Polling Brands and Partisan Non-response

• In Event: Methodological Innovations in the Study of Public Opinion

Abstract

The 2016 and 2020 election cycles threw pollsters and political observers for a loop. Polling error generated scores of headlines calling for the end of polling, creating a major PR headache for the industry. It was bad enough to warrant serious postmortem analyses. In the aftermath of 2016, a common theory among pollsters was that partisan nonresponse from Republicans — Trump supporters in particular — explained a significant portion of the polling error. They reasoned that because Republicans’ report low levels of media trust and relatively low trust in institutions generally, they were less likely to respond to surveys than Democrats. Clinton, Lapinski and Trussler (2022) found partisan nonresponse explained a significant portion of the polling error in 2020. Republicans and Independents were less likely to respond to surveys than Democrats. However, partisan nonresponse fails to account for all the polling error observed in 2020, leaving much unexplained. Using a survey experiment on the 2023 CES, we explore whether partisan non-response based on polling organization is more nuanced, where partisans on both sides are selective in the types of polling organization to which they want to respond. 

Authors

• Johanna Dunaway, Syracuse University

• Spencer Goidel, Auburn University

• Joshua P. Darr, Syracuse University