Are the Populist Attitudes Stable? Attitude Network Modeling with a Quasi-Representative Sample

Ideational approach studying the demand-side of populism suggests that populist attitudes are latent dispositions, which do not fluctuate. These dispositions, however, need to be activated, for example by populist rhetoric. Only after such activation, these attitudes become predictors of various outcomes like voting behavior or support of a populist movement. Despite growing experimental evidence, this proposition has not been tested rigorously. We adopt a novel new methodological approach to test the stability of the populist attitudes in a mixed-design (within and between subjects) experiment. Using a quasi-representative sample of German voters, we measured populist attitudes in two times during our study – before and after experimental treatments. In two conditions, we treated the participants with people-centric and anti-elitist slogans comparing these conditions to a control group. As an analysis strategy, we adopt attitude network modeling, which an emerging powerful technique in psychometrics. This method allows us to make formal comparisons on attitude strength and structure, which we exploit for our experimental analyses. Another advantage of this method is that it allows researchers to gain more insight into the nature of an attitude by simulations from the observed data. Results reveal that the populist attitude strength of males is more powerful compared to females, but the attitude structure did not significantly differ across gender. With regards to (self-reported) left-right political orientation, we find no significant differences in attitude strength or structure. Furthermore, the strength and structure of populist attitudes did not change significantly as a function of experimental manipulation. Simulations from the observed data reveal that populist attitudes have a continuous nature rather than a categorical one. This finding suggests that there are different degrees of being a populist on the demand-side; we cannot think of categories of populist vs. not populist individuals. Overall, we find evidence supporting the proposition of the ideational approach in populism studies. Attitudes on the demand-side of populism are stable, but we also discover that despite the widespread belief, it would be misleading to think of distinct categories of voters who are either populist or not populist. We discuss the implications of our findings for populism studies in general and suggest further research, especially with regards to the activation of populist attitudes leading to specific political behavior. Additionally, to the best of our knowledge, our study is the first to utilize attitude network modeling in an experimental design. We think this promising novel technique is powerful and flexible - we recommend the application of this method to political psychologists.