Deconstructing nature-vs-nurture charts

By Felix Salmon
September 2, 2009
Greg Mankiw sparked a blogospheric resurgence of the nature vs nurture debate, Brad DeLong and Tyler Cowen weighed in with very different views of the empirical data. Tyler featured an extremely provocative graph, and I waited for someone, maybe Yglesias, to respond.


" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

After Greg Mankiw sparked a blogospheric resurgence of the nature vs nurture debate, Brad DeLong and Tyler Cowen Alex Tabarrok weighed in with very different views of the empirical data. Tyler Alex featured an extremely provocative graph, and I waited for someone, maybe Yglesias, to respond.

tyler.png

In the end, the expected take-down came from finance whiz Mike Konczal, solidifying even further his status as the Italian Vogue of the econoblogosphere: the best that there is, read by everybody who matters, if nobody else. Mike drilled down deep in to the dataset used to generate this chart, and found:

  • The adoptees are four years younger than the non-adoptees: 28 years old, on average, compared to 32 years old. These are years in which most people’s income rises substantially.
  • 70% of the adoptees are female, compared to only 39% of the non-adoptees. Females earn less than males, and male heirs might well be better at inheriting their father’s income than female heirs.
  • The family income of the non-adoptees was $61,000 per year; the family income of the adoptees was only $42,000 per year.
  • The income of both adoptees and non-adoptees was measured by asking their mother how much she thought they were earning. Which obviously affects reliability.

On top of all that, there are racial issues which may or may not be relevant: the adoptees, in this group, were generally of a different race than their adopted parents. Put it all together, and the most we can learn from Tyler’s chart is, as Mike puts it, that there are “a lot of interesting questions for follow-up”. We certainly shouldn’t treat the study as showing anything solidly empirical.

Update: Karl Smith weighs in.

9 comments

Comments are closed.