Lying with Statistics

Huff Post carries an article with the shock title, "Who Gets Food Stamps? White People, Mostly." Saying nothing provably wrong, it provides a classic example of how to lie with statistics.

The article reports USDA figures on the race or ethnicity of SNAP, aka "food stamps," recipients. The numbers are as follows:

40.2%   White
25.7%   African-American
10.3%   Hispanic, any race
02.1%   Asian
00.7%   Multiple races
12.8%   Race unknown
07.0%   No household head listed

Meanwhile the table below reflects the percentage of the total U.S. population represented by each group, found here:

63.7%   Non-Hispanic White
12.6%   African-American
16.4%   Hispanic, any race
04.8%   Asian
02.9%   Multiple races

Note the Census lists no categories for "race unknown" or unlisted household head.

Is the claim made in the article that white people get more food stamps than any other group true? Yes, if the measure is total value of SNAP received.

The only identified group that gets more than their fair share of food stamps is African-Americans, something the Huffpo article specifically sets out to disprove. African-Americans make up an eighth of the U.S. population and receive a quarter of the food stamps.

Every other identified group receives a smaller percentage of total food stamps issued than their proportion of the U.S. population. Clearly, HuffPo believes that comparison is incongruent with the prevailing biases of its readership.