In the International Journal of Epidemiology, S. V. Subramanian and Jessica Perkins write that, after controlling for age, sex, race, marital status, religious service attendance, highest educational degree, and total family income, Republicans were 25% less likely than Democrats to report being in poor health. They find a key component of this to be smoking: after controlling for that above list of variables, Republicans were 15% less likely to be smokers.
The analysis is based on data from 1972-2006, and I think a lot more needs to be done before I’d know what to make of it. Subramanian and Perkins write:
The observation that republicans enjoy better health status may reflect the core republican value of individual responsibility, which could translate into increased adherence to health-promoting behaviours. . . . It may also be that republicans exhibit greater religiosity (beyond attendance) compared with democrats, which may lead to health promoting social conditions such as enhanced social ties and networks. Alternatively, it is possible that our measures of SES (income and education) are inadequate in terms of controlling for one’s SES. The effects of identifying with the republican party, however, did not alter substantially in unadjusted and adjusted models.
And they conclude:
Whether one’s political ideology is an independent risk factor, or a marker of something else, clearly requires further research. It does not seem implausible, however, that conservative values at the individual level may be health promoting.
It’s also been suggested that the causation might go the other way (“A Democrat is a Republican who got sick,” to update that old saying), but, although I can see that happening in an aggregate level–people moving toward the Democrats partly because of dissatisfaction over issues such as health care–it’s hard for me to believe this would be explaining the pattern at an individual level.
I think a lot more could, and should, be done here, for example using the full four-point scale for self-reported health and having a better understanding of the transition from raw comparisons to the regression model. But, whatever ultimately comes of these findings, I’m supportive of the general idea of looking at correlations from these databases. There are lots of possible dead ends and wild goose chases (and this might be one of them) but it’s certainly a step up from the bald speculation we often see, and it can sometimes lead somewhere interesting.