Fabio Rojas, sociology professor at Indiana University and my best friend from UC Berkeley, reviews my new book at Orgtheory. His more than fair intro:
Though this book is written by an economist, it's not another
cute-o-nomics pop text. It's a serious book about family planning
that's based on his reading of child development, psychology, genetics,
economics, and other fields. It's about one of life's most important
decisions, and this is what social scientsits should be thinking about.
The argument boils down to a simple point. If the evidence shows that you are over estimating the cost of having children, then, on the margin, you should probably have another child.
Fab then seems to criticize me for overstating my case:
In my view, twin studies tend to have two important limitations.
First, there is non-random selection of parents into adoption. Adopters
are, by definition, very unlike the rest of the population. Not in
income or demographics, but in personality...
The other limitation of twin and adoption studies is that they study
variation in existing parenting practices. It may be the case that
American parents simply don't know how to correctly socialize a kid to
reach some goal. Therefore, variations in family environment are just
variations in failed practices.
Here's a concrete example: child obesity. A hard core twin study
advocate would justifiably point to twin studies showing that weight or
BMI is more linked to shared parents than shared family environment... To be blunt, in a world where *everyone* eats bags of twinkies,
there won't be much of an effect of living in a home where people eat a
few more or less twinkies.
For that reason, it is too much of a jump to say that family
environment can't possibly affect weight. For example, parents who
remove all twinkies and switch to an all broccoli diet will likely
affect their children's weight.
What disappoints me about Fab's criticisms is that I specifically address and agree with virtually all of them in the text of the book. I repeatedly state that the best way to interpret the adoption evidence is, "If you're fit to adopt, you're good enough." And I repeatedly disavow the view that that "family environment can't affect X." All I ever claim is that family environment doesn't affect X (or has little effect on X). I devote entire sections to the latter point. So why does Fab write "Even if the argument is overstated" as if these sections didn't exist?
Admittedly, if First World parenting were extremely homogeneous, mere disclaimers might not be enough. But by almost any standard, parenting in the U.S. is diverse. We've got the whole range from fundamentalist to hippie, health nut to McDonald's "heavy users," ghetto poor to millionaire, TV-always-on to no-TV-at-all, Ozzie-and-Harriet to Rachel Berry's "two gay dads." Twin studies often undersample the poor, but not severely. Current adoption samples are fairly restrictive, but many of the best studies - like Sacerdote's work on Korean adoptees - look at kids adopted when the rules were much more lax. And Scandinavian twin and adoption studies are often based on national registries where sample bias isn't an issue.
At risk of pedantry, I'd really like Fab to re-read pages 41-42 and 83-86 and tell me how I failed to anticipate the key complaints in his review. How about it, old friend?