Truth Bias and Partisan Bias in Political Deception Detection

Every two years the International Association of Language and Social Psychology selects a Top Paper Award. This year the award was given to David Clementson for  “Truth Bias and Partisan Bias in Political Deception Detection,”

Abstract: This study tests the effects of political partisanship on voters’ perception and detection of deception. Based on social identity theory, in-group members should consider their politician’s message truthful while the opposing out-group would consider the message deceptive. Truth-default theory predicts that a salient in-group would be susceptible to deception from their in-group politician. In an experiment, partisan voters in the United States (N = 618) watched a news interview in which a politician was labeled Democratic or Republican. The politician either answered all the questions or deceptively evaded a question. Results indicated that the truth bias largely prevailed. Voters were more likely to be accurate in their detection when the politician answered and did not dodge. Truthdefault theory appears robust in a political setting, as truth bias holds (as opposed to deception bias). Accuracy in detection also depends on group affiliation. In-groups are accurate when their politician answers, and inaccurate when he dodges. Out-groups are more accurate than in-groups when a politician dodges, but still exhibit truth bias.


Effects of partisan bias on perceptions of evasion in a political news interview

Clementson, D. E., & Xie, T. (Grady PhD student) (2020, Nov.). Effects of partisan bias on perceptions of evasion in a political news interview. Paper to be presented at the 106th National Communication Association conference, Political Communication division, Indianapolis, IN.

Abstract: This paper applies truth-default theory (TDT) to political deception. TDT suggests that people detect deception through suspicion being triggered by a political reporter, which causes a politician to lose credibility. TDT also holds that in-groups have a particularly strong truth-default toward their own members. We report an experiment in which U.S. voters (n = 125 Democrats, and 125 Republicans) watched a real news interview featuring an allegation of evasion. Participants were randomly assigned to one of four conditions: a Fox News journalist accusing (or not accusing) a liberal of deception, or a CNN journalist accusing (or not accusing) a conservative of deception. Consistent with TDT, voters’ distrust arose through elevated suspicion, followed by perceiving deceptive messaging. In-group/out-group bias also drove perceptions of deception. However, moderated multiple mediator modeling indicated voters react non-differentially whether or not their in-group media outlet accused an ideologue of evasion.