The headline blares “Racism motivated Trump voters more than authoritarianism” but Thomas Wood’s piece at the Washington Post is a bit more nuanced than that. Before we talk about it, though, let’s look at how Wood got his data:

Last week, the widely respected 2016 American National Election Study was released, sending political scientists into a flurry of data modeling and chart making.

The ANES has been conducted since 1948, at first through in-person surveys, and now also online, with about 1,200 nationally representative respondents answering some questions for about 80 minutes. This incredibly rich, publicly funded data source allows us to put elections into historical perspective, examining how much each factor affected the vote in 2016 compared with other recent elections.

The factors he looks at are authoritarianism and racism, but also income level. Income level is self-explanatory, but the former factors require carefully crafted questions. To gauge people’s relative authoritarianism, they asked questions about child-rearing. The more people emphasize following rules and respecting elders over self-reliance and curiosity, the more authoritarian they are. To gauge people’s relative level of racism, they are asked indirect questions that really amount to giving an explanation for why blacks remain lower on the socioeconomic scale. This is called the “symbolic racism scale” or SRS.

The finding here is pretty straightforward:

Moving from the 50th to the 75th percentile in the authoritarian scale made someone about 3 percent more likely to vote for Trump. The same jump on the SRS scale made someone 20 percent more likely to vote for Trump.

This data can be compared to previous elections going back to 1988. What’s surprising isn’t that Trump voters are more racist than Clinton voters, because the same finding is there for people who voted for Romney, McCain, Dole and the two Bushes. In fact, on three of the four questions that test racial attitudes, Trump’s voters were less racist than their Republican predecessors (the fourth question was a tie).

The big difference is among Democrats, or Hillary Clinton voters, who are far less racist in their attitudes than the Democrats who voted in any recent election, including the two for Barack Obama. The implications are bizarre, suggesting that a lot of racially bigoted people were willing to vote for Obama against an opponent who didn’t appeal too directly to their racism, but who flocked to Trump when he made “political incorrectness” central to his pitch. To be explicit here, a lot of racist Democrats voted for Obama and didn’t vote for Clinton, and they did it because of racism.

This suggests that if you want a racist’s vote, you have to make an appeal directly to their racism. Without it, he or she just might vote for the racial minority.

Maybe this data would become a little clearer with questions about gender attitudes, but there would be no historical data for purposes of comparison.

In any case, yes, racism played a bigger role than authoritarianism according to this large survey, but what really stands out is the data about income. Here we have data going back to 1948, and it was always the case that people in the top income quintile vastly preferred the Republican. That changed in 2012, and it changed dramatically in 2016. Rich people preferred Obama in his reelection and they preferred Clinton.

Looking at the lowest income quintile is interesting, too, because prior to 2016 that group had voted at the national average or more strongly for the Democrat is every election except Nixon’s 1960 and 1968 campaigns (but not his landslide 1972 election). They strongly favored Bill Clinton and Barack Obama in all of their elections, but they turned on Hillary.

It looks like Trump resembles Nixon in more ways than one. They both did better among low-income voters than a typical Republican. I guess this is the Silent Majority pitch, and the appeal of the Southern Strategy. Perhaps lower income Protestants were the most likely to abandon the FDR coalition to oppose the Catholic John F. Kennedy, just as lower income Protestants were more likely to abandon the Obama coalition and go for the guy telling them that their racist attitudes were being stifled and marginalized by the political correctness police.

I still think gender attitudes played a role here, but that’s just my conjecture. It’s important to know the complete picture because it’s hard to craft a response without it.

Overall, it confirms my observations about the county patterns of voting. The election was lost because low income/rural white voters who voted for Obama decided to vote for Trump. Ironically, racism played a big role in the flip even though almost everyone expected the opposite to happen (that without a black candidate, the Democrats would do better with the racists).

It just goes to show that you can think you’ve got everything figured out, but you never do.