By Travis N. Taylor, Ph.D.[1]
Photo credit https://nappy.co/photo/XcMdQHJb6rI61jUdKDwGl
Does political identity affect how partisans approach jury service? Partisanship in America permeates nearly every aspect of life. Shared partisanship with a candidate is the most powerful predictor of vote choice.[2] However, in recent years, partisanship has become much more than that; it is now about more than politics. The political party with which we align ourselves now correlates with what we view as the meaning of life,[3] where we shop,[4] what we drive,[5] our religious preferences,[6] who we marry,[7] and even whether we marry at all.[8]
All of this is perhaps unsurprising to those who study social groups. Studies show that once groups are formed, they establish relatively stable norms,[9] are cohesive,[10] and exhibit strong preferences for in-group members over those in the out-group.[11] As Hogg summarizes, “group[s]…have such a powerful influence over the way we behave and the way we perceive ourselves and the world we live in.”[12]
Political science views social grouping by political party in two ways. First, as an in-group, complete with a group identity and allegiances. “[W]hen citizens talk about what they like and dislike about the political parties,” writes Huddy, “they comment on the shared beliefs, common traits, and the specific individuals who typify group members…political groups are seen as cohesive social entities with shared beliefs and common social characteristics.”[13] In simple terms, Republicans tend to like Republicans and Democrats tend to like Democrats.
There is also a natural competition between the parties. Yes, there is the electoral contests that pit Republicans against Democrats, but it is more than that. Through the lens of negative partisanship, partisans have come to not only compete against one another but to loathe members of the other party. Consider that in recent years, Americans in both parties “say that the other party’s members are hypocritical, selfish, and close-minded.”[14]
The question this paper answers is whether partisanship impacts jury service. Alexis de Tocqueville famously wrote that the jury is primarily a political body.[15] In addition, Jolly has theorized that, although partisanship has historically played a role in juries, today’s hyperpolarization elevates partisanship on juries to a “troubling” level.[16] He writes, “Partisans are prejudiced against opposing party members and take affirmative steps to discriminate against them based on these identities…The jury as an institution is not immune from this political malady.”[17] This paper puts Jolly’s partisan capture theory to the test. Given the recent high-profile court cases with explicitly political overtones, understanding how partisan jurors approach these cases is imperative for trial consultants and attorneys who use voir dire to identify bias and select an impartial jury.
I hypothesize that jurors will view themselves and a defendant more favorably when they share a partisan identity with the defendant in an explicitly political case than when they either are not co-partisans or do not know the defendant’s partisan affiliation.[18]
Data & Methods[19]
To test this hypothesis, I conducted an original online survey of a national sample of 1,429 jury-eligible respondents. To determine jury eligibility, respondents must: (1) be registered to vote, or (2) have a valid driver’s license or state identification card, and (3) not be a convicted felon. Respondents were recruited to participate from a non-probability, opt-in online panel. Research has shown little to no mode effects between non-probability panel samples and other probability-based modes of survey research.[20]
Embedded within the survey was an experimental vignette of a political corruption case. In the vignette, a politician was charged with political corruption and was on trial. Brief arguments from the prosecutor and defense were presented to the participants. The vignette’s experimental design varied only the defendant’s party affiliation, which was presented in a sentence about his having recently won a primary election. Respondents were randomly assigned to either a control condition, which did not contain the defendant’s partisanship or one of two experimental conditions—one in which the defendant was identified as a Republican and one in which he was identified as a Democrat.
Following the vignette, respondents were asked five questions about the case and their ability to serve as a juror on it. The questions explored (1) whether respondents felt they could be impartial jurors, (2) whether they could be fair to the defendant, (3) whether they believed he was guilty or not guilty, (4) whether they would vote to acquit or convict him, and (5) the punishment they felt was appropriate if the defendant were convicted. On each of these measures, respondents in the same party treatment group should evaluate themselves more highly (more impartial and fairer) than members of the control group and be more favorable to the defendant (less likely to believe in his guilt, less likely to convict, and less punitive in sentencing) than the control group.
Respondents were asked in the demographics section of the survey to state with which political party they most closely identified. The seven point scale, ranging from Strong Republican (coded 1) to Strong Democrat (coded 7) included options for respondents to self-identify as an independent or an independent who usually votes for Republicans (coded 3) or Democrats (coded 5). For the purposes of the analysis, the seven-point scale was collapsed into three categories: Republicans, Democrats, and independents. Consistent with political science research norms, independents who generally vote with one party over the other are coded as partisans.[21] Demographics, including party affiliation, were asked at the end of the survey to alleviate priming.
Table 1 shows the conditions created by the various configurations of Respondent–Defendant partisanship.
While these groups may benefit from methodological purity—random assignment to either a treatment or control group and no chances of pre-test exposure to the treatment—they do not reflect reality. When would you expect the partisan affiliation of a politician turned criminal defendant to be unknown, as in the experimental control? The probability of such a scenario approaches zero.
Fortunately, the data provide the opportunity for a more realistic test. Beyond comparing the treatment groups to the control group, I also compare the treatment groups to one another. Put another way, I will compare the responses of same-party Respondent-Defendant pairings to the responses of opposite-party pairings: in-group versus out-group. Compared to out-group respondents, in-group respondents will: (1) think of themselves as more impartial, (2) consider themselves to be fairer, (3) think of the defendant as not guilty, (4) lean more toward acquittal, and (5) be less punitive.
|
Table 1: Experimental Groups |
||||
|
|
|
Respondent Partisan Identification |
||
|
|
|
Democratic |
Independent |
Republican |
|
Defendant Partisan Affiliation |
Democratic |
Same Party Treatment |
Control |
Opposite Party Treatment |
|
Not Identified |
Control |
Control |
Control |
|
|
Republican |
Opposite Party Treatment |
Control |
Same Party Treatment |
|
To ensure a comparison-of-means test is appropriate, I conducted a series of means tests between the groups on the relevant characteristics. Across all of the characteristics, except for employment status in the same- versus opposite-party comparison groups, the groups are balanced; the mean levels are not statistically significantly different from one another. Thus, I use t-tests for all of the following analyses.
Analysis
Results indicate low levels of support for the hypothesis when treatment groups are compared to the pure control group. Of the ten analyses conducted, only one was statistically significant.
The results show no support for the hypothesis that respondents whose party affiliation was different from the defendant’s would treat the defendant worse than respondents in the control group. Across all five questions posed, there were no statistically significant difference between the treatment and control groups. Compared to a pure control condition, jurors do not appear to treat an out-group defendant any worse.
I also find limited support for the hypothesis that same-party jurors would treat the defendant better than jurors in the control group. In an initial analysis, same-party respondents were not significantly different than the control group on any of the dependent variables. However, once I narrowed the analysis, I found that, among jurors who leaned toward a conviction, jurors who share a partisan affiliation with the defendant believe a lighter sentence is appropriate (x̄=2.85, on a scale from 1 to 6) than do members of the control group (x̄=3.06), a statistically significant difference of 0.20 (p=0.03).
Turning to the realistic comparison, I find greater support for my hypothesis. In short, same-party juror-respondents are generally more favorable to the defendant relative to jurors who are of the opposite political party as the defendant. See Table 2 for the results.
|
Table 2: In-Group versus Out-Group Experimental Results |
|||||
|
|
Model 1: Impartial |
Model 2: Fair |
Model 3: Guilty |
Model 4: Convict |
Model 5: Punishment |
|
Out Group In Group Difference p-value |
2.79 2.83 0.04 0.33 |
2.96 3.09 0.12 0.04* |
3.78 3.75 0.03 0.39 |
2.65 2.51 0.14 0.01* |
2.99 2.87 0.12 0.08 |
|
*Statistically significant difference at p £ 0.05 |
|||||
While no significant differences exist between the two groups in their self-perceived ability to remain impartial, in-group jurors express significantly higher levels of fairness toward an in-group defendant than out-group jurors.
Turning now from the respondents’ views of themselves to their views of the defendant, we see a shift in the story. Although both groups are equally likely to see the defendant as either guilty or not guilty, members of the out-group (i.e., jurors of the opposite party of the defendant) are significantly more likely to convict him than respondents who share the defendant’s party.
Here, however, is where the results take a normatively troubling turn. From which respondents is the increased likelihood of convicting coming? It is not from jurors who see him as guilty. Among juror-respondents who believe the defendant is guilty, in-group and out-group respondents are equally likely to convict (D=0.11, p=0.13). However, among respondents who lean toward believing the defendant is innocent, out-group participants are significantly more likely than in-group participants to vote to convict (D=0.14, p=0.03). Put more bluntly, partisan jurors are willing to convict their political opponents of criminal charges even when they believe he is innocent.
Turning to the final model, we see that, overall, there is no difference in punishment between the in-group and out-group respondents. Digging deeper, however, there is again more to the story. Among respondents who say they would vote to convict, the consequences are significantly more punitive when coming from out-group juror-participants (x̄=3.12) than from in-group participants (x̄=2.85), a difference of 0.27 (p=0.01). When jurors vote to convict a defendant in a political corruption trial, partisanship strongly affects their decision-making vis-à-vis punishment.
Discussion & Implications
In People v. Trump, the New York State “hush-money” case against then-former-President Donald Trump, the trial judge, Juan Merchan, explicitly forbade counsel from exploring during voir dire the party affiliation of prospective jurors. In a letter to counsel, Justice Merchan instructed: “jurors [are not] asked [in the jury questionnaire] about their specific political party registration, though the answer to that question may easily be gleaned from the responses to the other questions. Counsel is forewarned not to seek to expand the degree of intrusion beyond what is relevant and has already been approved” (emphasis mine).[22] In other words, the judge appeared to think that jurors’ political affiliation was irrelevant to their views of a case in which a political actor (a former president and current candidate) was accused of a criminal act in furtherance of political goals (paying a bribe to protect the defendant’s political ambitions).
That view has been widely debated, mostly for partisan purposes. But now, we have experimental data which indicate that the judge’s view was likely wrong. As consultants, we know that people are bad at putting aside biases, even when a judge instructs them to do so or they say they can. It stands to reason that political identity is no exception. Partisanship matters, even to jurors who believe they can be impartial.
Partisan differences permeate American life including, as we see in this study, in juror decision-making. Compared to those of the same party, jurors who belong to the party opposite of the defendant in a political corruption trial say they are less able to be fair to the defendant, are more likely to convict him (especially if they view him as innocent), and will dole out harsher punishments once they have voted to convict. All of this after reading only a single-paragraph vignette. Jurors bring their partisan biases to the jury box with them, even when they claim they don’t.
Implications
The implications for trial attorneys are clear. When politics is likely to play a role in a case, arguments for uncovering partisan bias must be made to the extent permissible by law. This can be accomplished through questionnaires, voir dire, or even change of venue. When the defendant is a political actor, and his or her politics is known, the political leanings of prospective jurors must be evaluated for the purposes of empaneling an impartial jury. Such evaluation need not be conducted through only voir dire or a jury questionnaire—online background and social media research are also avenues which offer insights on political leanings.
For trial consultants, we must be cognizant of the political implications of a case. First off, we must identify and explore political themes in pre-trial research. To the extent possible (which, quite frankly, may be very limited[23]), we must look for ways to mitigate the political nature of a case and the political tensions that are likely to exist between a defendant and jurors. Moreover, during voir dire, we must be prepared to provide our attorney-clients with the empirical cover to root out political bias in the jury pool. We can support them with data they need to make the argument that for-cause challenges are appropriate to eliminate political bias.
Limitations
This research, as does all research, has its limitations. First, and foremost, survey research, no matter how well designed and rigorously analyzed, cannot replicate the deliberative environment of a real jury. The decisions made in this survey were made by individuals, not juries.
Second, this was a low-stakes decision. Even though the instructions made no mention that the case was hypothetical, respondents may have still treated it as such, knowing their response had no real-world consequence. And, while loosely based on a real case, the vignette was indeed hypothetical. A real-world decision would have warranted—and received from a real jury—much deeper consideration and thought.
Third, the case presentation was 156 or 157 words (depending on whether the respondent saw a treatment condition or the control condition). Such a short vignette does not provide jurors with the full picture of a case. While this is a limitation of the research, the results should sound alarms. If respondents, who admit they can be impartial jurors, are willing to convict a defendant even if they believe he is innocent because he belongs to the other party and are willing to hand out longer sentences to a defendant of the other party after reading only 157 words about the case, how much more lopsided will their view be after they have seethed in a courtroom during a three week trial?
Finally, this research is what might be called an “easy test” of the idea that partisanship matters to jury decision-making. It involved an explicitly political scenario in which most respondents were aware of the defendant’s partisan affiliation. The question that remains unanswered herein is whether partisanship impacts juror decision-making in less explicitly political cases, such as cases about climate change or involving police officers, or in cases that are wholly apolitical. Those questions are reserved for future research.
Conclusion
Americans are political creatures, and our partisanship is a core identity for many. For some individuals, partisanship even wields more influence in judgments than inherent characteristics such as race.[24] Moreover, the American jury is a political institution,[25] and while some have theorized that partisanship affects jury decision making,[26] there has been no empirical evidence of it until now. I have shown that jurors’ partisan identification significantly shapes their decisions in an explicitly political trial. Jurors who identify with the opposite party of a defendant say they will treat him less fairly, are more likely to convict him even if they believe he is innocent, and, once convicted, will punish him more severely than jurors of the defendant’s own party. This raises important questions about procedural fairness when courts disallow probing on prospective jurors’ political views. After all, to paraphrase a quote I once saw on the wall of a Baton Rouge lobbying firm: If there’s more than one person in the room, it’s political.
[1] President & Chief Research Officer, Strategic Insights Research, LLC. The author wishes to thank Dan Wolfe, Sarah Murray, Michelle LaRocca, and Clint Townson for their thoughtful comments and suggestions at various stages of this research. All remaining mistakes are my own.
[2] Campbell, Angus, et al. 1960. The American Voter. Chicago: University of Chicago Press; Lewis-Beck, Michael, et al. 2008. The American Voter Revisited. Ann Arbor: University of Michigan Press.
[3] Silver, Laura, and Patrick van Kessel. 2021. “Both Republicans and Democrats prioritize family, but they differ over other sources of meaning in life,” Pew Research Center. Available at: https://www.pewresearch.org/short-reads/2021/11/22/both-republicans-and-democrats-prioritize-family-but-they-differ-over-other-sources-of-meaning-in-life/
[4] Panagopoulos, Costas, et al. 2020. “Partisan Consumerism: Experimental Tests of Consumer Reactions to Corporate Political Activity,” The Journal of Politics 82(3): 996–1007.
[5] Hetherington, Marc, and Jonathan Weiler. 2018. Prius Or Pickup? How the Answers to Four Simple Questions Explain America’s Great Divide. New York: Harper Collins.
[6] Doherty, Carroll, Jocelyn Kiley, and Nida Asheer. 2024. “Changing Partisan Coalitions in a Politically Divided Nation: Party Identification Among Registered Voters, 1994–2023,” Pew Research Center. Available at https://www.pewresearch.org/politics/2024/04/09/party-identification-among-religious-groups-and-religiously-unaffiliated-voters/
[7] Wang, Wendy. 2020. “Marriages Between Democrats and Republicans are Extremely Rare,” Institute for Family Studies. Available at https://ifstudies.org/blog/marriages-between-democrats-and-republicans-are-extremely-rare
[8] Rothwell, Jonathan, and Frank Newport. 2024. “When and Why Marriage Became Partisan,” Gallup. Available at https://news.gallup.com/poll/646793/why-marriage-became-partisan.aspx
[9] See e.g., Sherif, M. 1936. The Psychology of Social Norms. New York: Harper.
[10] Friedkin, Noah E. 2004. “Social Cohesion,” Annual Review of Sociology 30(1): 409–25.
[11] Tajfel, Henri, M. G. Billig, R. P. Bundy, and Claude Flament. 1971. “Social Categorization and Intergroup Behaviour,” European Journal of Social Psychology 1(2): 149–78.
[12] Hogg, Michael A. 2010. “Influence and Leadership,” in Handbook of Social Psychology, 5th ed. Edited by Susan T. Fiske, et al. New York: John Wiley & Sons, Inc., 1176.
[13] Huddy, Leonie. 2003. “Group Identity and Political Cohesion,” in Oxford Handbook of Political Psychology. Edited by David O. Sears, et al. New York: Oxford University Press, 515.
[14] Iyengar, Shanto, Yphtach Lelkes, Matthew Levendusky, Neil Malhotra, and Sean J. Westwood. 2019. “The Origins and Consequences of Affective Polarization in the United States.” Annual Review of Political Science 22 (1): 129.
[15] Tocqueville, Alexis. Democracy in America.
[16] Jolly, Richard Lorren. 2025. “Red Juries & Blue Juries,” Arizona State Law Journal 57(1): 101–61.
[17] Jolly 2025: 161.
[18] By “explicitly political case,” I mean a case that involves a political defendant with known political affiliations involving issues that are, or can be viewed as, political in nature.
[19] A supplemental methodological appendix can be found online at https://www.yessirresearch.com/partisan-jurors
[20] Durand, Claire. 2024. “Polls the Week after the Convention,” Ah! Les Sondages (Ah! The Polls). Available at https://ahlessondages.blogspot.com/2024/08/polls-week-after-convention.html
[21] See, e.g., Petrocik, John R. 1974. “An Analysis of Intransitives in the Index of Party Identification,” Political Methodology 1(3): 31–47, and Petrocik, John R. 2009. “Measuring Party Support: Leaners are Not Independents,” Electoral Studies 28(4): 562–72.
[22] Merchan, Juan M. 2024. Letter to Counsel re. People v. Trump. Available at https://www.nycourts.gov/LegacyPDFS/press/PDFs/Letter-re-jury-selection-4-8-24.pdf. Accessed 27 January 2025.
[23] See, e.g., Lodge, Milton, and Charles S. Taber. 2013. The Rationalizing Voter. New York: Cambridge University Press, and Bartels, Larry M. 2002. “Beyond the Running Tally; Partisan Bias in Political Perceptions,” Political Behavior 24(2): 117–50.
[24] Westwood, Sean J., et al. 2018. “The Tie that Divides: Cross-National Evidence of the Primacy of Partyism,” European Journal of Political Research 57(2): 333–54.
[25] Tocqueville, Alexis. Democracy in America.
[26] Jolly 2025.