How can you tell when a government report found something interesting? When the caveat in the executive summary is shouted in loud italics: "It should be noted that this research was not designed to assess program impacts or in any way attribute differences observed between FSP participants and either group of nonparticipants to an effect of the program." This line certainly made me want to read the latest report on food stamps and nutrition, released last week by USDA's Economic Research Service.
You can see for yourself what might have raised eyebrows. Consider the mean body mass index (BMI) for food stamp participants and two comparison groups (BMI > 25 means "overweight" and BMI > 30 means "obese"). For women in particular, food stamp participants are more likely to be overweight:
But now consider the women's self-perception of their weight. Food stamp participants are less likely to perceive themselves as overweight.
The report's authors are surely right to avoid jumping to conclusions from one study, but at the very least this pattern raises the rank of comparatively burdensome policy options: heavier emphasis on nutrition education, for example, or targeted "green stamps" for fruits and vegetables, at least as part of the benefit package.
In any case, there has already been more than one such study. Economist Diane Gibson published an interesting article about the positive association between food stamp participation and risk of obesity a couple years ago. It seems to be available free online from the Journal of Nutrition(but please let me know if the link fails to work). Overall, the association between food stamps and obesity has been found in diverse data sources. The question is why? This is a question I will follow as this weblog develops.