by Patrick Murray
The release of Authoritarian Nightmare by John Dean and Bob Altemeyer raises important questions about the underlying values and motivations of the American electorate. A core part of their analysis is based on a survey of voters conducted with the assistance of Monmouth University’s Polling Institute. I discuss our participation in the project in another post, but here I take a deeper dive into the explanatory power of authoritarianism for American voter attitudes and behavior.
[Note: this is a lengthy “extra for experts” post aimed at those wanting to understand psychosocial dimensions of political attitudes and behavior – and whether it is even possible to measure these constructs. If you are only interested in polling to forecast the next election, look elsewhere. The analysis here is based on a survey of 990 registered voters conducted online from late October to November, 2019.]
The framework for the book’s analysis is found in a number of psychological scales developed by Altemeyer and others to measure perceptions of prejudice, social equality, morality and preferences for strong leadership. Altemeyer’s premise is that traditional religious values and authoritarian tendencies are interrelated. [However, I must emphasize that the following thoughts are mine alone and do not represent Altemeyer’s research or writing.]
The existence of a correlation between traditional social values and authoritarianism makes sense from a lay person’s point of view. People who value tradition are more likely to be threatened by changes to the social order they know – whether those changes are real or perceived. And if you are concerned about the world changing too rapidly, the more likely you are to cede control to a strong authority figure who will do whatever is necessary to stem or reverse that cultural shift.
In fact, many evangelical voters offer a similar rationale in their continued support of Donald Trump. His behavior may be antithetical to their stated belief system in many ways, but they can rely on him to fight for their priority concerns. And while none of them have actually articulated it in this way, it basically boils down to: “If some Constitutional norms need to be undermined to overturn legal abortion, then so be it.” The ends justify the means.
This description is admittedly a simplistic depiction based on one type of single-issue voter. For many other Trump supporters, though, the cultural shifts they hope to reverse are more amorphous than any particular policy. This type of person’s calculation is more about having to confront unknowns in their daily life – a sense of discomfort and discontent that they are not getting ahead while “others” are. In this context, the passive authoritarian is willing to cede control to a strong leader who can identify and vilify the “other.”
Altemeyer has been utilizing his scales in a variety of settings for nearly four decades, but this new book marks the first time they have been put to the test with a representative sample of the American electorate. As the discussion in the book illustrates, this new data supports many of his prior claims about authoritarianism.
How Does One Measure Authoritarianism?
As someone new to the scene, though, I examined his Right Wing Authoritarian (RWA) scale and asked whether it is a measurement of traditional values more so than it taps into a willingness to cede authority to a strong leader. A look at the 20 questions in the scale finds a mix of items, such as “Our country will be great if we… do what the authorities tell us,” as well as “God’s laws about abortion, pornography and marriage must be strictly followed before it is too late.” The scale includes constructs around strong leadership as well as items tapping into traditional views of sex and sexuality.
My question is whether we can tease out these two constructs. Altemeyer’s 20-item scale seems pretty solid as a measurement tool, with high reliability score of α=.96. I wondered if the scale hold up without the sexual morality component, so I excluded the most blatant religious value questions and replaced them with a couple of additional items in the survey that asked about adherence to strong leadership (see question list here). This new 11 item scale had a similarly strong reliability score (α=.91). And because academic social scientists have a tendency to overegg the sauce, I also created an even more efficient 5-item scale. This produced a similar level of reliability (α=.90).
Lo and behold, these new scales exhibit nearly the same distributions as the findings discussed in the Dean and Altemeyer book that showed increasing authoritarian tendencies among Trump supporters (Table 1). The mean authoritarian score of strong Trump approvers is more than twice that of strong Trump disapprovers across the board. The results here suggest that removing the most blatantly religious items from the RWA scale demonstrates a squarely different mindset among Trump supporters on the proper exercise of leadership to maintain “order.” In other words, there are greater authoritarian tendencies among Trump supporters regardless of whether the scale includes references to specific social “norms” of gender and sexual orientation.
| Table 1. | RWA Scale Means by Trump Job Rating | |||
| RWA scale… | Strongly disapprove (n=466) | Somewhat disapprove (n=43) | Somewhat approve (n=119) | Strongly approve (n=348) |
| Original 20-item scale (range=20-176; x̅=.84) | 54 | 82 | 101 | 119 |
| 11-item non-valence (range=20-178; x̅=.82) | 56 | 79 | 96 | 113 |
| 5-item non-valence (range=20-176; x̅=.87) | 56 | 81 | 101 | 125 |
Despite the different composition of these RWA scales, all three of them correlate significantly with a separate scale measuring religious fundamentalism, although they do so at notably different levels (Table 2). These correlations range from r=.83 for the original 20-item scale to r=.68 for the 5-item scale with the sexual norms items removed. It is also interesting that all three RWA scales correlate highly with a separate scale in the survey designed to measure racial and religious prejudice (between r=.78 and r=.84). These concepts – racial equality and religious piety – were not referenced directly in any of the RWA scale items. While the religious aspect may be implicitly tied to authoritarianism because of how the original scale was constructed, racial prejudice is not. Yet, the two scales are indeed highly correlated. [By the way, the direct correlation between the prejudice scale and the fundamentalism scale is r=.65. Prejudice is not as strongly related to evangelism qua evangelism as it is to authoritarianism. These findings hold even when the analysis is run among white voters only.]
| Table 2. | Scale Correlations (all are significant at p<.01) | ||
| RWA scale… | Religious Fundamentalism | Prejudice | Social Dominance |
| Original 20-item scale | .83 | .84 | .68 |
| 11-item non-valence | .72 | .78 | .62 |
| 5-item non-valence | .68 | .79 | .63 |
The survey included another scale – one that may be even more telling about the underlying psychosocial dimensions of political behavior. The 16-item Social Dominance Orientation (SDO) scale measures a belief in and preference for maintaining hierarchy in society in general. [Note: the developers of this scale note that it includes two distinct dimensions – group dominance and anti-egalitarianism – that are meant to be scored separately. However, the correlations of the separate components to the other scales in our survey are nearly identical, so for ease of discussion I am using a summary scale value for SDO.]
The SDO scale does not specify any group identities, but it still has a strong correlation with the Prejudice scale (r=.80). This is much stronger than its correlation with the Religious Fundamentalism scale (r=.50). SDO also correlates with RWA – between r=.62 and r=.68 depending on which RWA scale is used. Other research has shown a weaker correlation between SDO and RWA, but those studies were conducted primarily with university students, whereas this study was conducted with a representative sample of American voters. The gap between Trump supporters and other voters is apparent on both scales, although the social dominance dimension varies over a smaller ranger (Table 3).
| Table 3. | Scale Means by Trump Job Rating | |||
| Strongly disapprove (n=463) | Somewhat disapprove (n=43) | Somewhat approve (n=119) | Strongly approve (n=346) | |
| RWA scale | 54 | 82 | 101 | 119 |
| SDO scale (rebased to match RWA scale range) | 45 | 70 | 81 | 90 |
The Dean and Altemeyer book suggests that while authoritarian followers provide Trump’s base of support, social dominators in his camp may pose the bigger threat. Basically, they see authoritarianism as passive and social dominance as aggressive (e.g. think of QAnon). About 1-in-8 survey respondents scored in the top quartile on both the RWA and the SDO scales and among this group, 89% are solidly in the pro-Trump camp. Among those who score highly only on the RWA scale, solid Trump support stands at 74% and among those who score highly only on the SDO scale solid Trump support is 58% of this group.
We know these “Double-highs” – as Dean and Altemeyer call them – on the RWA and SDO scales are firmly in Trump’s corner. The question is how far are they willing to go to back him? One question in the survey is illustrative. It asked what should happen if Trump loses in November but he “declares the election was fixed and crooked” (Table 4). Nearly 1 in 4 (23%) double-highs agree that Trump should continue as president in this situation, while 53% disagree. Other high RWAs are not far behind (19% agree and 66% disagree) while other high SDOs are not as willing to back an unconstitutional extension of Trump’s term (5% agree and 79% disagree). That puts them in basically the same category on this question as voters who score moderately high on either the RWA or SDO scales (5% agree and 85% disagree).
| Table 4. | Trump should continue in office despite a loss if he declares election was fixed and crooked | ||||
| Double High (RWA&SDO) | High RWA only | High SDO only | Moderate RWA/SDO | Low on both | |
| Agree (7%) | 23% | 19% | 5% | 5% | 1% |
| Neutral (10%) | 24% | 15% | 16% | 10% | 1% |
| Disagree (83%) | 53% | 66% | 79% | 85% | 98% |
Now, I am not a big fan of hypothetical questions. There are too many instances where what people thought they would do in a given situation do not match up with their actions once they are faced with the reality of the situation. But these results do suggest that Trump does maintain a core base of high RWA/SDO voters who might take to the ramparts for him and another group of high RWA voters who will tacitly offer their support. The question is whether these two groups form a critical mass in American politics.
Altemeyer’s position is that a certain number of people in any given population will always have authoritarian tendencies. The question is whether those inclinations are validated and authoritarian behaviors are deemed acceptable by a critical mass in society. Such situations seem to require a perfect storm of social uncertainty and economic volatility, but also seem to depend on the willingness of political leaders to cravenly play on those fears – or stand idly by while others do this.
It is possible that much has changed since Altemeyer first developed these scales. Religious fundamentalism as a political force was in its nascent stage four decades ago. There may have been other belief systems that correlated with authoritarianism just as well. Regardless, these two constructs seem to be tightly linked today, at least in American politics.
This does not mean that all evangelicals or all Trump supporters are predisposed to authoritarianism. Nor does it mean that only those on the right of the political spectrum exhibit these tendencies. But the correlations are rather strong. I discuss these exceptions to the rule, as well as consequences for a constitutional republic, in the second part of this post.