Trego's Mountain Ear

I had a role in some bad research.  It was back in 2007, published in 2008, and a project my head of department authorized.  Donna’s listed as co-author, as I am – but the lead authors were grad students at the time – Trevor and Staci.  Rural Life Census Data Center Newsletter: South Dakota’s Food Deserts is the address of our blunder.  It’s still getting downloads.

I’m thinking of this because I read of Stanford’s president’s resignation this morning – over faulty research showing up in a 2009 article.  There are some differences – but it’s worth talking about the opportunities that exist to do flawed research.

The nice thing for me was that my department head, Donna Hess, started the project, and Trevor was the lead author – so I wasn’t the man who got the phone call from a grocer in Small Town South Dakota who was angry about his town, his county, being classified as a food desert.  Trevor spent over an hour learning just how screwed up his research and findings were – and, at the end, we still had to figure out what had happened and how we went wrong.

Food deserts had become a catchphrase.  The definition was “Food deserts are areas in which all residents “have low access to large food retailers” (p. 1); specifically, each person in a food desert lives more than 10 miles away from a supermarket (Morton and Blanchard 2007).”  Our article started with that phrase – but none of us had checked which piece of data Morton and Blanchard had used to indicate a food store that qualified the county as a food desert.  It was the number of employees in the retail food store – readily available from the Bureau of Labor and Statistics.  Shannon County (Pine Ridge) had a large turnover of part-time employees in their grocery store, so the Pine Ridge didn’t qualify as a food desert. 

I made a point of visiting the Small Town grocer who had been so angry about his town’s classification as a food desert.  The plus was that I had been included as co author because of a separate research project that was yet unpublished, and the article that he found so offensive documented that.  His grocery operation was ran by a large family who worked long hours and did not meet Morton and Blanchard’s criteria – but was definitely not a food desert.   It’s easier to make apologies when the guy you’re apologizing to knows it really wasn’t your blunder (being 6’3” tall may also help get your apologies accepted.)

Obviously, in rural America, the number of store employees didn’t correlate well with the diversity and quality of food in the store.  Just as obviously, the correlation was pretty good in urban America.  The thing was, our department was rural sociology – and the research techniques that worked well in cities with populations larger – much larger – than our state fell apart in counties with populations of three or four thousand people. 

All of the authors are gone – but the article continues to be available, and downloaded.