No strong evidence that authoritarian attitudes are driven by a lack of control

Kakkar and Sivanathan (1) provide new evidence of the link between economic crises and the preference for dominant leaders and strengthen previous evidence of the influence of economic crises on authoritarianism (2–5). However, the authors’ interpretation of why economic uncertainty favors dominant leaders entails two further predictions that are not borne out by our follow-up analyses of their dataset.

According to the authors, individuals choose more dominant leaders in times of uncertainty “to reduce the aversive state of low personal control when plagued by collective uncertainty”; in other words, “seek[ing] a dominant leader [is] a compensatory strategy to restore their sense of personal control” (1). This hypothesis predicts that people should prefer dominant leaders compared with prestigious leaders but also compared with having no leader at all. However, while study 1 reveals that economic uncertainty increases the preference for dominant compared with prestigious leaders, it has no impact on the probability of choosing a dominant leader compared with no leader at all (b = 0.24, P > 0.250). It thus appears that economic uncertainty only influences the type of leader that is preferred and does not trigger a compensatory strategy that would lead people to seek dominant leaders in general.

Another important point in Kakkar and Sivanathan’s demonstration is the predictive power of the lack of control in leader preferences (1). However, our follow-up analysis of their data does not provide strong evidence for their claim. Running a Bayes factor analysis on the World Value Survey data indeed fails to reveal evidence in favor of the model including sense of control compared with the model with control variables only or to those also including change in unemployment (Bayes factors < 1.19; Fig. 1). This analysis nonetheless confirmed the explanatory power of economic uncertainty as the model including change in unemployment was substantially more probable than the one that only included control variables (Bayes factor = 1.71e8). These additional analyses thus cast doubts on the role played by the sense of control in explaining the preference for strong leaders in times of economic uncertainty.

On the contrary, an increased perceived adaptiveness of dominant features, such as untrustworthiness, in uncertain contexts may provide a better explanation of the effect reported by Kakkar and Sivanathan (1). For instance, including interpersonal distrust increases by more than 10,000 times the probability of the model with the control variables only and that of the model including economic uncertainty (Bayes factor > 1.00e26). Importantly, further analyses confirmed that interpersonal distrust mediated the effect of economic uncertainty on leader preferences (b = 0.002, P < 0.001; proportion mediated: 9%).

Based on these new analyses, it thus appears that economic uncertainty influences individuals’ preferred type of leader by shaping their perception of the environment (3, 6) rather than by increasing the absolute need of a strong leader who would contribute to restoring people’s decreased sense of control.