When I read the results of this poll on the relationship between obesity and long-term unemployment (high blood pressure and high cholesterol are also positively correlated with long-term unemployment) my immediate reaction was to wonder whether obesity was not in fact a proxy for race. Consider this finding about those who are among the long-term unemployed from the Urban Institute:
we see that blacks, relative to other groups, are disproportionately represented among long-term unemployed and discouraged workers. They make up 22.6 percent of the long-term unemployed, 10.5 percent of the employed, 25.9 percent of discouraged workers, and 15.0 percent of newly unemployed workers. Hispanics make up a somewhat smaller share of the long-term unemployed (19.0 percent), the employed (15.7 percent), and the discouraged (20.2 percent) than of the newly unemployed (23.1 percent).
There are all sorts of factors that are in reality proxies for race which confound findings of this sort. Zip code can be a proxy for race. Level of educational attainment can be a proxy for race.