Article Summary
Introduction
Much of the literature on affective polarization is motivated by the idea that partisan hostility leads to undesirable political outcomes: those who are more affectively polarized are less likely to uphold important norms of democratic systems, hold their representatives less accountable, undermine bipartisanship, and so on. These ideas are almost entirely speculative, however. Despite the abundance of concern about affective polarization’s political consequences, there is very little direct evidence to substantiate this concern. These authors focus on exploring the supposed political consequences of affective polarization. They argue that people’s general like and dislike of the parties should not influence political choices. When people vote or engage in political action, there are many factors that are likely more relevant than their abstract feelings about the parties—how the economy is doing, how much they agree with a candidate, whether they historically vote for one party, etc. Thus, affective polarization should be mostly unrelated to concrete political judgments.
However, researchers often do not measure concrete political judgments in surveys. Instead, they frequently rely on abstract survey questions about what respondents might imagine doing in a hypothetical situation. Because these kinds of questions have no context (and, thus, none of the other factors that typically influence decision-making), affective polarization may shape how respondents answer them. These authors accordingly compare concrete questions with abstract questions to examine whether abstract questions might mistakenly lead to the impression that affective polarization shapes political judgments.
Analytical Approach
To explore the political consequences of affective polarization, the authors conduct five surveys. In each survey, they experimentally manipulate affective polarization and measure its various potential consequences. The first four surveys all use the same manipulation. The authors lead participants to believe that a member of the other party appears to trust them (or not) based primarily on their partisanship, generating a positive (or negative) experience with a member of the out-party.
This kind of manipulation is called a ‘trust game.’ Typically, trust games involve two players. Player 1 is given a set amount of money and asked to send some portion of it to Player 2. The amount sent to Player 2 will be tripled, and Player 2 then decides to send some portion of that money back to Player 1. In short, players who trust each other and engage in reciprocity will gain the most money. For this study, all participants were assigned to the Player 2 position, and Player 1’s behavior was randomly assigned to be either trusting (wherein Player 1 sent Player 2 a large portion of the initial money) or not trusting (wherein Player 1 sent Player 2 none of the initial money). All participants played three rounds of the game with the same conditions: participants assigned to the positive experience condition had three positive experiences, and participants assigned to the negative experience condition had three negative experiences.
Across the first four surveys, authors explore how this manipulation affected four main sets of outcomes: general social distance, support for democratic norms, perceptions of objective conditions, and electoral accountability. To measure social distance, they asked about participants’ abstract concerns about hypothetical interactions with members of the other party. To measure support for democratic norms, they asked about participants’ support for violating these norms, their support for partisan violence, and their preferred outcome in several specific vignettes illustrating situations in which opportunities benefiting the in-party might require violating norms. To measure perceptions of objective conditions, they reminded all participants that Trump had been president for the last four years, then asked them to estimate the unemployment rate and how many people had died from COVID-19.
The measures of electoral accountability rely on additional manipulations beyond the trust game. These manipulations were implemented in surveys 1 and 3. Participants were asked their positions on several issues. After that, they were shown how their real member of congress voted on three issues; participants were randomly assigned to see three votes they agreed with, three votes they disagreed with, or no voting information (as a control). The table below summarizes the treatment conditions, combining the trust game manipulation with the vote information manipulation. After both manipulations, participants were asked about their approval of the member of congress and their own issue positions again to examine whether participants adopted their member of congress’ positions.
| | Affective Polarization Manipulation | ||
| Positive experience with out-party | Negative experience with out-party | ||
| Vote Information Manipulation | Votes participant agrees with | 1 | 2 |
| Votes participant disagrees with | 3 | 4 | |
| No vote information | 5 | 6 | |
In survey 2, the authors implemented an additional manipulation to explore whether affective polarization affects support for bipartisanship. After the trust game, participants read about a real in-party member of congress, and they were randomly assigned to read about either that member’s party-line votes or their bipartisan votes. The authors then measured participants’ approval toward this member of congress.
| | Affective Polarization Manipulation | ||
| Positive experience with out-party | Negative experience with out-party | ||
| Congressional Bipartisanship Manipulation | Party-line votes | 1 | 2 |
| Bipartisan votes | 3 | 4 | |
Finally, survey 5 explored whether affective polarization could be manipulated with a different kind of experimental intervention to ensure that the other surveys’ results were not driven by the unique trust game manipulation. Here, the authors used a combination of depolarizing treatments from past research, correcting misperceptions of out-party demographic composition, correcting misperceptions about out-party political views, and asking people to reflect on a friendship with a member of the out-party. Participants either received all three interventions or none of them. In this survey, the authors focused on potential effects on social distance measures and support for democratic norms.
| Affective Polarization Manipulation | |
| Misperception corrections + reflection exercise | Control |
| 1 | 2 |
Main Findings
The authors find that the trust game manipulation substantially alters affective polarization. In particular, the positive experience condition dramatically improves feelings toward members of the out-party. This treatment reduces participants’ abstract discomfort at the prospect of interacting with members of the out-party. However, it did not affect support for democratic norms, perceptions of objective conditions, loyalty toward in-party legislators, responses to in-party legislators’ votes, people’s likelihood of adopting their legislators’ issue positions, or support for bipartisanship.
In sum, affective polarization does not lead to undesirable political outcomes, at least with regard to democratic norms and electoral accountability.
Implications
Despite widespread alarm about the political consequences of affective polarization, this research shows that there is no empirical evidence that affective polarization threatens democracy. At least, not yet! If researchers, practitioners, and legislators are interested in ‘solving’ the growing problems in democratic societies, simply reducing affective polarization may not be the answer. Accordingly, people concerned about these problems may be better served by refocusing on other potential causes of democratic ills. The authors highlight the importance of examining these questions with experiments. Much of the previous speculation about affective polarization’s supposedly negative consequences was based on observational, correlational data. However, without direct causal evidence from experiments, scholars could not conclude that affective polarization causes anything, and should be more cautious about doing so in future work. Moreover, the study shows that researchers need to be attentive to the ways they measure political outcomes with survey questions. The importance of affective polarization is likely to be inflated if researchers use only general, abstract questions; to more accurately simulate real political decision-making, researchers need to incorporate context and trade-offs into survey measures.
Questions left unanswered
The authors note that their manipulation was primarily successful in reducing affective polarization, not in increasing it. It is possible that studies aimed at increasing partisan hostility may reveal new consequences this study did not. Moreover, this study only successfully manipulated affect toward the out-party. The consequences of affect toward the in-party have not been clearly theorized or tested, and that may be a productive avenue for future research.
The authors also note that they focused on immediate outcomes of affective polarization. Do the effects of affective polarization accumulate over time, or do they perhaps have indirect effects on other variables (e.g., cynicism, trust in government) that these authors did not examine?
Methods and Analysis
Was the study and its analyses pre-registered?: Yes
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
The study does not use samples that are strictly representative of the U.S. population. As such, it is possible that some of the results may not generalize to the U.S. population at large. However, the authors show that the sample compositions are broadly similar to the U.S. population, so there is no real basis for specific, major concerns about non-generalizability. As part of their sampling strategy, the authors explicitly omit ‘true independents’—people who do not lean to one party or the other. These people are a small fraction (about 10%) of the U.S. population, but they are politically consequential nonetheless. What does affective polarization look like for these people? Is independents’ hostility toward the parties related to anti-democratic attitudes? These questions remain unanswered in this study and in the broader literature.
Open Data & Analyses
Does the article make the replication data publicly available?: Yes
Does the article make the replication analysis scripts publicly available?: Yes
Article Citation
Broockman, D. E., Kalla, J. L., & Westwood, S. J. (2023). Does affective polarization undermine democratic norms or accountability? Maybe not. American Journal of Political Science, 67(3), 808–828. https://doi.org/10.1111/ajps.12719
Bibtex
1
2
3
4
5
6
7
8
9
10
11
12
13
@article{https://doi.org/10.1111/ajps.12719,
author = {Broockman, David E. and Kalla, Joshua L. and Westwood, Sean J.},
title = {Does Affective Polarization Undermine Democratic Norms or Accountability? Maybe Not},
journal = {American Journal of Political Science},
volume = {67},
number = {3},
pages = {808-828},
doi = {https://doi.org/10.1111/ajps.12719},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1111/ajps.12719},
eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1111/ajps.12719},
year = {2023}
}