The debate over why so many white, middle-class Americans found the message of Donald Trump appealing has continued vigorously since the election, with economic anxiety the leading candidate among those most charitable toward Trump voters and racism a central reason among those less so.
Neither explanation appears to capture the full range of causes, which may be better thought of as a cocktail of internal and external factors that leaves a voter open to authoritarian appeal. Understanding the ingredients is urgent for reversing the lure of authoritarian leadership. As a research psychologist, I have been investigating this question with colleagues at psyML, a research firm focused on understanding the psychological states of people acting online.
Economic anxiety is inextricably connected, today, to the appeal of authoritarianism. In August before the election, I wrote a piece arguing that the movement of white, middle-class Americans toward Trump was the result of dysfunctional adjustments to financial stress. In research on over 6,000 Americans facing financial hardships, we consistently saw a profile similar to that of PTSD: nightmares, an inability to feel emotionally close, and cognitive blunting. Additionally, we found an association between high levels of external locus of control (ELOC) — the idea that success or failure is a result of external factors — and the intensity of financial stress. We hypothesized that high levels of ELOC are present in those who are drawn to authoritarian thinking and that faith in oneself may well be the antidote.
Consistent with this finding, the political scientist, Matthew MacWilliams, reported that a receptivity to authoritarianism was the clearest predictor of those who supported Trump in a national poll prior to the election.
But it is highly unlikely that there is a single variable that determines behavior as complex as a vote for president. While there has been compelling research tying attitudes towards race to the appeal of Trump, I do not consider these explanations sufficient.
Sean McElwee and Jason McDaniel, based on analyses of data provided by the 2016 American National Elections Studies (ANES), concluded that economic stress is not the most relevant factor in the election of Trump. They used regression analyses to consider the effects of symbolic racism (the belief that generations of discrimination has no relevance), the impact of African-American influence animosity (the perception of favoritism by the government for African-American people), and a scale assessing how someone determines the value of immigrants. When they adjusted their results relating to economic stress to account for racial resentment, or attitudes towards African-American influence or immigration, it reduced the predictive effect of economic stress on voting behavior to below a statistically significant level. As a result, the authors conclude “individual personal hardship” is not relevant to voting choice in 2016.

Photo courtesy of Shutterstock.

Thomas Wood of the Washington Post analyzed the same data with a focus on symbolic racism and authoritarianism. Using regression analyses the author found the effect of racism larger than authoritarianism and inferred that racism should be the focus of efforts to understand the Trump phenomenon. The assumption in both of these studies is that the strongest unique predictor is the single most meaningful variable.
In addition to our previous work on financial stress and externalization, there are other reasons to explore the role of financial stress, and chronic stress in general, in how individuals perceive authoritarian appeals. First is that financial hardship has been widely linked to the success of far-right candidates around the world. Another widely reported finding from the 2016 election that supports a role for stress is the fact that diabetes, lack of exercise, heavy drinking, and high body weight were found to be overrepresented among those who had swung toward Trump. Chronic stress is a precursor for each of these and the wear and tear on the body as a result of stress is applied to our understanding of the development of a sympathy towards authoritarianism.
Our team’s hypothesis is that there is a unique psychological profile that distinguishes white Trump voters. This profile is exacerbated by exposure to chronic stress, often financial in nature. There are many psychologically unhealthy themes contained in Trump’s message: racism, xenophobia, misogyny, anti-intellectualism and a lack of openness in general. Underlying many of these is a persistent disavowal of a consensus definition of truth in favor of this one man’s opinion. Historically, the widespread acceptance of obvious untruths is part and parcel with an acceptance of authoritarianism consistent with a primary component of authoritarianism — submission.
While our arguments around financial stress and psychological characteristics have grown from our research findings, Bruce McEwen, a molecular biologist from Rockefeller University, has developed a complex network model, called allostatic load, that is applicable to the development of this psychological state. McEwen identified a process where stress interacts with each person’s baseline physiological and psychological vulnerabilities to produce a cascading effect, with the result being increasing dysfunction in individual areas of vulnerability. While the human physiological system is incredibly adaptive to short periods of intense stress, our system is completely unprepared for chronic stress, stress that keeps our system from returning to physiological and psychological stasis.
Unresolved stress may well be one of the defining characteristics of western civilization. It is pervasive, along with its related posse of physical diseases — diabetes, heart disease, and autoimmune disorders, combined with equally deadly psychological dysfunctions that can lead to drug and alcohol abuse and even suicide.
The question is whether the psychological distress that develops with chronic stress also threatens our ability to engage in democracy. Specifically, when considering the core variables of the two studies described above (racism, financial stress, and authoritarianism) we hypothesize that the best causal model, taking into consideration the concept of allostatic load, is that both racism and financial stress are contributors to the development of a sympathy towards authoritarianism. We further hypothesize that openness to authoritarianism is the strongest direct predictor of Trump support.
The path from vulnerabilities to an authoritarian mindset is exacerbated not only by financial stress but stress of other varieties, including the stress associated with the perceived loss of entitlements and status. We have already noted that externalization of control, or a belief that you are controlled by powerful others, may also be a psychological factor that emerges on the path to authoritarianism, with increasing intolerance and openness to conspiracies also part of its development.
Using the core variables introduced by the above studies, in our own work we employed structural equation models to determine if the pattern of associations we hypothesize among these variables is the best predictive model of voting behavior. No measure of ELOC is available. Our broad model is that racism and stress contribute to an authoritarian mindset, which in turn will predict voting behavior. We do not include control variables used in other studies given a lack of reported effects.
We provide a model for whites who voted for either Trump or Clinton. As hypothesized, higher levels of financial stress and racism were associated with significantly higher scores on authoritarianism, while racism also has a significant direct effect on voting behavior. Authoritarianism measures also had a direct and significant effect on voting, with greater tendency towards racism and pro-authoritarianism associated with a vote for Trump.
While we find racism and financial stress to be relevant, it is noteworthy that in addition to this model, we ran thousands of regressions and path models including some related to personality and attitudes towards egalitarianism and traditionalism. There is an amazing amount of intercorrelation among all of these variables, and the current model must  be considered preliminary.
This model does provide a number of possible opportunities for intervention, for those interested in changing underlying attitudes and political orientation. Cognitive behavioral therapy methods have been demonstrated to be useful in minimizing the effects of financial stress and should be explored as a method to derail the development of an interest in authoritarianism. Motivational interviewing techniques have also been shown to be effective in enhancing self-esteem and internal locus of control.
We need to be working across disciplines and institutions to better understand why more and more people are open to an authoritarian message. We need to be collecting model-driven data now if we hope to understand why Americans, and the world, are moving toward an authoritarian mindset. And, we need to do all we can to help Americans build the self-esteem, openness, optimism, and sense of cooperation needed to be a full participant in democracy.

Galen Buckwalter is a research psychologist and chief science officer at psyML. He attended the Techonomy 2016 and Techonomy 2017 conferences. This research was conducted by a team at psyML.