The Different Faces of Public Opinion: Is the American Voter Tinted by Mode?

Abstract

Declining response rates and increasing costs of collecting public opinion survey data have led to an increasing usage of dual-mode surveys. Revisiting foundational theories of political knowledge, issue constraint, and issue voting, we gauge the theoretical implications this methodological change has in our understanding of the public opinion and voter behavior. We see different patterns in the 2012 and 2016 American National Election Studies and observe clear mode effects in an original dual-mode survey experiment of Florida registered voters. Overall, we find that web respondents appear to be more politically knowledgeable, ideologically constrained, and have greater correspondence between issue and vote preferences compared to face-to-face or phone respondents. Survey mode matters, even with a common sampling frame.