Via Andrew Gelman comes an interesting paper from S. V. Subramanian and Jessica Perkins suggesting that Republicans are healthier than Democrats, even when you control for such factors as race and income and education:
I'd be a bit careful with this data: Black Republicans are relatively rare, as are poor Republicans, and so you're probably not dealing with precisely like populations, even as they share certain characteristics. One of the big things I'd like to see controlled for is geography: The urban poor are different than the rural poor, but urban poor tend to be Democrats while Republicans do better among rural populations. But even with those caveats, the data is provocative, and fairly hard to explain.
I'd be a bit careful with this data: Black Republicans are relatively rare, as are poor Republicans, and so you're probably not dealing with precisely like populations, even as they share certain characteristics. One of the big things I'd like to see controlled for is geography: The urban poor are different than the rural poor, but urban poor tend to be Democrats while Republicans do better among rural populations. But even with those caveats, the data is provocative, and fairly hard to explain. 
