The headline accompanying a new TIME interactive map of food stamp usage across the U.S. declares that “Republicans [Are] More Likely to Have Constituents Who Use Food Stamps.” The implication isn’t hard to discern. Congressional Republicans who want to pare the growth of food stamp spending — program enrollment has increased by nearly 50 percent since President Obama was inaugurated in 2009 — aren’t just heartless jerks. It turns out they don’t even care about their own starving constituents.
Unfortunately for TIME, the data don’t really support the sensational headline. The authors relied on incomplete data and an arbitrary (and inconsistent) definition of what constitutes “high levels of participation” in order to buttress the headline. Here are the first two paragraphs of the article:
When the House voted last September to cut $40 billion from the federal food stamp program over 10 years, all but 15 Republicans supported the measure while not a single Democrat did so.
But according to a TIME analysis of county-by-county food stamp enrollment data compiled by the non-profit Feeding America, it appears that House Republicans represent more districts with high levels of participation in the program than House Democrats. Of the 350 Congressional districts in which TIME was able to estimate the percentage of people enrolled in the Supplemental Nutrition Assistance Program, 76 had levels of 20 percent or higher. Of those, 43 are held by Republicans while 33 are controlled by Democrats.
The first thing you’ll notice is that data from 85 congressional districts — nearly 20 percent of the country — are missing. In fact, roughly a dozen states lacked complete county level data. And never mind that you need nine-digit zip code data, not county data, in order to completely isolate one congressional district from another.
Five states included in the analysis don’t appear to have used any county level data at all. Which states, you might ask? One blue state with above-average food stamp usage and four red states with below average food stamp usage: Oregon (21% usage), West Virginia (19%), Nebraska (10%), Utah (9%), and Idaho (7%). Four out of Oregon’s five House lawmakers are Democrats.
According to the article, the national average for food stamp usage is approximately 15 percent of the population. So why didn’t the authors use that as the cut-off and inform their readers which party’s congressional constituents were more likely to have above average food stamp usage? Good question. An even better question would be why the authors chose to make 20 percent their cut-off for the “high levels of participation” metric they used as headline fodder? Why not 25 percent or 30 percent? Without having access to the raw data it’s hard to know, but the astute reader can probably guess the answer.
It’s not until paragraphs 9 and 10 of the article that the authors finally reveal the true confession of the tortured data.
Because of limitations in Feeding America’s data, which is missing county-level data for about a dozen states, it is not possible to make a definitive correlation between SNAP enrollment in a district and the party of the member who represents it in Congress. (It is also difficult to translate county data into congressional districts in small or highly gerrymandered districts, leading to some margin of error.) Based on the available figures, there does not appear to be a significant difference in food stamp usage in Democratic or Republican districts, with both averages in the neighborhood of the national mean.
Yes, you read that correctly. “There does not appear to be a significant difference in food stamp usage in Democratic or Republican districts.” There goes that talking point. But what about this notion that “House Republicans represent more districts with high levels of participation in the program than House Democrats”?
[C]ongressional districts with the absolute highest levels of enrollment are more likely to be represented by Democrats[.]
Oh, so the congressional districts that are most dependent upon food stamps are represented by Democrats? Fascinating. That could also have made for a great headline, but it apparently wouldn’t have communicated the right message.