Bryan Caplan  

Has Harris Done It Again?

PRINT
The Verdict on V for Vendet... Slate Produces Two Good...

I'm a huge fan of Judith Harris' The Nurture Assumption, which powerfully debunks the idea that how your parents raised you has a large effect on how you turned out. Now she's got a new book, No Two Alike, which presents her theory of personality formation.

Harris deliberately sets the bar high: her explicit goal is to explain why even identical twins raised in the same home are STILL different from each other.

Has she succeeded? I'm afraid not. While the book is well worth reading and full of interesting insights, she hastily dismisses several promising approaches, and her own story is poorly fleshed out. Main problems:

1. As far as I can tell, Harris overstates how different identical twins are from each other. Even a single individual will test somewhat differently on personality tests at different times. How do the personality test differences between identical twins tested at the same time compare to the test differences between one individual tested at different times? (I think some data does correct for this problem, but it doesn't seem like Harris makes this distinction).

2. Harris is too quick to dismiss genetic-environmental interaction effects. Yes, the evidence in favor is crummy. But it's also very difficult to get. Contrary to Harris, the jury is still out.

3. Harris is too quick to dismiss socialization as a cause of identical twins' differences. Yes, if there is only one culture, culture can't explain differences. But if there are many cultures and sub-cultures, and twins don't always belong to exactly the same one, socialization could easily explain differences.

4. In the end, Harris zeroes in on the pursuit of status to explain personality differences. But how does this help to explain differences between identical twins? In the end, she has to appeal to random events that build on each other. Fine, but what makes her think that random events relevant to status are so special? Why not just say "random events"?

5. Harris barely tries to relate "the status system" to any of the main personality traits that psychologists study. For example, how does greater or lesser Openness or Neuroticism have anything to do with status? Maybe there's an answer, but it's not in this book.

6. Harris knows a lot of disciplines, but she should have added economics and game theory to the mix. If we think about the distribution of personalities as a mixed strategy Nash equilibrium, then we should expect that - at the margin - all personalities will be equally effective. From this perspective, variation is not very puzzling: Maybe we see a lot of variation because variation is practically free.

Is this a bad book? Hardly. I'd much rather read a book that asks a thoughtful question and fails to answer it than a book that asks a boring question and gets it right. But if you want to see Harris at her best, start with The Nurture Assumption.


Comments and Sharing





COMMENTS (8 to date)
Fabio writes:

Bryan says: "Harris barely tries to relate "the status system" to any of the main personality traits that psychologists study. For example, how does greater or lesser Openness or Neuroticism have anything to do with status? Maybe there's an answer, but it's not in this book."

Hi, Bryan. I haven't read either of Harris' books but from your description, but I might guess Harris' hypothesis can be stated as:

Different societies/peer groups have different systems for rewards, which lead to status and/or position. These provide incentives to suppress or cultivate different aspects of personality. Therefore, the "status system" explains some variance in personality.

This claim was postulated in 1950s sociology by functionalists like Parsons when he claimed that institutions, like schooling and work, encouraged the adoption of certain values. Robert Merton made similar claims in his seminal essay on individual behavior and status/reward systems. This sort of claim was never systematically investigated, but it's been around for a while.

To connect to psychology, you might simply say that the institutions and practices that reward individual behavior encourage the expression of certain traits. For example, American capitalism encourages risk taking, which is probably highly correlated with stuff studied by psychologists.

This might be an important genetic-environment interaction. I can believe that not everyone has a "risk taking" gene. I can also believe that social environment might somehow encourage the expression of the gene. Although I don't remember where I heard it, I do remember someone making the claim that exposure to violence during adolesence seems to promote agressiveness as an enduring personality trait.

So maybe the fruitful approach to culture and personality is to say: (a) much of your basic personality is due to genes & (b) during adolesence, your peer group and culture provides incentives for the cultivation of personality traits, long as they are consistent with your basic genetic make up.

It is tempting after reading Pinker & twin studies to say it's "all genetics." But as any decent geneticist will tell you, including Pinker himself, genes are only the beginning of the story. The environment can trigger some genes and not others, and the environment can suppress traits that are caused by genes.

Just think of the gardner - same seeds, but he can make one plant grow enormously while he stunts the other.

Steve Sailer writes:

Robert A. Heinlein's "Time for the Stars" is narrated by the subordinate member of a pair of identical twins. Why does he let his brother be the idea man? Well, it saves a lot of time on arguing, especially when what his twin wants to do is usually pretty much what he wants to do. (They are, after all, genetically similar.) But, there are still a lot of secret resentments generated by his not being the dominant twin.

Another reason why identical twins end up different is because there often is room for two people to fill identical roles. For example, Horace Grant had to learn to play power forward in high school because his brother Harvey Grant played small forward in high school. Horace didn't look like a natural power forward, but he ended up having an even better career than Harvey.

You see this a lot with very tall twins, where one plays center and the other plays power forward, or in high school football, where one twin plays quarterback and the other receiver.

Also, position and other differences in the womb can mean a lot in later life.

Scott writes:

Harris is too quick to dismiss socialization as a cause of identical twins' differences. Yes, if there is only one culture, culture can't explain differences. But if there are many cultures and sub-cultures, and twins don't always belong to exactly the same one, socialization could easily explain differences

Harris addressed this during her exposition by focusing on siamese twins. Since siamese twins are connected, they share not only genes but also environment - even down to the most minute detail. Therefore, the differences between Laleh and Laden could not be accounted for by differential socialization. And the differences were large enough for them to risk their lives on surgery that would make them separate.

Zac writes:

I think you're on to something with the mixed strategy nash equilibrium, but the rest of your theories are dead ends (particularly the one about socialization).

Ted Craig writes:

My biggest issue with Harris has always been that she bases her work on twin studies. Twins by their nature are abnormal and I often wonder how much this skews her work.
By abnormal I mean their experience and the way they are treated is different from other children. The mother has to instantly split her time between the time from birth, for example, which is different from the experience of the average person.
I'm not saying she's wrong. I just believe this adds to Bryan's assertion that there isn't enough good data, so the jury has to be out.

Jim Clay writes:

Her premise (as I understand it)- that you would expect identical twins in the same environment to turn out the same- seems foolish to me. There might be something to that if they were "independent variables", but they aren't. The twins I have known are often intentionally different. They are so tired of the sameness that they purposely veer off onto different paths.

Yes, I know a lot of it is superficial stuff like the clothes they wear, but I think the intentional differentiation goes deeper than that.

David Friedman writes:

I also am a fan of Harris--I had a post about her first book on my blog a month or two ago. Like Fabio I haven't read the new book, but I can offer a guess about the special importance of status.

Arguably, status differences are reinforcing. Once it is established which of the twins has superior status, that superior status then results in interactions that increase the status difference--a system of positive feeback. So small initial causes could explain substantial final results.

Alan Reifman writes:

I think Ted and Jim (above) are on to some important points. Not only are identical twins a very specialized population, but identical twins reared apart are an even more specialized subset. As critics of twin studies have long noted, whether "apart" actually represents "different" environments is questionable, at least in some cases (e.g., when one twin is sent to live with his or her mother's sister).

Also, the possibility that twins would respond to self-report questionnaires by constrasting themselves to their co-twin, was noted by a team of researchers about five years ago. These researchers instead had independent judges rate the behaviors of twins who were videotaped individually in a variety of situations.

Having thus removed the impact of self-report bias, the researchers found shared environment to account for 25% of the influence on behavior. This finding, to some degree at least, would appear to undercut Harris's premise of even identical twins from the same environment being so different in personality.

The study I've referred to is:

Borkenau, P., Riemann, R., Angleitner, A., & Spinath, F. M. (2001). Genetic and environmental influences on observed personality: Evidence from the German Observational Study of Adult Twins. Journal of Personality and Social Psychology, 80, 655-668.

Comments for this entry have been closed
Return to top