Do latte drinkers really vote for Obama?
The media is abuzz with speculation that Barack Obama is an elitist. A few weeks ago, at a rally for Hillary Clinton, one of her more strident local boosters proclaimed the following about Obama supporters:
“I’ve got news for all the latte-drinking, Prius-driving, Birkenstock-wearing, trust fund babies crowding in to hear him speak! This guy won’t last a round against the Republican attack machine. He’s a poet, not a fighter.”
-Tom Buffenbarger, Youngstown, Ohio
You may have cheered, you may have been offended, but hearing Mr. Buffenbarger’s words you must also have wondered: is he right?
Having previously drawn political fire for suggesting the connection between steak, sushi, and voting habits, your plucky statisticians at Urbanspoon decided to roll up their sleeves and dive into this conflict.
Do latte drinkers really vote for Obama?
Is there really a connection between sipping your double tall breve and voting for Obama? We’ll leave political analysis to the professionals, but this is the kind of food question we’re equipped to investigate. Unfortunately, we can’t directly measure how much latte everyone is drinking. But as an approximation, we looked at the number of Starbucks stores per capita on a state-by-state basis. Compare this to how states voted in the primary:
The blue line measures the percentage by which Obama beat (or lost to) Clinton. The green dots represent the number of Starbucks stores per million people for each state. The black line is the trend line of Starbucks stores, drawn to make it easier to see the relationship between voting and latte sipping.
My statistics professor might be rolling over in his grave to hear me say it, but there’s a mild but real correlation here!* States with more latte-purveying Starbucks stores are more likely to have gone for Obama.
Is it possible Mr. Buffenbarger was on to something? We should probably call it a day, but let’s leave our comfort zone (food) and look at the rest of his statement.
Toyota won’t tell us how their buyers are voting, so again we’ll need an approximation. Fortunately, the folks at hybridcars.com and R.L. Polk came to our rescue. They track how many hybrids are sold per capita in each state. Comparing that with voting preferences, we can see how states with lots of Prius drivers vote, and by inference who those individuals support.
Huh. There isn’t the faintest whiff of a correlation here. I suppose it’s possible that all Prius drivers are, in fact, Obama supporters, but that’s sure not reflected in the way hybrid-friendly states are voting. Let’s move on.
Where the Prius failed, perhaps Birkenstock will prove a better measure of elitism. Birkenstock was kind enough to provide some sales statistics so that we could once again map against voting this primary season:
Wow. Nothing. Nada. No correlation.
Birkenstock asked me not to publish the specific numbers, but if you’re wondering about the outliers — the big winner, with the most Birkenstocks per capita was a shock to me: Wyoming. Not exactly the typical “elitist” state.
Trust fund babies
This is a tricky one, because there’s no National Association of Trust Fund Babies to turn to. We could just assume all trust fund babies are sipping latte at Starbucks and refer to figure 1, but that wouldn’t serve the interests of science. Instead, let’s assume that trust fund babies have one thing in common, money, and look at how wealth stacks against voting habits. Let’s look at per capita earnings per state, and compare that to how each state voted.
Again, what’s really striking here is how little correlation we see.
Stick to the latte
Based on this exacting analysis, what can we say about the upcoming Pennsylvania primary? Well, apparently it’s meaningless that Pennsylvania has fewer Priuses (Prii?), fewer Birkenstocks, and a lower per capita income than the national average.
However, they also have fewer Starbucks than average, and that could make a difference. If you have only Mr. Buffenbarger’s criteria to go by, Mr. Obama might be crying into his latte come April 22nd.
On a personal note: I don’t drink lattes, don’t drive a Prius and I certainly don’t have a trust fund. I do, however, have a great pair of Birkenstocks. Can you guess who I voted for?
* For all those statisticians out there, the correlation coefficient relating the number of Starbucks locations to Obama voting is 0.31. You might argue the outliers are exerting too much influence here, but removing them did not markedly change the correlation.