I’m a bit late to the pile-on over Thomas Frank’s recent attack on political scientists and data-oriented journalism. I won’t attempt to improve on what folks like Ed Kilgore, Jonathan Bernstein, and Jonathan Chait have already capably written. I just wanted to provide a bit of context that I think helps explain Frank’s rather caustic view towards political science.
Back in 2005, political scientist Larry Bartels wrote up a thoughtful paper (pdf) critiquing Frank’s highly successful and enjoyable book What’s the Matter with Kansas?. The book’s central thesis is that the Republican Party is basically duping working class white voters in places like Kansas by running on cultural promises (gun rights, abortion bans, etc.) and then largely ignoring those issues once in office, turning their attention instead to cutting taxes and gutting social services, which only ends up hurting those voters. As Bartels demonstrated with a detailed analysis of ANES data, that book rests on a faulty assumption. By most reasonable definitions of the term, the “white working class” in Kansas and elsewhere outside the South is voting Democratic and has been doing so for some time. What’s more, there’s little evidence that these voters are being duped on cultural issues; voting on economic issues has become more pronounced over time, not less so.
Here’s how Frank responded to Bartels’ critique:
The fundamental assumption animating Bartels’ attack on What’s the Matter With Kansas? is that studies like mine—based on movement literature, local history, interviews, state-level election results, and personal observation—are inherently inferior to mathematical extrapolations drawn from the National Election Surveys…. My own feeling, after watching him steer his science around the proving ground, is that this vaunted research device is in reality a rickety and most unreliable contraption. To begin with, consider the barren landscape of American politics as Bartels describes it—a featureless tundra swept of history, ideology, and any hint of the raw emotional resonance that everyone knows politics possesses. His NES America is not a place that I recognize. It might as well be the moon….
Frank’s attitude seems to be that political scientists like Bartels are pointing to mathematical symbols like Ï€ or e and claiming “That’s America!” But of course that’s not what’s happening. These “mathematical extrapolations” come from an analysis of a detailed survey of American voters that’s been running since 1948, asking questions about partisanship, income, union membership, religion, beliefs, education, ideology, and more. These are the issues Frank claims to care about. And yet when he’s presented with evidence of the way Americans truly think and vote and behave in a thorough and unbiased format, he completely rejects it in favor of his preferred narrative. “It might as well be the moon.”
Frank is a very good storyteller. It is regrettable that he sometimes tells stories not just absent evidence, but in direct contradiction to evidence. And it’s all the more regrettable that in such situations, he rejects the evidence as alien.
[Cross-posted at Mischiefs of Faction]