The Nature of Affective Polarization: Disentangling Policy Disagreement from Partisan Identity

The forthcoming article “The Nature of Affective Polarization: Disentangling Policy Disagreement from Partisan Identity” by Nicholas Dias and Yphtach Lelkes is summarized by the author(s) below. 

Republicans and Democrats loathe each other now more than ever measured in surveys. This chasm between in- and out-party affect—called affective polarization—encourages citizens to judge politicians and political outcomes in emotional and biased ways, compromising their ability to hold elected leaders accountable. It leads voters to reward politicians that eschew compromise in favor of grandstanding. Worse still, affective polarization’s effects are not limited to politics. It also motivates social segregation along party lines and potentially distorts economic markets.  

But why do Republicans and Democrats hate each other? Proposed causes include a decline in cross-cutting social identities between the parties, or Americans’ increasingly extreme or sorted policy preferences. These accounts assume partisanship influences interpersonal affect via one of two mechanisms: an evolved tendency to dislike social out-groups (partisan identity) or the parties’ disagreements about salient policy issues (policy disagreement). Yet, answering this question has been complicated given that the effects of partisan identity and policy disagreement are often empirically indistinguishable.  

Our study overcomes this methodological challenge to disentangle the effects of partisan identity and policy disagreement. We find partisan identity is the primary mechanism of affective polarization, and that party-branded policies mediate the party-affect relationship by signaling party loyalty. Thus, the party-affect relationship is not spurious, as others have argued. These results suggest that partisanship is not merely a “running tally” of policy preferences and past political experiences. 

We begin with a representative survey (Study 1), which shows the parties have easily distinguishable stances on policy issues used in past studies of affective polarization. This underscores the need to test the effects of policy preferences that do not cue partisan identity. Then, in four experiments, we separate the effects of policy disagreement and partisan identity by having participants evaluate personal vignettes with randomly manipulated features: partisanship, the statement of policy preferences, and what stated preferences signal about party loyalty. 

Study 2 finds party-branded preferences (which signal party loyalty and policy disagreement) nearly erase the party-affect relationship. But salience-matched, unbranded preferences (which only signal disagreement) have a much smaller effect. This gap in policy effects is especially large among strong partisans. Study 3 shows unbranded preferences act as party-branded when randomly tied to a party. Moreover, party-disloyal preferences particularly deplete the party-affect relationship. We replicate this finding (Study 4) and show that party-disloyal preferences do not work via inferences about other policy preferences (Study 5). 

Though policy disagreement’s role in affective polarization has been overestimated, disagreement still matters to affective polarization. Unbranded preferences diminish the effect of partisanship on feeling thermometers by around a third. Also, the total effect of policy preferences—that is, including mechanisms unrelated to partisanship—on thermometers exceeds that of partisanship. However, the total effects of partisanship and policy preferences on social closeness are comparable. 

Altogether, our findings underscore how the effects of partisanship and policy preferences flow through one another. The policy-affect relationship is direct. The party-affect relationship is large, but mediated by party-branded policies. These patterns can give the false impression that policy preferences matter more than partisanship. 

About the Author(s): Nicholas Dias, Ph.D. student in Communication and Political Science, Annenberg School for Communication, Department of Political Science, School of Arts & Sciences at University of Pennsylvania and Yphtach Lelkes, Associate Professor of Communication, Annenberg School for Communication at University of Pennsylvania. Their research “The Nature of Affective Polarization: Disentangling Policy Disagreement from Partisan Identity” is now available in Early View and will appear in a forthcoming issue of the American Journal of Political Science.