Saletan states that 78% of the differences in picky eating can be attributed to differences in genes (the very definition of heritability). Fine. He then goes on to say that none of the similarities in picky eating can be attributed to being in the same household (the very definition of shared environment, the flip side of which is to say that none of the differences in picky eating are due to differences in households). Also fine, though counterintuitive like so much of the behavioral genetics data.

However, he then gets it wrong when he states "The remaining 22 percent influence is from raising genetically similar kids in different households, so parenting style might make a difference." Not so. The remaining 22% of variance falls to "nonshared environment" which is defined as "differences in environmental factors between individuals who share the SAME household". Once you factor out genetic and shared-environment influences from a trait, you're left with nonshared environment. An example of nonshared environment is this: my brother and I are being raised together, but when I make it to third grade I get a great new teacher who wasn't there when he was in third grade. Because of this, I learn to do math better than he did.

Nonshared environment is most easily calculated by quantifying the differences between identical twins raised in the same household...those differences can't be attributed to different genes, or different households, because both were the same for each twin. Therefore, they must be due to nonshared environment.

If "nonshared environment" was what Saletan thinks it is, it would be "shared environment". There is no room for "parenting style" (something which we assume has some consistancy within a household) in nonshared environment variance.

D'oh! 300 pages of Judith Harris wasted on me.

We're posting a correction with a link here. Thanks for the catch.

Very gracious of you. Thanks! Sorry for any lack of tact in my response...I tend to think of the Fray audience as hard-to-convince readers, not authors receptive to feedback. Behavioral genetics is kind of a pet topic for me, since I teach developmental psychology and make a very big deal of Judy Harris, the pervasive nurture assumption, and why the Blank Slate is such a terrible model of cognition. I always try to give my undergrads a good grasp of what heritability studies mean, even though heritability is an incredibly slippery concept.

Thanks for picking such interesting topics...I enjoy the column very much!

Ah, you teach developmental psych. I knew you had to be in the sciences; you've posted some pretty knowledgeable takedowns of my layman's errors before.

I came to this beat thinking I could spout off just like I had on the politics beat. Whoops! Then after my early mistakes, I tightened up and tried to be as careful and knowledgeable as scientists in their own fields. Now I'm realizing that I can't expect to meet that standard, so my current rules are 1) encourage experts to peer-review what I screwed up and 2) don't make the same mistake twice.

Any of you out there who have scientist friends, please invite 'em to patrol this board. We get a lot of hot topics in here, and we need the reality checks.