Article Summary
Introduction
Partisanship is an incredibly influential identity - shaping vote choice, information processing, and feelings toward politically dissimilar others. One explanation for the strength of partisan identity is that parties are understood as collections of other, long-standing groups (i.e. sorted by race, class, sexual orientation, etc), and that therefore people’s attitudes about those constituent groups shape their attitudes about political parties. However, a puzzle for this theory is that the demographic composition of the two major parties in the U.S. is quite similar – majority White, middle class, and heterosexual. This paper describes the extent to which Americans misperceive party composition in the U.S., and why that matters.
Analytical Approach
Study 1 surveyed 1,000 Americans through YouGov in order to acquire descriptive data on the perceived percentage within both political parties of stereotypically constituent groups. Respondents were asked to estimate the percentage of Democrats who are Black, atheist or agnostic, union members, and gay, lesbian, or bisexual; they also estimated the percentage of Republicans who are evangelical, 65 or older, southern, and earning over $250,000 annually.
The second study addressed three alternative explanations for overestimating constituent group size: expressive responding, innumeracy, and ignorance of base rates. There were four experimental conditions – the standard condition simply asked the same estimation questions as in Study 1. The expressive responding condition addressed the concern that participants were exaggerating the party percentage of constituent groups in order to express a dislike for the party by providing a financial incentive for correct estimates. The innumeracy condition asked participants to make estimates for the percentage of multiple groups in a category (e.g. “What percentage of Democrats do you think are: Black? White? Latino? Other?”) and required that responses sum to 100%. The base rates condition addressed ignorance of base rates (the prevalence of a particular group in the population at large) by starting sliders at the true population base rate for each group and telling participants to adjust it to reflect the group’s percent representation in the party.
Turning to the consequences of misperceptions of the out-party, they administered an additional observational survey which measured perceptions of constituent group sizes, beliefs about outpartisans’ policy preferences, and allegiance to one’s own party. Finally, an experiment probed the effects on partisan animus and social distancing of three conditions: Ask (same estimates of group sizes within parties as previous studies), Tell (correct participants overestimates of constituent group size before proceeding to the rest of the DVs), and Control (participants make group size estimates after the other DVs).
Main Findings
The first key insight of this paper is that people strongly overestimate the percentage of both parties that is composed of stereotypical constituent groups. For example, participants estimated that 38.2% of Republicans earned over $250,000 annually (only 2.2% of Republicans earn that much) and that 31.7% of Democrats were LGB (only 6.3% are).
Second, they experimentally tested alternative explanations. Participants given financial incentives to respond correctly were equally biased in their estimates as those in the control condition, providing evidence against expressive responding. Similarly, the guide rails in the innumeracy condition did not strongly affect estimates of party composition, suggesting that inflated estimates were not pure innumeracy. Interestingly, providing base rates of group prevalence in the population made participants even less likely in their estimates of party composition. Ultimately, there was no strong evidence for any of the proposed alternative explanations.
Last, this paper examined the impact of misunderstanding who makes up the opposing party. In the observational survey, greater overestimates of the size of stereotypical constituent groups were associated with perceiving opposing partisans as holding more extreme political positions. In the experiment, correcting overestimates of the party share of stereotypic constituent groups resulted in more accurate understanding of party composition, decreased perceptions of the opposing party as ideologically extreme, decreased affective polarization, and decreased social distancing from outpartisans. Generally, the correcting misperceptions intervention had a stronger effect on participants whose initial estimates were more inflated.
Implications
This research provides strong evidence that people vastly overestimate the percentage of each party composed of stereotypical groups. It argues that because people have strong attitudes toward those constituent groups already, overestimating their proportion in the opposing party leads to greater dislike of outpartisans, as well as perceiving them as more ideologically extreme. This helps explain how high levels of polarization can coexist alongside two political parties that are in many ways demographically similar to each other.
Questions left unanswered
This paper provides strong evidence that people misperceive the composition of the opposing party, and that these misperceptions are tied to perceiving the opposing party as more extreme and disliking them more. One lane of future investigation would be to put the fact that people strongly overestimate the prevalence of constituent groups in dialogue with the fact that there are significant differences in the demographic composition of the Republican and Democratic parties. Meaning, while fewer Democrats are LGB than Republicans believe, significantly more Democrats than Republicans are LGB. This is in contrast to overestimates of the income of Republicans, where high income earners are more evenly distributed between the two parties. Further probing how people understand the relative prevalence of social groups between the two parties might provide additional insight.
Methods and Analysis
Was the study and its analyses pre-registered?: No
Did the study rely on proxy variables to measure polarization?: No
Were standard p-value thresholds used (p<.05 or 95% Confidence Intervals that don’t overlap zero)?: Yes
- Largest p-value presented as significant: 0.05
Were correlational results interpreted with causal language?: No
Limitations / Weaknesses
This paper replicates its central finding that people overestimate the percentage of the opposing party composed of stereotypical groups numerous times and does an excellent job ruling out alternative explanations. However, there is some ambiguity regarding causality - finding that correcting misperceptions decreases partisan animus does not necessarily show that those misperceptions cause partisan animus to begin with. Additionally, it is possible that providing a small financial incentive would not be sufficient to deter expressive responding within the highly polarized domain of politics.
Open Data & Analyses
Does the article make the replication data publicly available?: No
Does the article make the replication analysis scripts publicly available?: No
Article Citation
Ahler, D. J., & Sood, G. (2018). The parties in our heads: Misperceptions about party composition and their consequences. The Journal of Politics, 80(3), 964-981.
Bibtex
1
2
3
4
5
6
7
8
9
10
@article{ahler2018parties,
title={The parties in our heads: Misperceptions about party composition and their consequences},
author={Ahler, Douglas J and Sood, Gaurav},
journal={The Journal of Politics},
volume={80},
number={3},
pages={964--981},
year={2018},
publisher={University of Chicago Press Chicago, IL}
}