Typecast: Roots of Misperceptions About Partisan Composition

• In Event: The Psychology of Partisanship

Abstract

Both Democrats and Republicans dislike, distrust, and discriminate against their out-party counterparts (Iyengar et al. 2012; Iyengar and Westwood 2015; McConnell et al. 2016). Party stereotyping is a source of this animus: Americans vastly overestimate the share of party-stereotypical groups in the parties, and clearing up these misperceptions reduces animus toward the opposing party (Ahler and Sood 2017).

But what causes party stereotyping? Why do people overestimate the shares of party-stereotypical groups in the mass parties? For example, why do Americans think that a third of Democrats are gay and a third of Republicans belong to the economic one percent? We shed light on the roots of these perceptual biases in this paper.

People think about political parties in terms of longer-standing social groups associated with them (Green, Palmquist and Schickler 2002). Groups become associated with parties not because they are most heavily represented in a party, but rather because they are most heavily sorted into one party. This is because people think of parties in terms of their most distinguishing traits—that is, they use the representativeness heuristic. This bias, in turn, causes them to overestimate the share of “representative'” groups in the parties. For example, just 4% of Americans identify as LGBT, but people think the share of Democrats who are LGBT is nearly 33%.

We begin by presenting data that show that people see heavily sorted groups as party-stereotypical, ignoring the fact some of these groups constitute only a small fraction of the population and thus cannot constitute a large share of the party. Next, we provide experimental evidence that representativeness is one of the underlying psychological mechanisms behind this cognitive bias. We assign survey respondents to one of three conditions: one in which they are informed of, and made to remember party-stereotypical groups' base rates (e.g., 4% in the example above), one in which they are informed of and made to remember those groups' “party rates” (e.g., 63% of LGBT Americans being Democrats), or a control condition. Compared to the control condition, participants in the “base rates” condition become only slightly less biased in their perceptions of party-stereotypical groups’ shares. By contrast, participants' perceptions become significantly more biased in the “party rates” condition, especially for groups that sort heavily into one party—consistent with a representativeness-based explanation for party stereotyping. We conclude by discussing the implications of these findings for partisan relations in the era of President Trump and a new identity politics.

Authors

• Douglas Ahler, Vanderbilt University & Florida State University

• Gaurav Sood, Independent