Red State, Blue State, Rich State, Poor State: Why Americans Vote the Way They Do

By ANDREW GELMAN

Andrew Gelman’s Red State, Blue State, Rich State, Poor State can be summed up with extreme concision: Rich people vote Republican, but rich states vote Democratic. Poor people vote Democratic, but poor states vote Republican. That’s pretty weird.

But to Gelman, it’s worse than weird. It’s unknown. Even by the people who are paid good money to know things like that. And it’s that troubling fact which seems to have spurred Gelman to write this book. His introduction begins with a quote from pundit Tucker Carlson: “Here’s the fact that nobody ever, ever mentions — Democrats win rich people. Over $100,000 in income, you are likely more than not to vote for Democrats.” As Gelman goes on to explain, there are no fewer than two large falsehoods in that statement. First, lots of people mention that “fact.” Gelman quotes a number of them at length, ranging from left-wing populists like Thomas Frank to right-wing culture warriors like Michael Barone. Second, the statement is false. As Gelman says, “If poor people were a state, they would be ‘bluer’ even than Massachusetts; if rich people were a state, they would be about as ‘red’ as Alabama, Kansas, the Dakotas, or Texas.”

There are a variety of reasons for the confusion, but the empirically interesting cause comes in the form of what Gelman calls his “supergraph.” This graph is important for the book, so it’s worth describing in some detail. One axis tracks the income of voters; the other axis tracks the probability of voting for Bush. Gelman plotted three states across these metrics. One, Mississippi, is a low-income state. The next, Ohio, is a middle-income state. And the third, Connecticut, is high-income. “Our original thinking,” writes Gelman, “was that some states were more Republican than others, but that there would be a consistent pattern of income and voting within each state.” It was not to be. Connecticut exhibits a gentle slope. The rich are more Republican, but not by much. Ohio is a steeper incline. The rich are a fair bit more Republican than the poor. And Mississippi shoots sharply upward. The rich are far more Republican than the poor. Why should rich states be less sharply polarized than poor states?

It’s an interesting puzzle, and one that Gelman spends much of the book trying to solve. The basic answer is that Marx was wrong: Class ain’t everything. “With growing political polarization,” writes Gelman, “parties have started to take ownership of social and religious issues, such as abortion, that used to straddle the political divide?low-income voters are pretty similar in their political attitudes, but high-income voters show much more geographic variation, with rich red-staters being clearly more conservative on social issues than rich blue-staters.” The rich in red states, as it turns out, are likely to be evangelical Christians. The rich in blue states are not. This matters. Race is a big part of it, too. African Americans are disproportionately poor. They’re also disproportionately Democratic. When you’re looking at Mississippi, say, this will make the class divide look incredibly sharp. But, in part, you’re seeing a racial divide.

“The division,” concludes Gelman, “is not simply between rich and poor voters, or religious and secular, or rich and poor states. Each of these factors is important, but the true differences lie in subsets of the population. Rich versus poor voters in poor states, high-income religious versus high income secular voters, red states versus blue states among rich voters?the rich-poor divide is a superposition of many changes.”

At the most basic level, this is an argument for complexity. The country is not as simple as some would have it, and if that means political discussion segments need to be lengthened from two minutes to four minutes, then tough. But it’s also an argument for data, and for increased rigor among the chattering class.

Gelman, for one, is bothered by the fact that so simple a question — How do rich people, on average, vote? — is so frequently misunderstood. Searching for an answer, he settles on the existence of “availability bias.” Availability bias is the term used for the human tendency to generalize based on nearby information. For instance: If I were governor of Alaska, a state that runs on oil revenues, I might think economic governance in general was quite easy and required few tradeoffs and I was fully prepared to do that task on a larger scale. I’d be wrong, but it would have tracked with the nearby information. Similarly, most journalists are well educated, have high incomes, and live in rich states (broadly speaking, California, New York, Maryland, DC, and Virginia). And most of their friends and colleagues vote Democratic. But this is not the norm. As Gelman notes, “richer counties tend to support the Democrats within the media center states of Maryland, Virginia, New York, and California but not, in general, elsewhere. And richer voters support the Republicans everywhere, but this pattern is weaker — and thus easier to miss — within these richer states.”

What Gelman does not engage is that this mistake is also quite convenient. Journalists, after all, have access to the same exit polls Gelman is using. They could look this up. So, too, could producers and editors. Those who misstated the facts could be censured. But they are not. Rather, they advance rapidly through the ranks. This is due to the profession’s preference for narratives it understands, knows, and promotes. One such narrative is that Democrats are the party of elites. Start with the presumption that that’s true, and though most rich people actually vote for Republicans, the idea that they vote for Democrats is, in Stephen Colbert’s memorable term, truth-esque. That’s not availability bias. It’s professional bias. And you can’t solve it with a graph.

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>