult population in the United States are sociopaths, and sociopathy is highly heritable. 30% of those serving jail terms are sociopaths, and the rate is higher for violent offenders (interestingly, sociopaths are over-represented in positions of authority, such as among pastors and CEO’s)."

My understanding is that sociopathy is not a binary condition. Up to a certain point, increased sociopathic tendencies are a benefit.

I'm not a sociologist, however. So that could be completely wrong.

Taco - Yes sociopathy may contribute to fitness. One feature about this trait is that sociopathic tendencies probably contribute the most to fitness in societies with a low frequency of sociopaths. This leads to hawk-dove or predator-prey dynamics. A society consisting solely of sociopaths wouuld be so dysfunctional that it would probably be destroyed by other groups. On the other hand the lower the frequency of sociopathy in a society the more advantageous it becomes to be a sociopath. The msot likely result is quasi-cyclical fluctuations in the level of sociopathy in a society.

If a small percentage of males aren't properly constrained by social norms, and you kill them, seems like the underlying trait that made them violate social norms should reduce in frequency pretty quickly.

Higher fertility rates in any given year or over a lifetime? A man who was executed at 20 yrs old who has fathered 4 children since he was 16 has a higher fertility rate than a 20 yr old man who only fathers one child every four years, but being able to live to the ripe old age of 40 or so would allow him to outspawn his more impulsive brother.

Are we talking about a 'gene for murder' that was reduced in the population by a long period of high executions or are we talking about conditioning the population. Imagine a town setting up a camera to automatically ticket anyone who goes through a red light. After a few years of thousands of tickets, I would expect the rate of running that light would go down yet no one would entertain a eugenic removal of a 'red light running' gene as the cause.

Fertility isn't quite the right measure either. Poor children were much less likely to survive, and the wife of an executed murderer was likely to become poor. (Wasn't confiscation of all property often part of the deal too?)

We are talking about a gene for murder, yes. Clearly the conditioning you mention would also occur. The conventional position seems to be that this would explain 100% of the decline; I doubt anyone thinks it's 0%.

If there's a heritable component to violent crime, and the justice system is even halfway accurate, then there has to be some effect of executions (or very long prison terms with no conjugal visits--if you don't get out of prison till you're 80, you probably don't leave any more progeny than if they hang you at 20). The interesting question is the size of the effect, right? If commiting serious enough crimes to get hanged is mostly your own innate tendencies, then it's probably rather strongly heritable. If that's mostly driven by circumstances, then it's probably more weakly heritable.

It does seem a bit backwards here. A gene for murder? Wouldn't a gene for killing other people be more likely? If there was such a gene it could find expression *either* in the form of criminal murder or in the form of executing people in socially approved circumstances. In the 1500's there were lots of socially acceptable ways to kill people. There were lynch mobs, the job of executioner, soldier's of fortune and so on. As albatross points out, there's also an assumption here the Justice system was accurate in determining who actually committed murder and executing those people only. It's my impression that modern forensic science has overturned a lot of erronous assumptions about investigating crimes.

It seems like you are postulating a common gene that expresses itself both in homicidal rage and in other forms. Particularly in the form of a propensity to become a lawyer, then a judge and then a judge who executes people at a higher rate than a judge without the gene, enough so that it has a statistically meaningful effect on the population.

This strikes me as somewhat unlikely.

I think the assumption here is not that there's a single gene or a small number that drive people to become murderers, but rather that there are some large number of genes that all contribute a little to a tendency to violence. The prevalence of those genes will go down over time if murderers tend to have fewer children than non-murderers.

How much selection you get from executing murderers depends on stuff like how important the genetic component is to determining whether you'll be a murderer--perhaps most people can be put into a situation where they'll commit a murder, but few find themselves in such a position--the genetic component would determine how exceptional the situation had to be to trigger a murder. Alfred will only commit a murder the day he finds his best friend and wife in bed together the day he's been fired and his dog has died; Bob will only commit a murder the day someone looks at him the wrong way on the street. Executing murderers gets rid of more Bobs than Alfreds.

A 'rage gene' has a problem in that it could express itself in a lot of ways other than murder. A soldier, gladiator, 'ultimate fighter', debator, and so on could use a 'fit of rage' to professional advantage without running afoul of local laws. So it isn't clear that a program to quickly arrest and execute murders before they could have lots of children would necessarily decrease the presence of a 'rage gene' in the population.

Likewise if the 'murder gene' is really just a rage gene, it wouldn't have any impact on premeditated, highly logical murders like killing someone for an insurance payout, Mafia hits etc.

Boonton, why do you assume that whether rage is channeled into a socially acceptable or unacceptable form is divorced from genetics and heritability?

Two possibilities here: A 'rage gene' that just makes one more prone to anger and rages and a 'murder gene' which makes one more prone to kill other people in socially unacceptable ways. Quick and certain execution of murderers would be an evolutionary filter against a murder gene but its hard to see how you could get a murder gene to begin with. I'm finding it hard to see a plausible evolutionary advantage such a gene would have during any point in our history.

A 'rage gene' on the other hand may have advantages in lots of times and places. But execution of murderers doesn't seem like a very certain evolutionary pressure against such a gene. As long as rage can be expressed in ways that don't run afoul of laws, such a gene would remain free to multiply.

There is also the problem of countries that implode (Afghanistan, Syria, Somalia, Rwanda, Bosnia etc.) We have a lot of history where an area was quite peaceful and civilized for long stretches of time but then with the collapse of the social order you get a fast orgy of violence, ethnic cleansing etc. The eugenics argument here is implying that we've been altering human nature. By punishing murderers for so many generations we actually alter human nature creating a population that is genetically less likely to murder. If that was the case then a collapse in social order should not produce much violence since the population has already been 'bred' to be nonviolent. Yet that does not seem to be the case.

I think more likely our 'genes' carry Pinkers set of 'angles' and 'demons' and we cultivate one more than the other depending upon our social context. That to me sounds like it has a much greater evolutionary advantage than trying to be the farm on a 'murder gene' or 'law abiding gene'.

albatross - The very large differences in the rates of violent crime between whites, blacks and East Asians points to a strong genetic component of violent behavior. In general virtually all human behavior has a strong (50% or more) genetic component.

Obviously both can play a part.

Alfred is too busy writing C&W hit tunes to commit murders.

Child abandonment was widespread for centuries. See John Boswell's book The Kindness of Strangers: The Abandonment of Children in Western Europe from Late Antiquity to the Renaissance. With dad dead the mom(s) had even lower chance keeping their offspring alive. Many had to give up trying.

We are talking multiple genes for murder.

I'd be fascinated to know how they contrived their crime statistics (or did they merely crib from some other professor whose work is a black box?).

The psychologist Paul Vitz offered some years ago that he'd been around long enough to witness excessive investment in notions of environmental influences and excessive investment in notions biological influences.

This is also not PC, I know, but I've always wondered if you had a similar phenomenon with the world wars in Europe and Japan. Germany and Japan in particular had a much higher amount of military deaths in WWII relative to civilian deaths. I have no idea if this hypothesis "works," but I can't be the first one to think of that, right?

Nope not the first one to think of it. Not to mention all of the other wars. That along with capital punishment and people prone to violence have really had a hard time of it.

Wouldn't the same effect show up after WW1? I'm pretty sure a single event would have to be pretty apocalyptic in a really selective way to have any noticable effect on a heritable trait (like some widespread disease killing almost everyone with a certain allele), and 20th century war with conscript troops probably doesn't qualify.

If I'm doing the calculation right, assuming a .5 narrow-sense heritability for IQ, if you killed everyone in the US with an IQ higher than 100, you'd only drop the average IQ about 6 points long term (less than half a standard deviation). (After killing everyone with an IQ of 100, the population average would drop about 12 points, but the next generation would regress about halfway to the mean, after which they'd be more or less stable.)

So what you are saying is that if we were to sterilize everyone below the IQ of 100 we'd permanently increase the IQ of the US by 6 points?

But, sterilization of that many people would never be politically acceptable. I wonder if there is some other means to reduce the progeny of the less intelligent portion of population that would be politically acceptable?

Stopping to subsidize them would be a good start but even that is a politically hard sell.

I know, i know, Mrs Planned Parenthood! Abortion?

It happened in Cambodia. Pol Pot exterminated the intelligentsia.

"Did Pol Pot Lower Cambodia’s National IQ?" http://humanvarieties.org/2014/06/12/hvgiq-cambodia/

http://h2oreuse.blogspot.com.au/2009/04/what-happened-in-cambodia.html "Anyone educated or "intellectual" was killed. The country's 60,000 Buddhist monks were killed. All the teachers were killed. Ethnic groups were attacked including the three largest minorities; the Vietnamese, Chinese, and Cham Muslims, along with twenty other smaller groups. Fifty percent of the estimated 425,000 Chinese living in Cambodia in 1975 perished."

BUt "James Flynn recently addressed this very question in his book Intelligence and Human Progress (2013, pp. 42-44) and answered ‘not much’"

But then the IQ of Cambodia college students 1995:
"The oldest study I’ve located for Cambodia is Fergusson et al (1995 ); Cattell’s Culture Fair Intelligence Test was administered to 23 students at the Institute of Economic Science in Phnom Penh and to 47 students at the Maharishi Vedic University in Prey Veng province. Remarkably, the IQ of these college students was 70.9."


I have no idea whether Pol Pot managed to lower the Cambodian average IQ, but I think that kind of selective apocalyptic mass death is the only way you're going to get a one-time event to have a big effect on something like IQ.

Even in a conscript army people who like combat and fighting will be more likely to chances and more dangerous tasks than the unwilling.

That seems plausible, but I don't know how big the effect is. An awful lot of guys who really didn't want to go to war found themselves with a rifle in their hands and people shooting at them in the two big world wars.

albatross - A drop of 6 points in the average IQ of a population can have very significant effects. In the US an IQ 0f 115 is about the minimum required for professionals such as engineers, doctors, lawyers etc. This is about 16% of a population with average IQ 100 but only about 8% of a population with an qaverage IQ of 94. The decline in the really high IQ "smart fraction" would be even greater.

On the other hand Linda Gottfredson has stated that in the current US economy a person with an IQ below 75 is essentially useless. That is about 5% of a population with an average IQ of 100 but about 10% of a population with an average IQ of 94. So a reduction in IQ has a "double whammy" diminishing the highly productive proportion of the population and increasing the economically useless proportion.

I doubt that a population with an average IQ of 94 would have much chance of being a really First World nation. The difference between the IQ of say Turkey and the countries of Northwest Europe is roughly six points or maybe a little more.

It's harder to get that sort of effect with a one-time thing (which it basically was for Japan at least). You'd need to get a REALLY large fraction of the populace dying. Something that happens over centuries will be more effective.

Weren't most of the armies in those massive wars (Napoleonic, US Civil War, First and Second World Wars) mostly conscripts? Also, aren't most war-related deaths among the civilian population? At best, war kills indiscriminately. At worst, it discriminates _in favor_ of those most capable and willing to kill.

Try Serbia in WWI

The Kingdom of Serbia lost more than 1,100,000 inhabitants during the war (both army and civilian losses), which represented over 27% of its overall population and 60% of its male population.[5][6] According to estimates by the Yugoslav government (1924) Serbia had lost 265,164 soldiers, or 25% of all mobilized people. By comparison, France lost 16.8%, Germany 15.4%, Russia 11.5%, and Italy 10.3%.

Why is that result ascribed to eugenetic?? how about "dangerous criminals aren't around to raise and (mis)educate children, so on average children are educated better , with less extremely negative impacts on their behaviour, and that leads to less homicides?"

The purported genetical improvement could be an educational (ie, enviromental) improvement.

They only investigate the genetic aspect, i.e. could the drop be explained by genetics. They make no attempt to show any causation or that there actually was such an effect. As such, educational effects would be additive in their model.

I don't understand your logic. How does an environmental component nullify the genetic component?

Looking at the time frame, I would make another argument.

From 1500 till the early 18th century the UK executed most of its troublemakers without effecting the crime rate. After the War of Spanish Succession, Queen Anne's War in the US, it started exporting some of its troublemakers to North America, while for the first time in a century it found its main population centers free of war and successful revolution. As the 18th century continued it continously increased its export of dangerous males AND females until the American Revolution as transport replaced execution. Australia quickly replaced North America and transport continued until the mid 19th century, accompanied by considerable executions.

Simultaneously with this process an apparatus of regular policing was created starting with John Fielding in the mid 1700s and culminating in the creation of the Metropolitan police force. Which allowed far more rigorous social control.

In short you have the end of internal political and religious violence, the export of disruptive elements, and the creation of the first modern police force.

You don't need eugenics, you just need the removal of disruptive elements while state power is consolidated and order instituted. But even if there is a eugenic factor, despite continued executions the rate of murder did not start falling until after a generation of transport.

Btw, if this is a eugenic argument why is Australia so peaceful?

As well as Americans of European decent.

Oz has the highest violent crime rate in the modern world.

Correlation? Australia's highest violent crime rate persists even after [gasp] the governemnt disarmed (guns confiscated) the civilian population.

Plus, the main anti-PC aspect of the study is that it doesn't advance the narrative. .

This is nonsense. The homicide rate in Australia is equal to or less than the first world average. As are the rates for manslaughter, serious assault, robbery etc.

"Btw, if this is a eugenic argument why is Australia so peaceful?"

Australia has the highest rate of reported rape in the OECD. As in over 3 times the US rate. And surprisingly almost 3 times the New Zealand rate. It's over 10 times the German rate.


Australia is not peaceful.

Well I was mistaken on that data point, I will physically assault my obnoxious moralizing Aussie colleague after work today!

As to the US, my idea is correct it might explain both the high US rate of violent crime, especially among anglo saxons, as opposed to say German and Scandinavian Americans, while simultaneously confirming the stereotypes that all colonials have about the English being effete. So, as a non anglo saxon American, I am game.

There are two very distinctive population groups in Australia, one descended from former British convicts and one not. The statistics do not appear disaggregated between them.

Assuming Australia's Anglo population is in line with their cohort in the rest of the Commonwealth, I suppose the hypothesis would be that disruptive individuals will tend to go extinct in an environment that puts a premium on social cooperation.

Those figures are misleading. Most likely because different measures for sexual assault are used in different jurisdictions. For example the report compares the total rate of all reported sexual assault, of any kind, in Australia to the number of convictions for rape in Germany. Using the same standard the rate per 100,00 of rape convictions is Australia 12.6 and Germany 10.7. Attempting to use broadly similiar standards the reported sexual assault rate per 100,000 in Australia is 28.6, in the USA 28.6, in New Zealand 29.6, in the UK 28.6. The best example of the difficulty of intercountry comparisons is that currently Sweden has the second highest reported rate, after South Africa, of any country at 66.5/100,000.

Limit to nesting of replies?

Reply to duxie: Cattell's "Culture Fair" Intelligence test says a sample of university students has an average IQ of 70.9. Really? And this doesn't send alarm bells ringing in your ears? If it's truly a "culture fair" intelligence test, these students are able to feed and clothe themselves, but not prepare their own meals. They probably can't handle class schedules that change by the day. (Yes, that's what an IQ of 70 means.) They certainly can't handle reading college level textbooks. No math or science. (They can probably categorize, the "stamp collecting" part, but "why" is beyond them.)

I don't know how much fun you and your classmates had making fun of the liberal arts, but in fact the liberal arts are the things a free person needs to know. Enough law, economics, sociology, psychology, literature, religion, and science to get by in the society they live in. (The key is "get by", not "be an expert". Maybe "stay free" would be a better description. Then everyone would want to learn the liberal arts.) This isn't possible with an IQ of 70, 70 is stupid. 70 is genetically determined to be taken advantage of by every unscrupulous salesman, tradesman, con artist, slimy evil practitioner of bait-and-switch capitalism at its lowest on the planet.

In short, it doesn't pass the smell test. Something is wrong. It isn't a random sample. Their definition of "university" isn't our definition of "university". There was a misprint, or a copying error, or a grading error, or something. Something just doesn't fit.

This isn't my field at all, but I am super skeptical of IQ comparisons across countries as different as (say) the US and Cambodia. There are so many differences in schooling, upbringing, day to day situation, experiences, etc., that it's easy to imagine those swamping the signal even in a test designed to be culturally neutral.

You don't "need" eugenics of course. This paper asks if eugenics COULD be responsible.

Ah, yes. The logic of "what if".

Violent criminals were hanged until dead.

The kinds of criminals being exported to North America/Australia tended to be pick pockets, fraudsters, debtors, etc.

quite so. not that many murderers were actually transported to australia or north america: http://hbdchick.wordpress.com/2014/04/28/not-totally-slacking-off/#comment-62190

How can the authors discount all the other possible causes for such a decline over this 250 year period, which include, in no particular order, the Protestant Reformation, the establishment of the American colonies, the East India Company, and the Atlantic slave trade, a bushel of wars, revolution, counter-revolution, and 250 years of economic development? Seems unlikely. Seriously, would anyone think it makes sense to compare the homicide rates from 1760s NYC with those of today, much less explain any differences between the two by referencing a single factor?

This is exactly how natural selection works. It's so bizarre how most people turn into creationists the second you talk about humans being subject to evolutionary pressures. Being a murderer, whatever the root cause, leads to a loss in fitness. You can't procreate if you're dead or in jail. People who are inclined to not murder others will have a benefit to their fitness.

Even if murderers have offspring before they are dead or incarcerated, they have less chance of having offspring or as many offspring as someone who doesn't murder.

So far, I haven't seen anyone denying evolution can happen to humans or can happen above the neck--I've just seen people here propose other mechanisms they think are more plausible explanations for the drop in violence, and pointing out reasons why the natural selection explanation seems questionable to them.

The thing is that evolution doesn't work so quickly with a constant population flow and a very large population. Combine a selection that is worling erratically and that geneologists suggest that the lowest level of societies have a tendency to contribute less than their population share to the gene pool in early modern societies, and it makes me very suspicious that this was all pulled off in 15-20 generations in such a large population.

And that 15-20 generations is closer to 350 years than 200

Don't rates of sickle cell anemia in a population change pretty quickly when removed from malarial areas?

'Being a murderer, whatever the root cause, leads to a loss in fitness.'

Let me introduce the German term Schreibtischtäter - in other words, the modern murderers on an industrial scale aren't the people unloading the cattle cars, it is the people that designed the entire process. And let us be honest, at least when referring to last major experiment in using execution as an intentional tool for eugenic goals - most of the people involved got off considerably more lightly than their victims, speaking in the eugenic terms that the perpetrators believed in and practiced. That's right, they had children - their victims didn't. In other words, one could say that mass murder is a eugenically viable strategy, at least when looking at a fairly recent attempt go beyond a mere Gedankenexperiment.

Of course, a certain GMU econ professor has also written 'ultimately the Nazi connection will be seen as a bump in the road' in regards to eugenics ( http://marginalrevolution.com/marginalrevolution/2013/10/further-small-steps-toward-designer-babies.html ). One can wonder if his experience living in Freiburg is too far in the past to remember such words as 'Schreibtischtäter.' After all, those eugenic paragons who had the triumphal will to carry out their beliefs also had programs designed to eliminate the criminally insane from their midst.

'In early October all hospitals, nursing homes, old-age homes, sanatoria were required to report all patients who had been institutionalised for five years or more, who had been committed as "criminally insane", who were of "non-Aryan race", or who had been diagnosed with any of a list of specified conditions. These included schizophrenia, epilepsy, Huntington's chorea, advanced syphilis, senile dementia, paralysis, encephalitis and "terminal neurological conditions generally". Many doctors and administrators assumed that the purpose of the reports was to identify inmates who were capable of being drafted for "labour service". They tended to overstate the degree of incapacity of their patients, to protect them from labour conscription – with fatal consequences.[60] When some institutions refused to co-operate, teams of T4 doctors (or in some cases Nazi medical students) visited them and compiled their own lists, sometimes in a very haphazard and ideologically motivated way.[61] At the same time, during 1940 all Jewish patients were removed from institutions and killed.' http://en.wikipedia.org/wiki/Action_T4#Killing_of_adults

That's both creepier & more obsessive than normal, p_a.

The creepy and obsessive p_a is on the right track for a change. Only in the wrong hemisphere. Has anyone studied the effects of premature state-imposed death on the original Americans?

It's a good paper. We should have more like it published in economics. I say that because of its deficiencies, not its virtues. As commenters have pointed out, there are alternative explanations (bad parenting, deterrence) and there are robustness checks (importance of the no-kids assumption) that could be done. An econ journal would make the author spend two years handling all that. What's better is to get the paper out, well-written so people can understand it, and then have other people publish follow-ups if it's a good idea. That's efficient because most econ papers do *not* have a good idea, and so are not worth the author and referees spending so much time on. First publish a pilot paper, and then if anybody reads it, publish a more solid version. PIketty's book is actually another example of this--- a book with a huge number of flaws, but that has stimulated a lot of other people's work.

Capitalism and the paradox of the British murder

This also assumes there were minimal false positives. That's pretty unlikely when you consider that 1) most prosecutions were by private citizens not state officials (so that large share of accusations were themselves acts of personal violence), 2)the jurors were usually powerful & wealthy (jury by 'peers,' as in nobles with peerages) who decided by class first, facts second, and 3) most state prosecutions were against political enemies, not violent criminals.

"3) most state prosecutions were against political enemies, not violent criminals."


Prosecutions for murder would have been public prosecutions: 'Rex' or 'Regina' v the accused. "Peers" in this context means "equals" not "nobles". Most juries would have been composed of tradesmen.

Oh goodness! This paper is bad history and bad science. History: There is a very large body of research on English/British and European criminal law during the relevant period, with lots focused on homicide. Google "historical murder rates in medieval and early modern europe" to find as much as you could possibly want to read on the latter, including extensive discussions of methodological issues. The authors show no familiarity with this literature. Their second assumption, that murderers on average kill only once (since a murderer cannot kill zero times, that means only a de mimimus number of murderers can kill 2 or more times): where is the research that shows this, particularly for medieval and early modern periods? The third assumption of no offspring: oy vey! Where is their research on, at least, age ranges of executed malefactors? More bad history: where is their research on the offenses for which people were executed? It was not only for murder and there is plenty of research on this issue. Does infanticide count? Even more: do the authors seriously contend that all those poor people dying in medieval English jails were being held for murder?! Where is their research on how many murders were actually prosecuted? How many murder prosecutions resulted in executions? Without data on these latter two questions, the authors cannot even pose the question they purport to answer. I could go on and on and on.

As for the science: heritability of a propensity to murder? Really? Where is the evidence for that? Aggressive/antisocial behavior does not equal murder. I have no idea whether a propensity to murder is heritable. I do not know whether or how strongly aggressive/antisocial behavior correlates with murder. But neither, apparently, do the authors. They estimate a factor of "0.69 for the heritability of aggressive behavior." This figure is meaningless.

There are multiple cultural explanations for the dramatic decline in murder rates in the West over the last 1000 years that have been postulated. It is in the nature of such explanations that they are not scientifically verifiable (of course, any such explanation not consistent with the data is very unlikely to be correct). Occam's razor.

The authors may be fine scholars in their fields and kind to their families and strangers. But their paper is junk.


This is no more than an exercise in arithmetic.

One of the authors is a professor of Anthropology, the other seems to be a notable archaeologist and the paper has a pretty extensive reference list.

I know nothing about the field, but I'm curious as to what expertise allows you to classify the paper is junk?

I don't want to be mean. It is just that the authors do not appear to have any knowledge of the very extensive scholarship on violent crime and executions, none of which appears in their reference list. The last 20-30 years has seen an explosion of research on these topics. The data accumulated in this research has to be incorporated in any claims about execution rates. In addition, the methodological problems with the article cannot be cured. Apart from the fact that the authors appear not to be familiar with the historical research, they posit a factor of "0.69 for the heritability of aggressive behavior." This is a very high number upon which their posited genetic explanation depends. Even if correct, the authors cannot, repeat cannot, use that factor as the heritability of the propensity to murder. Aggressive behavior and murder are not the same and any claim that the correlation between aggressive behavior and propensity to murder is 1.0 requires data. I would note that the authors do not even acknowledge that they make this substitution.

BTW, the proceedings of the Old Bailey, London's criminal court, have been digitized and are online. Check out the Ordinary's Accounts (the Ordinary was the preacher at Newgate Prison) on oldbaileyonline.org. These are accounts of the executions ordered at Old Bailey from 1676-1772 with details about the crimes and the criminals. It is easy to see that a large majority of executions were for property crimes, not to mention that many of the condemned malefactors reproduced before their demise. These accounts are a fun read (ignoring the real world consequences to the individuals involved), as are the individual trial reports. The site also has lots of interesting social history.

The number of multiple murders in a single incident is a quite modest share of the total. Also, the 1 murder floor per murderer is probably wrong - often a whole gang would be executed for a single murder if all participated under the felony-murder rule, and many (most?) executions in this time period were of non-murderers. Certainly there would be some multiple murderers, but this assumption is unlikely to be so far off from reality that it blows the rest of the study's usefulness.

The study is useful for providing an order of magnitude, back of napkin, estimate of how much of an impact eugenic effects could have (a material effect, but not a full explanation), but is hardly definitive or precise either.

"useful for providing an order of magnitude, back of napkin, estimate of how much of an impact eugenic effects could have (a material effect, but not a full explanation) "

The way I read it, that's how it's intended.

The authors' single murder assumption was for the lifetime of the bad guy assuming he or she had not been executed, not for a single event (i.e. multiple victims at one time). Your reference to the felony-murder rule raises another problem. If 3 highwaymen hold up a toff and one shoots the poor guy, we really cannot say that the other two are murderers for the purposes of the authors' analysis (which requires a genetic propensity to kill). I do not know how many people were executed solely on the basis of the felony-murder rule, but that would have to be taken into account in coming up with the numbers for the authors' formula.

I cannot agree that the article provides a back-of-the-envelope estimate of anything at all for the reasons I gave previously.

I more intrigued not by the death penalty but the murder rate decreased because being caught and convicted for murder increased during the same period as well. According to the author most death penalties were administered to thieves and other crimes while murder were often not solved. (Or likely as well the murdered body was not discovered or not proven to jury as there might not be witnesses.)

Otherwise, the authors did little to control for the Reformation and various Civil Wars during the period. Maybe, the murder dropped a lot because England was simply at more stable in the 1700s.

I remember Greg Clark discussing something similar, although he also mentioned that demonstrating literacy was enough to get one excused for crimes in the medieval era (since only the clergy were expected to be able to read and faced separate courts).

Tabarrok seems a bit confused about the concept of "eugenics".

Peter Frost discusses the selection effects of certain social policies. He describes the selection effects as "genetic pacification". Whether one regardes these effects as "eugenic" will depend on one's values. Elswehere, Frost argues that these same selection effects of genetic pacification in the Roman Empire resulted in its inability to defend itself and in its collapse. So some might regard these effects as "dysgenic", rather than "eugenic".

"So some might regard these effects as “dysgenic”, rather than “eugenic”."

I think that's a pretty good point.

Further considerations: the authors don't seem to have categorized the executions. Isn't it likely that most of these executions were political? Think Thomas Moore and Cramner. Not exactly homicidal maniacs.

I dunno, Thomas More like burning protestants at the stake.

1. Underestimatation of fertility as noted in the original post. Execution would almost always reduce lifetime number of children per man executed, but would rarely reduce that number by 100%.
2. However, there would also be an indirect fertility effect as children with an executed father were probably significantly less likely to live to adult reproduction than men whose fathers were not executed in that time period.
3. Executions from 1500-1750 were for a much broader array of capital crimes than homicide. Essentially any serious felony was punishable by a death sentence. Some were violent, others were not. On one hand, this means that many murderers who were not caught for murder were executed for other crimes. On the other, many non-violet thieves, for example, were executed.
4. The death rate for violent young men in this cohort may be greatly underestimated. Today, there is a strong demographic overlap between men who commit murder and men who are murder victims, especially in places where murder rates are high. A lot of homicide is retaliatory and a lot of homicide deaths could be de facto extrajudicial executions that may have more eugenic effect than those under color of law.
5. The counterexample of Australia - a large portion of people exiled to Australia were guilty of capital crimes but had their sentences commuted to exile. Yet, the violent crime rates in Australia, even early on, do not reflect that level of violence.
6. Regular, swift and certain executions for serious crimes may have had a deterrent effect that impacted the overall British culture in a way that had long lasting effect even after the deterrent effect was no longer present.
7. We know to a high degree of certainty that the genetic connection to violence is not that simple and that there are powerful non-genetic factors (e.g. traumatic brain injury and lead exposure in childhood) that have a profound effect on violent crime commission rates.

I think this problem/question was better addressed by Pinker in The Better Angels of Our Nature: Why Violence Has Declined . One of the useful things he did there was try to look at the rate of violent death in general and see how it has changed over time. Doing this eliminates trying to making distinctions between deaths from wars, 'just wars', executions, lynch mobs, crime and so on.

Looking at this graph:

We see a generalized decline in violent deaths over time. The 'murder gene' argument doesn't work very well. A hard nosed justice system that promptly removes murderers from the gene pool does logically seem like a good evolutionary filter that would diminish a 'murder gene', if one existed. But such a filter wouldn't diminish other types of killings like wars, executions, lynch mobs etc. In these killings someone with a 'murder gene' could still prosper, they would only murder in ways that are socially accepted.

In other words, this genetic aspect could take two forms:

Bender's "Kill All Humans" gene - This would be a gene that makes someone more inclined to kill other people whether or not society classifies that as murder to be punished or an act that is praiseworthy or blameless.

Break the Rules gene - This would be a gene that would make someone inclined to kill only when it would serve to violate society's norms and rules.

If society embarked on a campaign to eliminate murderers from the gene pool, I suspect the second type of gene would become less rare. But why would such a gene exist in the first place? And the only decline we'd see would be in death by murder. Presumably death by authorized killings, like in wars, mobs, and other collective activities, would see no decline. Yet all those deaths have declined as well.

The Bender gene, in contrast, could exist as an evolutionary advantage *but* again why wouldn't the gene respond by seeking 'approved' types of murder?

Pinker's theory is to argue for a balance of forces. Humans have instincts both towards and away from violence. A society will get more or less violence depending upon the way it cultivates one set of instincts over another.

This makes a lot more sense to me in terms of evolution. A real genetic advantage wouldn't be having a 'murder gene' or a 'peaceful gene' but instead a gene that favored looking at your environment and being inclined towards whichever strategy works best for the context you are in.

It also explains a lot of examples we see like Nazi Germany, Rwanda, Bosnia, Serbia, and ISIS dominated Syria/Iraq. History is filled with examples of countries that had relative domestic pease for long periods of time but when a major upheaval happens, like a strong central gov't suddenly becoming weak, things like ethnic cleansing can explode literally out of nowhere. The 'murder gene' theory says this shouldn't happen since the 'bad genes' have mostly been culled from the herd, it would take generations for a breakdown in government to unleash mass violence. But Pinker's theory predicts it. All that needs to happen is for a culture to suddenly switch from promoting 'angels' to 'inner demons' and you will get a huge amount of violence.

Consider the evolution of the characters on The Walking Dead. After only a few years in the zombie world and most non-dead Americans are characters with huge moral compromises on them. All of them were more or less well adjusted good Americans a few years ago. What happened was their context turned upside down where 'inner demons' more often than not had the right answer and the 'better angels' were wrong.

"We see a generalized decline in violent deaths over time. The ‘murder gene’ argument doesn’t work very well."

my guess is that what was selected out of the north european population over the course of the medieval period were genes for *impulsive* violence -- losing one's temper over a slight or an insult to one's "honor," like being bumped into or hearing someone say that your mother is a such-and-such.

Yes, but what about the people with malfunctioning/mutations on "look at the society and see what behavior works best there" genes? Such mutations would be heritable. Some of the behaviors caused by such malfunction might be beneficial in terms of reproductive fitness, at least within the context of a society that won't prevent them from reproducing.

Perhaps the contextual upheaval that would ultimately upend that peaceful society will be that the proportion of people with the malfunctioning "look at society..." genes becomes large enough that the old behaviors no longer work for people with standard "look at society..." genes. Then the solution becomes violence once again.

Both culture matters and genetics matter. There is no doubt. But the extent to which people's behavior is influenced and constrained by culture will be heritable. If there is no huge downside to being unconstrained by culture (execution, expulsion, what have you), then it may be advantageous to have a combination of genes that makes culture a far less significant contributor to your behavior.

Yes, but what about the people with malfunctioning/mutations on “look at the society and see what behavior works best there” genes? Such mutations would be heritable. Some of the behaviors caused by such malfunction might be beneficial in terms of reproductive fitness, at least within the context of a society that won’t prevent them from reproducing.

If you had such a malfunctioning gene you could go either way. You could try to rob, steal and kill in a society that is peaceful, law abiding and rewards playing by the rules more than breaking them. Or you could end up trying to play by the rules in a society with no rules. It seems to me that it would be pretty hard to make a case that misreading your society is an evolutionary advantage.

And remember society too is not static. Consider a society with no rules, ruled by criminals and strongmen. What would Tony Soprano say about being honest, keeping your word, playing fair etc.? No doubt he would say these are things you should do. After all, if everyone else is honest, it makes his lies more effective. So society has an interest in getting people to misread their environment, getting them to think playing by the rules is rewarded and cheating punished more often than it really is. I don't think society ever has a motive to fool people into thinking the opposite. So if there was a malfunctioning gene to read society, it would probably result in someone being more rule abiding rather than less IMO.

Both culture matters and genetics matter. There is no doubt. But the extent to which people’s behavior is influenced and constrained by culture will be heritable. If there is no huge downside to being unconstrained by culture (execution, expulsion, what have you), then it may be advantageous to have a combination of genes that makes culture a far less significant contributor to your behavior.

Maybe, but then in 1500 how exactly did England get a population with a lot of "immune to culture" genes that were then filtered out by generations of executions? Before 1500 was not caring about the culture you were living in an evolutionary advantage? That doesn't seem to make any sense. Successful people have always been successful by leveraging the culture they are in, learning how their society works and how to find the best possible place in it.

H.L. Mencken on this topic - "Jack Ketch as Eugenist" first printed in the American Mercury, July, 1925 -

"Has any historian ever noticed the salubrious effect, on the
English character, of the frenzy for hanging that went on in England
during the Eighteenth Century? When I say salubrious, of
course, I mean in the purely social sense. At the end of the Seventeenth
Century the Englishman was still one of the most turbulent
and lawless of civilized men; at the beginning of the
Nineteenth he was the most law-abiding. What worked the change
in him? I believe that it was worked by the rope of Jack Ketch.
During the Eighteenth Century the lawless strain was simply
chokecl out of the race. Perhaps a third of those in whose veins it
ran were actually hanged; the rest were chased out of the British
Isles, never to return. Some fled to Ireland, and revivified the decaying
Irish race; in practically all the Irish rebels of the past century
there have been plain traces of English blood. Others went to
the Dominions. Yet others came to the United States, and after
helping to conquer the Western wilderness, begat the yeggmen,
Prohibition agents, footpads, highjackers and other assassins of

Mencken is fun but he is not much of an historian. The massive criminalization of property crimes in the 18th century in England did not lead to massive increases in the numbers of executions. Besides, all this stuff about "murder genes," as if there is any evidence at all that they exist: how exactly do they factor into the massive decline in murder rates in the U.S. in the last 20 years? I'm starting to lean to leaded gas as the explanation.

Pure nonsense. Too many confounding factors combined with a very minor selection forcing function without dramatic heritability (hard to separate the cultural from the genetic).

With some males competing by methods other than violence (being smarter, richer, nicer, etc.) and winning even more, the cultural spread of these traits would be much faster with greater penetrance of nice guy genes.

If someone wanted to make a genetic claim, with several generations of the smart and educated marrying the smart and educated, we should be getting more of an intelligence (very heritable) stratification in our society.

"If someone wanted to make a genetic claim, with several generations of the smart and educated marrying the smart and educated, we should be getting more of an intelligence (very heritable) stratification in our society."

Someone did make that claim:


"They also argue that those with high intelligence, the "cognitive elite", are becoming separated from those of average and below-average intelligence."

You are right, I now remember. In todays PC world, such a book would be even less acceptable to the political leadership, but the problem is getting worse by the generation.

The irony is that a big fraction of the problem is social/cultural but we can't even correct or say that. You can't point out that some sub cultures are non-functional in the modern world and some can start in absolute poverty and rise to the top in one generation like the Vietnamese immigrants in Orange County Ca, but with their kids is wasn't a question of whether they would go to high school but which major university would they go to for their advanced degree. Meanwhile the other cultures remain in poverty and ignorance.

We can't even point out that the black culture in the better areas of LA is the same as the White and Asian cultures at similar levels with similar discussions and interests, but we have a huge black culture that is every bit as non-functional as the white trash cultures.

but we have a huge black culture that is every bit as non-functional as the white trash cultures.

Putting aside the dubious phrase "white trash," violent crime in nearly every poor black community is considerably worse than in the poorest, most "non-functional" white neighborhoods.

The trouble with crime statistics is that in a world of rampant overcriminalization, they're only comparable if you control for people committing crimes that wouldn't have been crimes in earlier generations. And of course in a world where the state is so willing to be used as a method of robbery, you also need to try to include cops and politicians who took the job in order to start what amounts to a successful career of crime.

There are multiple cultural explanations for the dramatic decline in murder rates in the West over the last 1000 years that have been postulated. It is in the nature of such explanations that they are not scientifically verifiable (of course, any such explanation not consistent with the data is very unlikely to be correct). Occam’s razor.

Indeed. I think the Frost/Harpending highlighting of executions is ONE factor in that mix of explanations, as I note on their blog. The controversial part is not so much death penalty applications (which were much more frequent back in medieval times- for a much wider range of offenses). There is some data that capital punishment over a long period of time, and as the state grew strong enough to monopolize violence (versus the private vendetta) did have some deterrent effect, and some impact in reducing homicide rates, including some cultural changes spillovers to behavior as the deterrent too effect. It is not the only factor but one of the significant ones.See Eisner 2001, Modernization, Self control and Lethal Violence).

But that is not the main controversy. What is, is the notion that such executions reduced "the criminal classes" and thus their impact on the European gene pool- producing a 's "pacified" or more peaceful, kinder, gentler, more virtuous European. In essence, executions helped "cull" the bad genes for violence among Europeans. But there are several weaknesses with this notion- as others show above. If "culling" did indeed take place, then it might be well said that Europe's numerous wars also exercised a "culling" effect- in "selecting against" the "bad genes" for violence entering the European gene pool.

Pat of this argument relates to the nature of European military personnel over the long period measured- Medieval to early modern times. Back in those days sometimes, the military was at least SOMETIMES used as a dumping ground for "undesirables" or misfits, into the lower ranks. Vagabonds, beggars etc were often impressed into the military, in both the naval and land forces, though again, as time went on military forces throughout Europe became better paid and more professionalized. The Ottomans, Germans and the British back in Medieval times and later all to some extent, used forced draft military recruitment to flush "undesirables" from the ranks of the better folk. Even in Russia one study notes:
[i]"In Russia, communities and landlords used conscription to send off criminals, troublemakers, drunkards and men deemed disobedient, unruly or simply lazy. It is hardly surprising that armies time and again complained about the quality of the personnel that was provided to them in this way." [/i]

So to some extent European militaries, at some level drew off the "dregs" of society, variously defined in their own eras. Casulaty rates in war would no doubt impact these at a higher rate, cutting off their participation in the gene pool.

But there is also another side to this reasoning. Sure more centralized European states may have suppressed private vendettas, and monopolized violence, making executions more common and regulated across a wider area, but did this "pacification" make Europeans less violent? Internal homicide rates in a territory may have went down, but violence over a wider zone went up. The newly centralizing hegemons crushed "wild" unregulated violence, and harnessed the resources gained into aggression against other weaker groups and neighbors- hence the numerous succession, dynastic, predatory and religious wars in Europe since medieval times.

If the "mo betta pacified Europeans" approach is correct, then European elites may have used both methods (cull by execution, or cull by warfare) to ensure that only the better people, those more malleable, and submissive would, in the long haul, win the gene inheritance sweepstakes. As regards the warfare track, it could be said that the European elites gained a two-fer bonus. If dirty work needed to be done, then the lower end people could be deployed to good effect, whether in Europe or in various colonial enterprises. If a more submissive, higher quality remaining population needed to be deployed they could do just as well. Hence as Christopher Browning’s study “Ordinary Men” shows, ordinary German accountants, bakers, store clerks etc, good submissive bourgeois types, in the "second string", unglamorous SS police battalions, were among the most zealous, vicious killers during the Holocaust. Internal pacification made possible more efficient external violence directed against neighboring peoples. Those at the top got the dirty work done, no matter who was used.

Jim says:
The 2-R allele of the MAOA gene has been found in a number of studies to be associated with higher levels of violence. It is present in about 5% of US black males and about .1% of US white males.

Re the alleged "black violence gene" you mention, the touted certainties are less than earth-shattering. Other studies show more links with environmental conditions such as childhood maltreatment. rather than embedded "genetic causes" that supposedly "promote" violence. QUOTE:

"The interaction with childhood maltreatment also proved robust to sensitivity analyses and generalized across studies of either cross-sectional or longitudinal design and studies in which maltreatment exposure was assessed by family (self, parent) report only or included independent informant sources. It is noteworthy, too, that MAOA variation interacted with childhood maltreatment to predict outcomes referenced to both childhood/ adolescence and adulthood, dependent measures of both con-tinuous and categorical distribution, and both violent and non- violent antisocial behaviors. The latter finding suggests that the low-activity MAOA genotype heightens maltreatment-dependent risk for a range of conduct problems and not aggression or criminal violence specifically."
--Byrd and Manuck 2014. MAOA, childhood maltreatment and antisocial behavior

Likewise Haberstick et al 2005, and another study using a clinical sample of males referred for substance abuse or conduct disorder problems [Young et al., 2006] [show] a direct effect of maltreatment on subsequent antisocial behaviour, but not MAOA genotype, was reported.

In addition, results from analyses using a measure of violent victimization during adolescence in place of childhood maltreatment [Haberstick et al., 2005] also did not replicate the interaction hypothesis of Caspi and colleagues (2002). (Taylor 2013).

Comments for this post are closed