Abstract
This article explores people’s susceptibility to political deception. Participants watched a news interview in which a politician either answered all the questions or deceptively evaded a question. In Study 1 (n = 202), deception is perceived through the dodge being irrelevant for voters who do not identify with the politician. In Study 2 (n = 618), partisan voters consider the politician more deceptive, and acting more deceptively, when the politician has their opposing party affiliation, independent of whether he dodges. When the politician shares their party identification, voters identify with the politician more and consider his responses more relevant. Findings are consistent with theoretical positions of identification, the cooperative principle, and social identity.
Notes
1. Power analysis for an estimated effect size of at least .3 (which is a moderate effect size according to Cohen, Citation1988) in regression (f2) with three predictors (one of which, the independent variable, is dichotomous), an alpha (α) level set at .05, and statistical power (implied β) of .8 (denoting an 80% likelihood of detecting an effect, as recommended by Cohen, Citation1988) requires a minimum sample size of 111.
2. The “partial completes” and/or “terminates” in the Qualtrics panel were not recorded. However, an estimated 32% qualification rate was recorded. Thus, we can estimate that roughly 400 people were terminated by Qualtrics.
3. Power analysis for an estimated effect size of at least .3 in regression with four predictors (two of which, the independent variable and a moderator, are dichotomous), an α-level set at .05, and statistical power of .8 requires a minimum sample size of 260.
4. “Partials” and “terminates” were screened out. Here is the breakdown of respondents who failed an attention check: 230 respondents were filtered out in an opening questionnaire block after the consent form, 65 were filtered out in a later attention check, and 23 were filtered out from a third attention check. Only “good completes” from the Qualtrics panel were reported to the researchers for the dataset, resulting in the 618 participants.
5. Here is the breakdown of those filtered out who failed the manipulation check, based on their treatment condition: 103 randomly assigned to the Democratic (D) politician Dodge condition failed, 112 randomly assigned to the D No Dodge condition failed, 103 in the Republican (R) politician Dodge condition failed, and 88 in the R No Dodge failed.