Whole Foods and Elections: Proof That Political Data Can Be Absurd

Jeff J Mitchell/Getty Images

Jeff J Mitchell/Getty Images

Let’s say you’re attending a horse race and you have a nice thick wad of cash rolled up in a rubber band; you’re ready to place your bet, go home a winner, and quit your job at the post office (because in this hypothetical you’re Charles Bukowski, apparently). There are a myriad of different things you can consider to help you decide which restless and snorting steed to risk your life savings on; the jockey’s skill, the horse’s record, its breeding, knowledge of equestrian anatomy, and of course any insider tips you’ve picked up from your bookie friends at the bar.

Political elections are something like a horse race. Polling groups, prognosticators, and analysts sift through all of the available signs and data in hopes of recognizing the best bet. Like a mildly toasted gentleman eying up the horses in his Sunday best, polling and data groups like Gallup and FiveThirtyEight can conduct phone surveys; speak to Americans to get an idea of candidate reception by race, gender, and socioeconomic demographics; look at political climate based on other members of the same party; or compare present day reactions to historical political events to see if they contain any predictive hints.

However, they also look at a wide variety of statistics and current events you’d never in a million years consider relevant, and sometimes it leaves you wondering: “Is this really worth thinking about?” Case in point: FiveThirtyEight recently critiqued a NBC “Nerd Screen” election breakdown, which sought to compare the Democrat versus Republican battle to “rural ‘Chick-Fil-A country‘ vs. urban ‘Starbucks country.’”  The NBC report claimed that areas with a concentration of each business would be where you’d see the toughest congressional battles this fall, with Starbucks representing Democrats and anti-gay Chil-Fil-A representing conservatives.

FiveThirtyEight’s David Wasserman pointed out that the geographic locations of each business don’t line up with the demographic assumption made by NBC’s Chuck Todd. Chick-Fil-As are heavily located in many liberal states, including one in New York University’s residence hall in Greenwich Village, and Starbucks (as we all know) are basically everywhere. It’s the glitter of the beverage industry, so even if its CEO Howard Schultz’s support of Obama in 2012 makes for a convenient partisan example, it doesn’t mean its locations are restricted to liberal city centers.

Then, Wasserman went on to reintroduce the argument, this time with Cracker Barrel Old Country Store and Whole Foods Market. His critique was not that NBC had sought to study elections via restaurant placement, but that NBC had chosen the wrong restaurants to do so. He argued that not only were Whole Foods and Cracker Barrel more representative of the regional political breakdown, but both also represented a “culture gap” that has been growing, as shown in the table below.


It’s a little bit laughable — tantamount to hitting up the Indy 500 in the 1960s, taking Mario Andretti’s blood pressure, recording the air pressure in each race car’s front left tire, and using this information to decide whether or not to bet on Seabiscuit. Here’s the thing, though: he has a valid argument. You can compare these things. You can argue, or at least try to argue, that the “Organic vs. Nostalgic” gap increase is indicative of “a harbinger in midterm cycles.” That’s what’s so marvelous about the data playground — there are so many different things you can look for correlations between, and in today’s online age, you can access an incredible backlog of information and statistics.

Too many studies have precautionary admissions like this one from FiveThirtyEight: “Granted, frequenting the Whole Foods hot bar doesn’t make you a liberal Democrat and pulling over to eat at Cracker Barrel on your next road trip doesn’t mean you’re a raging Republican … And counting up the number of each chain’s locations won’t tell you who will win the tight Senate contests in Colorado, Iowa, or North Carolina.”

So what’s the point, then? The Cheat Sheet is, of course, guilty of this as well, because realistically all publications are driven by a desire to pick apart the information at hand. That’s the point: it’s a knee-jerk nerd reaction in the best possible sense, and it’s worth looking at because it’s a new angle from which to gaze. Sometimes you dissect, analyze, and discuss to no avail, and sometimes you find a golden nugget of a thesis that isn’t really true or false, just suggestive — as we see with FiveThirtyEight. But in the end, you’re never going to turn down hearing a tip on the next horse race, even if you aren’t sure what you’re willing to place your bets on.

Follow Anthea Mitchell on Twitter @AntheaWSCS

More from Politics Cheat Sheet: