USDA report raises issue of food programs and obesity

USDA's Food and Nutrition Service this month published a substantial report about food assistance programs and obesity. At times, the genuine difficulty of this delicate issue seemed to tie the writers into knots. Here is the first main conclusion (bold in the original):

The panel concluded that to be able to determine the relationship between obesity and food assistance program participation, it is necessary to consider the difficulties and complexity of separating the effects of poverty from the potential effects of food assistance on any health or social outcome including obesity. The challenge is in controlling for the effects of poverty, which is highly correlated with program participation. Because of the strength of the existing research on this point, it is imperative that this established association be presented as the first point in their consensus statement.

But at other places, the report seemed to call for serious research to evaluate both the positive and the potential negative effects of food assistance programs. On food stamps, the report concludes, "The published research on Food Stamps indicates there is some association between program participation and overweight/obesity; but there is no evidence of causality." The report does not endorse random assignment research designs (whereas I have favored ethical random assignment designs), but instead recommends strong longitudinal research designs (which I think could be adequate if done well). It offers astute observations on the potential for making use of recent policy variations, such as the institution of food stamp outreach efforts, to better control for selection bias. I hope something comes of it.