'Food Deserts' or Mirages?
A second look at the data
29 Jul 2011 in Regulatory Science, Regulatory Economics, Information Quality
Yesterday's post on the federal government's new program to eradicate "food deserts" provided an obvious example in which the database yields nonsensical results. A large US Army installation, one with its own grocery store, no unemployment, and essentially no poverty, nevertheless appeared to meet all of the government's (arbitrary) criteria for inclusion.
A closer look shows that there is yet another reason why this particular "food desert" is a mirage. Soldiers who live within this census tract actually do live within one mile of a second, large grocery store--albeit one that is in an adjacent census tract.
The same Food Locator map is reproduced below, this time with two additions. First, a circle with radius of 1 mile has ben centered on the most distant housing unit within the census tract. Second, a large grocery store located in the adjacent census tract is plotted with a red x. This store lies within the 1-mile radius.
Even if the residents of this census tract were poor (they aren't) and lacked a large grocery store within their census tract (they don't), they'd still live within one mile of a large grocery store.
