Always Silent? Voter Intention Non-response in the 2020 US Presidential Election

• In Event: Challenges of Forecasting with Polls

Abstract

Non-responses to vote choice questions notoriously affect the quality of electoral predictions. In the US, this issue has gained particular visibility in the aftermath of the 2020 Presidential Elections. Indeed, the failure of many major pollsters in predicting election results in several key states stimulated a renewed attention for the so-called "Shy Trump Supporter" hypothesis, according to which Trump supporters would be more likely to hide their vote preference in electoral surveys due to social desirability bias. Although empirical research on this topic is methodologically diverse, it tends to focus mainly on individual-level data, overlooking the role that the actual socio-political environments of the respondents could play in the decision to disclose (or not) one's own political preferences.

This paper aims at contributing to the voting behavior prediction literature by testing two possible mechanisms behind non-responses to presidential voting questions in the US: social desirability bias, due to holding minority political preferences at the local level, and exposure to cross-cutting political pressures in one's own local context. Hypotheses are tested by means of logistic multilevel regressions on pre-electoral data from the ANES 2020 Exploratory Testing Survey, matched with prior election results (i.e., 2018 mid-term elections) at the county level, which are used to reconstruct the political composition of respondents' local environments (source: MIT Election Data and Science Lab).

Authors

• Lewis Alexander Luartz, University of California, Riverside

• Stefano Camatarri, Waseda University

• Marta Gallina, Waseda University