Preference Falsification in Authoritarian Regimes: Evidence from Russia

• In Event: Public Opinion Under Autocracy

Abstract

Misrepresentation of preferences among those responding to public opinion surveys is a fundamental obstacle in public opinion research under autocracy. Respondents in authoritarian settings have a number of potential reasons to misrepresent their support for an autocrat (Shen and Truex 2020). We fit these reasons into three broad categories: concern about personal retribution and repression (what we refer to as ‘political desirability’), a general desire to conform to positions and views that are widely held in society, and social desirability effects vis-à-vis survey interviewers. The latter two potential sources of bias are presumably found in survey research in many contexts, while political desirability, stemming from concerns, or fear of repression, is commonly seen as specific to authoritarian regimes. Understanding which ‘type’ of preference falsification is at work is important for how we interpret survey data in authoritarian politics. For example, if many respondents report supporting an autocrat simply because they think that is what the interviewer wants to hear, falsification might be less widespread than if respondents fear for their safety upon giving the “wrong” response in an interview. Perhaps more importantly, distinguishing between these reasons for preference falsification also matters for how regimes generate an image of strength and invincibility. Regimes can enforce perceptions about "acceptable" views and norms through socialization, social interactions, schooling, and the media. If social rather than political desirability encourages individuals to misreport ambivalent, or hostile attitudes toward the ruling regime stem from individuals who are willing to misreport their views in order to conform with the acceptable views in society, then perhaps regimes do not need to invest so much in spreading fear in order to inflate their reported levels of support. These reasons for preference falsification also matter for how regimes break down: if social conformity is driving falsification, regime support might evaporate suddenly, even as the regime’s (perceived) coercive capacity does not change.

In this paper, we investigate these different forms of desirability by combining priming and list experiments. With the priming treatments, we investigate the degree to which support for an authoritarian president, Vladimir Putin of Russia, is contingent upon 1) (views about) the degree of confidentiality of the survey and 2) perceptions of his general popularity in society. The first prime allows us to investigate whether being primed to think about one’s personal security in the face of potential regime retribution affects reported support for the president (political desirability). The second is a set of two primes, one signaling that the president's popularity levels are lower than the respondent might otherwise believe, and the other that they are higher. This set of primes allow us to investigate the degree to which the president's popularity is contingent upon perceptions of general trends, and whether or not the effect of these perceptions is unidirectional (i.e. whether this effect is entirely the result of learning the surprising information that the president is less popular than one thought). Finally, we incorporate list experiments as our measurement strategy when capturing outcomes from the priming experiments. List experiments allow us to estimate the degree to which these priming effects reflect actual changes in support for the president by eliminating traditional social desirability effects relative to the interviewer. To the best of our knowledge, this is the first survey to combine survey experiments with list experiments.

Authors

• Ora John Reuter, University of Wisconsin, Milwaukee

• Katerina Tertytchnaya, University College London

• Noah Buckley, Higher School of Economics; Trinity College Dublin

• Kyle L. Marquardt, University of Bergen