Article Summary
Introduction
What drives emotional investment in politics? What leads people to participate in the political process? Why do many voters care intensely about who wins the next election? Political scientists have long recognized party identification as a crucial variable in these processes, but they disagree about what precisely party ID represents.
On one hand, partisanship may be a product of policy preferences. If people understand the differences in how each party will govern, perhaps their electoral hopes and fears are an extension of preferring one party’s policies — or rejecting the other’s. To care about one’s party, in other words, is to care about the policies that this party will enact when in power. Scholars refer to this as the “instrumental” partisanship perspective. On the other hand, perhaps politics matters to voters because it becomes a part of their identities. By influencing partisans’ sense of who they are, parties engender a sense of belonging and lead voters to connect their self-esteem to the relative status of their political tribe. Such is the theory of “expressive” partisanship, which de-emphasizes voters’ policy views. Instead, support for a political party operates similarly to sports fandom: Once someone identifies with a team, victory simply makes them feel good, and defeat makes them feel bad.
Huddy, Mason, and Aarøe (2015) test these competing explanations. They employ four surveys — and two survey experiments — to gauge whether political emotions and activity are more associated with strong policy stances (instrumental partisanship) or with strong partisan identity (expressive partisanship).
Analytical Approach
The empirical test of expressive partisanship compares the correlation between expressive partisanship and various outcomes and instrumental partisanship and various outcomes. This paper analyzes four datasets: an online survey of political blog readers conducted in 2008 (Blog Study), a New York telephone survey fielded in 2011 (NY Study), a November 2011 wave of YouGov’s national panel survey (YouGov Study), and a 2010 survey of undergraduate students (Student Study).
Instrumental partisanship is represented by ideological issue intensity, which combines ideological consistency with opinion strength. Each study asks respondents 3-5 specific policy questions, and the authors then combine each subject’s responses to gauge a) how consistently conservative or liberal the respondents are, and b) the strength of their policy opinions (see Table 1 for a list of policies).
Table 1
| Study | Policy Topics |
|---|---|
| Student Study | Health care Same-sex marriage Reduce government spending by providing fewer services |
| NY Study | Health care Same-se marriage Increase taxes or cut spending |
| YouGov Study | Immigration Health care Abortion Same-sex marriage Reduce budget deficit or reduce unemployment |
| Student Study | Abortion Same-sex marriage Prayer in public schools Health care Taxing the wealthy Education spending |
Expressive partisanship is measured with a four-question partisan identity scale consisting of items such as “When talking about [Democrats/Republicans], how often do you use ‘we’ instead of ‘they’?” The YouGov Study also applies this scale to ideology, providing a measure of ideological identity.
As for political engagement, the Blog, YouGov, and Student studies look at current campaign activity, using respondents’ self-assessed likelihoods of contributing money and volunteering for candidates and political organizations during the upcoming election season. Similarly, the Blog and NY studies ask about prior electoral activity. Finally, the Blog and Student studies include experiments designed to instill feelings of political threat or reassurance. In the Blog Study, respondents are assigned to read a fake blog post that predicts long-term victory or defeat for their party. Subjects receiving good news (from their partisan perspective) were coded as receiving the party reassurance cue, whereas recipients of bad news received the party threat cue. Subjects’ levels of anger, hostility, and disgust are then measured, followed by their hope, enthusiasm, and pride. The Student Study used the same party reassurance/threat cues, as well as new treatments conveying reassurance and threat for specific policies — such as health care, gay marriage, and government spending levels — rather than parties. The issue reassurance conditions are, from the subject’s ideological perspective, optimistic about future policy outcomes, while the issue threat conditions are pessimistic.
These measures allow the authors to assess how well instrumental and expressive partisanship each predict political engagement and emotional responses to partisan threat and reassurance.
Main Findings
The authors’ analyses consistently indicate that expressive partisanship is more influential than instrumental partisanship. First, they show how partisan identity is consistently and strongly predictive of (past and current) electoral activity, whereas ideological issue intensity has a much weaker, sporadically significant relationship with such activity. Ideological identity also shows a moderate, significant relationship with electoral activity, even when accounting for partisan identity, which the authors interpret as further evidence bolstering the case for expressive partisanship (while acknowledging that ideology can also represent instrumental motives).
Second, the two experiments consistently suggested that expressive partisanship better predicts subjects’ emotional reactions to political news than instrumental partisanship. Specifically, the blog post treatments have greater effects on respondents with strong partisan identities, who show more anger in response to the threat condition and more enthusiasm following the reassurance condition. This pattern is also present among those high in ideological issue intensity, but the relationship is always weaker and only sometimes significant.
Finally, the Student Study’s partisan-based treatments appear to be more impactful than the issue-based treatments — at least for those strongest in expressive and instrumental partisanship, respectively. Based on the results, partisan-based threat and reassurance significantly affect respondents strong in partisan identity, whereas no such effect is detected for issue-based threat/reassurance among those high in ideological issue intensity. (However, a simple comparison of how the partisan- and issue-centered treatments affected all respondents in the Student Study is unavailable, which means there is no evidence as to whether party-based or issue-based messaging is more impactful for the overall electorate.)
Implications/Unanswered Questions
Political scientists have long accepted that party identification heavily shapes voters’ political beliefs and behavior, while also debating whether partisanship represents more of a social identity or a set of issue preferences. This paper provides evidence bolstering the expressive partisanship side of the debate. Partisanship can be thought of as a social identity — one with significant effects on voters’ behavior and psychology. As the authors recognize, this raises questions about the quality of representation that parties will provide to voters motivated by expressive partisanship. If party ID stems less from issue preferences and more from tribal loyalty, then politicians may have less incentive to serve their voters’ policy interests and more reason to pander to symbolic identities and dislike of the other side. However, expressive partisanship may also promote voter engagement, as well as lead politicians to show respect for their constituents’ identity-centered concerns. The authors leave the weighing of these pros and cons for future research.
Beyond the normative implications, this paper raises additional questions. For one, researchers should examine the relationship between partisan identity and other phenomena, such as politically motivated reasoning or support for democratic norms. Also, what is the substance behind partisan identity? Is it a product of a sense of linked fate with one’s co-partisans? Of polarization against the opposite party? Of associations between political parties and other social groups? Of built-up investment after experiencing the ups and downs of politics? This paper’s definitions are compatible with all of the above. Finally, while this paper suggests a relatively weak role for instrumental partisanship, it still shows that policy concerns still can be significantly predictive of political engagement in certain contexts. These results ought to prompt additional research examining when (and which) policy concerns can impact political engagement.
Methods and Analysis
Was the study and its analyses pre-registered?: Study was conducted before 2015
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: 0.1
Were correlational results interpreted with causal language?: Yes
In the last paragraph of the Why Differentiate Instrumental and Expressive Forms of Partisanship section (“We find that partisan actions taken during an election campaign, such as donating money and working for a candidate, are driven by partisan identity…”). Also, there are some instances in the Campaign Activism: Party Identity and Issues subsection on page 10 (examples from the 2nd/3rd-to-last paragraphs: “In the blog study, a strong position on gay marriage increased past … and current involvement”, “a strong partisan identity was far more likely than strong ideology or commitment to a single issue to drive campaign activity”).
Limitations / Weaknesses
The authors provide persuasive evidence that expressive partisanship is related to mass political behavior. Of particular import is their success in replicating this finding across several different samples. However, the paper may underestimate the impact of instrumental partisanship, due to two limitations that each could artificially weaken respondents’ estimated ideological issue intensity. First, the studies do not measure the impact of any issue positions not included in Table 1. These neglected issues include contentious topics ranging from gun policy to racial equality. Similarly, the experimental blog posts only discuss health care policy, gay marriage, and the size of government. As a result of this limited scope, the paper’s measure of ideological issue intensity may not adequately represent the impact of policies on voters, simply by leaving many issues out. Second, some respondents (such as single-issue voters) may only be driven by a subset of the included issues. If so, the approach of combining all issue positions could make those respondents look more apathetic about policy than they are in reality. After all, the authors mainly measure ideological consistency, but some people may place no priority on their ideologically incongruent positions. Such voters may still be as issue-driven as perfectly consistent ideologues, but this paper’s data would indicate otherwise. The authors try to account for this possibility in two of the studies by incorporating a measure of issue importance. But this measure, while commonly used, may lead respondents to overstate how important they consider issues to be, according to new research (Ryan and Ehlinger 2023). Overall, the paper’s gauge of instrumental partisanship is less airtight than its measure of expressive partisanship, which may help explain why the latter appears to be more impactful than the former. Finally, expressive and instrumental partisanship were not exogenously assigned. As these variables may be endogeneous to one another (as well as other variables), caution must be taken when making causal inferences and the relative impact of these variables.
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
Huddy, L., Mason, L., & Aarøe, L. (2015). Expressive Partisanship: Campaign Involvement, Political Emotion, and Partisan Identity. American Political Science Review, 109(1), 1-17. doi:10.1017/S0003055414000604
Bibtex
1
2
@article{Huddy_Mason_ Aarøe_2015, title={Expressive Partisanship: Campaign Involvement, Political Emotion, and Partisan Identity}, volume={109}, DOI={10.1017/s0003055414000604}, number={1}, journal={American Political Science Review}, author={Huddy, Leonie and Mason, Lilliana and Aarøe, Lene}, year={2015}, pages={1–17}}