Monday, February 11, 2013

Liberals biased towards Trader Joe's? Conservatives for Walmart?

A recent story on the Pacific Standard web site -- which has since been repeated on the Huffington Post and many other aggregators’ sites -- purports to ‘explain’ liberals’ bias towards stores like Trader Joe’s, and conservatives’ bias towards Walmart.

Color me skeptical.
The underlying academic study was published by a researcher named Vishal Singh in the journal Psychological Science. It was, in fact, not a study of Trader Joe’s vs. Walmart per se, or the two chains’ customers. Rather, it associated religiosity and a tendency to vote Republican in conservative-leaning counties with those counties’ consumers’ tendency to favor branded over generic products. 

The research was a bland piece of statistical meta-analysis. Journalists who picked up the story turned it into a meme by leaping to the conclusion that it provided scientific confirmation of their own opinions about Trader Joe’s and Walmart customers.

From my perspective as an expert on Trader Joe’s, the conclusion is flawed. It was wrong to cast Trader Joe’s as a supplier of generic products, for the simple reason that to Trader Joe’s customers, ‘Trader Joe’s’ is a highly-desirable brand in its own right.

I’m not surprised that the ‘journalists’ who repeated Pacific Standard’s flawed conclusion believe that liberals are biased towards Trader Joe’s. Until very recently, almost all Trader Joe’s stores were located in blue states. Likewise, Walmart stores tend to serve consumers who live outside of the U.S.’ largest urban centers. (There’s not a single Walmart anywhere on the San Francisco peninsula, nor is there a Walmart anywhere in Manhattan or Brooklyn.) If there are no Walmart stores in such big liberal enclaves, it’s not surprising Walmart customers skew conservative.  

The truth is always more complex. 

The Trader Joe’s store where I worked, in Kansas City, served a diverse population. Two miles to the east, there was a neighborhood that was poor, high-crime, and mostly African-American. I served lots of customers who Mitt Romney would have happily characterized as congenitally Democrat -- members of the 47% of the population subsisting on hand-outs. Two miles to the west of my store was the posh suburb of Mission, KS. You can be sure shoppers coming from the west voted Republican. 

I’m not surprised, either, by the sketchy merit of the underlying study, in spite of Psychological Science’s claim to be one of the top-10 peer-reviewed journals in its area.  This study’s not based on actually observing the purchasing habits of conservative consumers, or even on surveys of such consumers. 

Singh’s study analyzed “a comprehensive database that tracks weekly store sales of thousands of products. Focusing on 416 counties which collectively represent 47 percent of the U.S. population, they calculated the market share of generics in 26 categories, including coffee, deodorant, and peanut butter.” Reports don’t suggest the researchers are capable of parsing the subtle distinction between ‘generic’ and ‘own-brand’ products.

I don’t know enough about the database to comment on the likelihood that it reflects total consumption. But even if the database included all grocery purchases -- which it certainly doesn’t -- Singh merely proves that counties in which a high percentage of the population self-report as religious and/or Republican are also counties in which sales of generic products (and newly introduced branded products) lag. Singh’s research doesn’t -- and can’t -- prove that those religious and/or Republican consumers are the very same people eschewing generics.

In fact, the research doesn’t actually prove that anyone is rejecting generics. What it proves is that academic psychologists don’t know anything about the grocery business.

Walmart, for example, is the largest grocery retailer in the U.S., and it concentrates on selling national brands and deemphasizes generics. As noted above, Walmart skews rural-to-small-city. Religiosity and Republican voting patterns also skew that way. So, an inverse relationship between generic products and religiosity/Republicanism is an expected association, and doesn’t necessarily mean those religious, Republican consumers are consciously avoiding generics. It only means that the stores where they shop push brands, not generics.