Poll Respondents Lack Confidence in False Beliefs

WASHINGTON (August 1, 2022) – Most polls claiming to show that Americans believe in falsehoods should not be treated as a representation of the firmly held beliefs of respondents, according to a new study published in the journal American Political Science Review. Pollsters systematically overestimate the degree to which survey respondents believe in false claims, often leaving the public misinformed about the effects of misinformation.

Matt Graham, now an assistant professor of political science at Temple University, studied the meaning of surveys that claim to measure belief in falsehoods during his time as a postdoctoral research scientist at the George Washington University Institute for Data, Democracy & Politics (IDDP). He conducted surveys covering government budgets, politicized controversies, the economy, science, and the COVID-19 pandemic. Graham discovered that survey respondents who claim to hold false beliefs show considerable response instability as time goes on. Even those who reported total certainty in a falsehood were significantly less likely to express that belief as confidently in follow-up surveys. The results imply that a typical person who shares a false belief with a pollster is usually just guessing and may just believe the falsehood is plausible because of previously held beliefs.

The study suggests that pollsters and journalists may be overly eager to interpret survey results as evidence of extreme political polarization. Rather than interpret survey responses as value statements founded in conviction, they should be viewed as spur-of-the-moment thoughts from people who possess insufficient information. Most people who endorse false beliefs in polls feel a substantial degree of doubt about their replies. 

“It’s not that the person on your Facebook feed with wild and crazy beliefs isn’t real,” Graham said. “It’s just that those people are way less representative of the average person than coverage of public opinion polls tends to claim.” 

The study was partially funded by IDDP, which launched in 2019 with the support of the John S. and James L. Knight Foundation. IDDP’s mission is to help the public, journalists and policymakers understand digital media’s influence on public dialogue and opinion, and to develop sound solutions to disinformation and other ills that arise in these spaces.

-GW-