[cfcef]

Typical epi-nonsense: ignore the huge and omnipresent genetic and other confounds which lead to long pre-existing differences, find congenial event to blame as origin, then leap to the most exotic methods of transmission possible; bonus points for each mention of Holland, rats, or mice.

The South as a region is very distinct from the rest of the country: settled by different ethnic groups (no, not all 'whites' are genetically identical twins; consider the different genetic BMIs by country in http://www.nature.com/ng/journal/v47/n11/full/ng.3401.html ), different cuisines, different industrial patterns, different weather and climate... It is eyebrow-raising, to say the least, to bring in the Civil War as at all relevant and speculate even more wildly that it has intergenerational effects.

If they want to talk about that, they should be trying some sort of regression-discontinuity or other natural-experiment design involving borders or Civil War activity, not, effectively, saying 'look at how different the South is! the Civil War?!' Uh, yeah, I think we all know the South differs. That's not very helpful.

[creshal]

> It is eyebrow-raising, to say the least, to bring in the Civil War as at all relevant and speculate even more wildly that it has intergenerational effects.

I wonder what the results would be if you applied that method to, say, Poland, or north-eastern France.

[cfcef]

You'll pick up a lot of post-WWII differences, but you can still try. There's an interesting paper exploiting the meandering of river-defined national borders to do a discontinuity design for long-run effects of the Austro-Hungarian Empire, for example, which is at least not obviously wrong.