Trump voters, authoritarianism, and legalism: Don't jump to conclusions

Update (4/17/17): Since writing this, I have read a few more relevant articles and studies, which I might incorporate into this post in the future. Each one is discussed in the National Review article 'Racism' (Still) Didn’t Elect Trump.

I was recently asked for my thoughts on an op-ed from January 2016 about Trump voters, written by Pastor Paul Prather in Kentucky.

The op-ed makes a few points:

• The common denominator of Trump supporters is, according to research, an authoritarian tendency.
• "Authoritarianism" maps neatly onto a theological preference for law over grace.
• The author favors grace over law, but believes that it's best when the two are held in balance.

Reading this, my first reaction was skepticism. The title of the op-ed was, "The mystery of Christians' support for Donald Trump is solved." But does the cited study have as much explanatory power as the author believes? I had several questions:

• What is an authoritarian personality anyway?
• Is there really such a strong correlation between Trump support and authoritarianism that it explains the Trump phenomenon completely?
• Can you really make the connection between authoritarianism and legalism (or whatever you want to call the Christian tendency he connects it to)?
• The op-ed was written during the primaries. Even if it applied then, could it really be applied to the general election?

Spoiler alert: after researching these questions, I remain skeptical. Let's take these one at a time.

What is authoritarianism?

Different writers and scholars have defined it differently, but the various articles from the election season seem to all go back to the scholarship of Karen Stenner. She says that the term "conservatism" in contemporary political discourse unhelpfully lumps together three distinct tendencies, which are substantially different and not significantly correlated with each other. In her article "Three Kinds of 'Conservatism'," which summarizes her book The Authoritarian Dynamic (brought to my attention via Jonathan Haidt), she says that the three different tendencies are:

• Laissez-faire conservatism
• Defined as "a persistent preference for a free market and limited government intervention in the economy."
• Measured by "respondents' positions on whether incomes should be made more equal (or allowed to vary as individual incentive), on private versus collective ownership and management of business and industry, and on whether government 'should take more responsibility to ensure that everyone is provided for.'"
• Status quo conservatism
• Defined as "an enduring inclination to favor stability and preservation of the status quo over social change."
• Measured by "two items gauging (on 10-point scales, anchored each end) the extent to which respondents agreed that 'one should be cautious about making major changes' (vs. 'you will never achieve much in life unless you act boldly'), and that 'ideas that have stood the test of time are generally best' (vs. 'new ideas are generally better than old ones')."
• Authoritarianism
• Defined as "an enduring predisposition, in all matters political and social, to favor obedience and conformity (oneness and sameness) over freedom and difference."
• Measured by "respondents choosing (from a proffered list of 11) those 'especially important' qualities 'that children can be encouraged to learn at home'; counting 'obedience' and 'good manners' as reflecting authoritarian tendencies; and 'tolerance and respect for other people,' 'independence,' and 'imagination' as indicative of libertarian inclinations (the second component reversed and equally weighted in the final measure)."

Again, Stenner goes to great lengths and pains to emphasize that the three are not significantly correlated with each other across time and space. She even puts "conservatives" in quotes to emphasize that she thinks authoritarianism is a totally different animal from the other two things commonly called "conservatism." Nor is there much correlation between being authoritarian and being a self-identified conservative. On a personal note, I unquestionably identify as a "status quo conservative."

Does authoritarianism explain the Trump phenomenon?

The op-ed relies on research by Matthew MacWilliams and Marc Hetherington. The best explanations and defenses of the cited research that I've found are an article in Vox and an article by MacWilliams himself, both from March 2016. They measured authoritarianism similarly to Stenner, and found that authoritarians are significantly more likely to favor Trump than to favor other Republican candidates.

However, this research was contradicted by another study by Eric Oliver and Wendy Rahn in March 2016, which found that Trump support was not very well predicted by authoritarianism, but extremely well predicted by populism. You'll notice that this is a new term, not found in my summary of Stenner's article. Rahn and Oliver defined and measured authoritarianism similarly to the others, but defined populism this way:

A type of political rhetoric that casts a virtuous "people" against nefarious elites and strident outsiders. Scholars measure populism in a variety of ways, but we focus on three central elements:

• Belief that a few elites have absconded with the rightful sovereignty of the people;

• Deep mistrust of any group that claims expertise;

• Strong nationalist identity

So who was right, Hetherington and MacWilliams or Oliver and Rahn? I don't know enough about their methodology to say that either was right or wrong, but I'm including the latter argument because:

• It seems more honest to include competent competing claims.
• Both the authoritarian and the populist explanations make intuitive sense.
• It provides context for an argument I'll make later.

However, if I have to pick one explanation over the other, I'll pick the authoritarian explanation. Matt Braynard, an internal pollster and strategist during Trump's primary campaign, revealed later that (while he objects to the term "authoritarian") some of the strongest indicators of Trump support were consistent with authoritarian tendencies.

So let's say authoritarianism correlated strongly with Trump support during the primaries. I'd give two words of caution: firstly, don't be too quick to conclude that because it explains a lot of his support, that therefore all Trump supporters in the primary could be explained this way. Secondly, don't be too quick to conclude that this explains his general election support. We'll return to both of these points later.

Are "Law Christians" authoritarian in the sociological sense?

I'll note a few problems with Prather's analysis:

• While explaining why "Law people" are a Christianity-specific strain of authoritarians (or at least are correlated with being authoritarian), he actually gives a description that sounds more like status quo conservatives: "They insist on doing things, whatever those things might be, the way they’ve always been done, whether or not that happens to make sense anymore, just because that’s the way they’ve always done them."
• He seems to commit the statistical fallacy of assuming that when X is correlated with Y and Y is correlated with Z, X must be correlated with Z. In this case, X, Y, and Z would be a Law focus, authoritarianism, and Trump support.
• In general, his description of "Law people" sounds an awful lot like run-of-the-mill social conservatism (though caricatured a bit). However (I'm having trouble finding actual data on this, so if someone could help out I'd appreciate it) I seem to recall social conservatives being one of the last Republican constituencies to come around to Trump. Maybe here too it's problematic to consider "social conservatism" monolithically, but I don't know of a way around that problem offhand.

In addition to those problems with the original op-ed, I'll note that it's problematic to pass it around post-election as a way of explaining Trump's general election support, or his post-election support.

Does authoritarianism (or populism) explain Trump's general election support?

This may be surprising, but it doesn't look like authoritarianism or populism explain Trump's general election support in any meaningful way.

In May 2016, Adam Enders and Steven Smallpage studied ideological and temperamental indicators of general election support. From what I can tell, the data is unavailable to the public but there are some useful data visualizations in the article. Some conclusions from the article:

• Partisan identification (Republican versus Democrat) was one of the strongest predictors of Trump or Clinton support, much like it had been in 2008 and 2012.
• Ideological identification (conservative versus liberal) was another strong predictor, much like in 2012, though in 2008 "extremely conservative" types weren't much more likely than moderate conservatives to support McCain.
• Racial resentment is another very strong predictor, so much so that the Washington Post made it their headline. I note that in 2008 and 2012, lack of racial resentment was a strong predictor of support for Obama, but its presence was not a very strong predictor of support for his opponents, unlike in 2016. I also note that a New York Times article that made the rounds during the primaries concluding that racial prejudice predicted Trump support in the primaries was bunk.
• Authoritarianism was hardly an indicator of support whatsoever, much like in 2008 and 2012. Surprise! Turns out there are authoritarians in both parties, and they largely fell in line after the primaries.
• Populism was hardly an indicator of support whatsoever, but it was a new measure in 2016.

Conventional wisdom said that Trump's strongman appeal, or populist appeal, was key to his November victory. Enders and Smallpage dismantled that. According to them, most Americans voted for their usual party, and once Trump sealed the nomination the prejudiced liked him more. Maybe populism and/or authoritarianism played a role in the primaries, but they didn't seem to in the general election. Similarly, it seems that racial resentment played a role in the general election, but it did not play a role prior to his nomination.

It seems reasonable to conclude also that laissez-faire conservatives and status quo conservatives supported Trump in similar proportions to how they usually support the Republican candidate.

What other sources can help explain general election support? How about exit polls? If you compare the 2016 exit polls with those of 2012 and 2008, you may notice some minor shifts, but to me the big story is just how similar the demographic data is between the three elections. Across racial, religious, and socioeconomic categories, the best way to predict the level of support for the Democratic or Republican candidate in one category in one election, is to view the level of support in the other two elections and estimate it as being very close.

So what does this mean? I think it means that partisan identification, disappointingly, is by far the most significant predictor. If you've voted Republican or Democrat in the past, you will in the future no matter how problematic the candidate. That's not true 100% across the board, but my point is that the data supports this conclusion in most cases. Or at least that a voter will sooner abstain from voting than cross party lines. (Arguably, this is basically what I did, since I'm a historically Republican voter who voted independent in 2016; sometimes, regretting Trump's win, I wish that I had voted Clinton, but I'm usually satisfied with my decision. Either way, it's bygones.)

What explains Trump's general election support from evangelicals, if not authoritarianism?

You mean besides partisanship, coupled with polarization? I don't know. Conventional wisdom says that it had a lot to do with abortion and the Supreme Court, but the data doesn't seem to back that up.

(Lack of) conclusions

Why was I so skeptical? Because one of the most important conclusions of 2016, I think, is that even the experts (or, if you're a populist, especially the experts) don't know the electorate all that well. So caution is warranted. (Incidentally, caution is one of the cardinal virtues of my "status quo conservative" tribe! My team scores!)

But I think my skepticism proved warranted in this case. What does this mean going forward? I can think of a few lessons:

• If you know someone who says they're conservative, don't assume you know what kind of "conservative" they are, or whether they support/supported Trump. If you want to know, ask.
• If you someone who support or supported Trump, don't assume you know what their reasons were. If you want to know, ask.
• Especially don't assume that, just because there is a strong correlation between one variable and support for Trump, that any individual Trump supporter has a strong presence of that variable. I'm thinking especially of racial prejudice; assuming that all (or even most) of your political opponents are overtly racist is poisonous to public discourse, and assuming that all or most are implicitly racist is unproductive to discourse. I confess that I'm not sure of the way forward on this point, since asking directly is unlikely to have good results.
• Don't assume that data explaining what drives a candidate's primary support will also explain what drives the candidate's general election support. In fact, if you don't have time to investigate it, assume that it doesn't.
• Don't be quick to latch onto one silver bullet narrative for why an election or other event happened the way it did. Or instead of saying "don't be quick to do so," I should just say don't. If you do, you'll undoubtedly be thinking too simplistically.

And here's the chief lesson. If we want to understand the other side (whichever side of whichever debate that happens to be), we can't be content with sweeping generalizations or with armchair psychology. We must be willing to listen to the other side.