The headline on yesterday’s New York Times piece on the demographics of Tea Partiers read: “Poll Finds Tea Party Backers Wealthier and More Educated“—than the general public, that is. The nutgraf adds that they’re also likely to be older, whiter, maler, and (shocking!) more conservative.
Now, the obvious question for me is: Why wouldn’t you try to disentangle a bunch of factors that we know full well are strongly interdependent? After all, we know that older white men are likely to have higher incomes and more schooling than a random sample of the population—and, of course, education and income are themselves related. Obviously, the headline gets a lot less interesting if it turns out to amount to: “Tea Party Backers Exactly as Wealthy and Educated as You’d Expect for Their Demographic.” Though without someone doing a proper regression, we don’t know if that’s true either.
Of course, you can flip the direction of explanation. Suppose, plausibly enough, that a movement concerned with the level of taxation and debt might disproportionately appeal to higher earners. A group like that will tend to skew whiter than a random population sample even if race per se is not motivating people to join up.
I have neither the time nor the chops to sort this out, but it does seem odd that nobody at the Times decided to commission an analysis. The interesting result is not whether the Tea Partiers are whiter or older or richer or more educated than the general population—if you know they differ along one dimension, it’s reasonable to expect them to differ along the others—but which features stick out once you’ve controlled for the others. It is, after all, perfectly possible to be wealthier than “average” but substantially poorer than the average 50-year-old white male. Anyone feel like running the Times‘ data and trying to get something more useful out of it?