Are Republicans healthier than Democrats?

From S. V. Subramanian and Jessica Perkins.


P.S. See John’s comment below. He seems to have a good point. More here from Steve Kass.

My bad in not screening this more carefully before posting. In defense of Subramanian and Perkins, they sent me the paper and it was my idea to blog it. They were planning all along to do more systematic analysis of the raw data (which they haven’t yet received).

6 thoughts on “Are Republicans healthier than Democrats?

  1. Hmm. Don't they seem to be overinterpreting correlation as causation? I can just as easily make a case for reverse causation — just as it used to be said that a conservative was a liberal who had been mugged, a Democrat is a Republican who needs health care.

  2. "Self-rated health", hmmm, some biases possible there? Check out for a lightweight summary of research into various biases in self-opinion that authoritarian personality types have relative to non-authoritarian.

    As for smoking, I find it incredible that only about 5% of strongly Republican people in these three sub-demographics smoke, unless the survey was taken in Utah or Idaho. Another self-reporting bias, perhaps? I really need to see the data on this one.

  3. I'd say John found a smoking gun:

    I generated a crosstab of {smokes, doesn't} vs. {Black Strong Rep, Black Strong Dem} from the tool Subramaniam and Perkins say they used (Reference 3:
    <a href="; rel="nofollow"><a href="; rel="nofollow"&gt <a href="http://;” target=”_blank”>;

    You can look at the graph I got. It's very different:
    <a href="; rel="nofollow"><a href="; rel="nofollow"&gt <a href="http://;” target=”_blank”>;

    Subramanian and Perkins also say "We also found that more than twice as many Democrats smoke (16.7%) compared to Republicans (7.1%)." Significantly more than 16.7% of adults smoke. It looks like S. and P. misused the GSS data tools.

  4. I think I've got an explanation. (I also posted it to your other blog. Feel free not to approve both copies.)

    The graph presents real data, but it's mislabeled and incorrectly explained. Specifically, the captions "Percent in Poor Health" and "Percent who Smoke" should say "In Poor Health" and "Smokers." The graph's bars do not represent the prevalence of smoking, poor health, etc. The bars represent the prevalence of strong Democratic and strong Republican party identification.

    Here's an explanation of the leftmost pair of bars in the graph. The other pairs are analogous.

    In the GSS data set, 1,687 respondents reported both poor health and less than high school education. Among these 1,687 respondents, party affiliation went as follows:

    Strong Dem: 27.0%
    Not-Strong Dem: 23.6%
    Independent near-Dem: 9.1%
    Independent: 14.9%
    Independent near-Rep: 4.6%
    Not-Strong Rep: 11.2%
    Strong Rep: 8.0%

    Subramaniam and Perkins display the Strong Dem and Strong Rep percentages.

    In other words, among people reporting poor health and less than a high school education, 27% call themselves strong Democrats and 8% call themselves strong Republicans.

    Even if they'd explained it correctly, the data they present isn't useful without more context and comparisons, and I'm not convinced there's much enlightening here.

    Among black smokers, about 36% call themselves strong Democrats and about 2% call themselves strong Republicans. Is anyone surprised?

    The best thing about the article for me (and it's a really good thing) is that it pointed me to the very cool online data tool on the GSS web site.


Comments are closed.