Are courts “different?” Experimental evidence on the unique costs of attacking courts

Abstract

U.S. courts have long been thought to be held in special regard by the American public, and public support is theorized to protect institutions from interbranch aggression. At the same time, recent research underscores that institutional fealty and public reaction to court curbing is shaped by partisan concerns. Drawing on a survey experiment fielded in the U.S., we evaluate whether (1) the public is uniquely punitive toward incumbents who seek to undermine a court rather than an agency and (2) the extent to which these penalties are dependent upon shared partisanship with the proposer. We demonstrate that the public is less supportive of efforts to strip judicial power than analogous efforts to strip power from an executive agency, but that this penalty for court curbing dissipates in the face of copartisanship. This substantiates previous claims regarding the role of partisanship on shaping public attitudes about high courts but underscores that the American public may still hold the courts in unique regard.

Americans generally like their courts more than legislatures or executives. Courts benefit from a perception that they operate “above politics” (Lipset and Schneider, 1987: 69) and are separate from the politicking that characterizes “the Washington system” (Hibbing and Theiss-Morse, 1995). This gap in support appears early as children are socialized into an understanding of politics emphasizing judicial independence and impartiality (Easton and Dennis, 1969) and results in a positivity bias in adulthood as the norms, rituals, and symbols that accompany the judicial process makes it unique in the minds of citizens (Gibson and Nelson, 2018). These symbols “reinforce [Americans’] predispositions to regard the Court as special” and worthy of more support than other branches of government (Ura, 2014: 113).

This public positivity bias is thought to have important political consequences because higher levels of public support confer political advantages to institutions. Public support helps institutions to secure acquiescence and compliance with their decisions (Gibson et al., 2005). Higher levels of public esteem may also help institutions withstand attacks on their power and authority, thereby helping to maintain the balance of power among political institutions (Caldeira and Gibson, 1992).

However, because scholars have traditionally analyzed support for courts and judicial reform in a vacuum, empirical evidence on the consequences of this public esteem advantage is scant. Where scholars have compared the public’s response across institutions, the evidence has been inconclusive. Some suggest that courts are uniquely able to persuade the public to support or adhere to certain policies (Bartels and Mutz, 2009; Gibson et al., 2005; Hoekstra, 1995), while other studies find no evidence of this effect (e.g., Bishin et al., 2021). To compound the issue, we lack systematic evidence on the second advantage conferred by public esteem: whether courts are better able to weather assaults on their institutional power than their peers (c.f. Carey et al., 2020). This ability is particularly important for courts whose supposed counter majoritarian role requires them to be able to make unpopular decisions without fear of reprisal and is especially relevant during a period of time in which concerns about democratic backsliding have come squarely into view.

Relying on an experiment fielded on a diverse national sample of Americans, we investigate whether the public’s outsized regard for judicial institutions corresponds with a differential willingness to punish incumbents for proposals of interbranch attacks. Bringing data to bear on the assertion that “courts are special” and acknowledging the powerful role partisanship plays in modern American politics, we hypothesize that (1) the public will be less supportive of institutional reform proposals that undermine judicial power, relative to those that would change other institutions, and (2) these penalties will be weakened in the face of shared partisanship with the proposer. We find support for both hypotheses. Our research corroborates a growing literature suggesting that the public’s respect for the judiciary is affected by partisanship (Bartels and Johnston, 2020; Driscoll and Nelson, 2023), and has important implications for those who care about maintaining the stability of political institutions at a point in history when attacks on the authority of political institutions are commonplace.

Are courts unique? The public’s evaluation to proposals of reform

Political scientists have long suggested that, so long as institutions have a threshold level of public support, the public may retaliate against incumbents who seek to meddle with their structure or function (Caldeira and Gibson, 1992). In this way, the electoral connection protects political institutions from overt incumbent efforts at politicized reform. American courts have long been regarded as “special” in this regard, as the public’s widespread and enduring esteem has afforded judicial institutions a durable base of institutional authority, even as they make displeasing decisions (Ura, 2014). If high public esteem is thought to “shield” courts from political interference, and U.S. courts have been more warmly regarded than other political institutions, then we should expect that court curbing might elicit an especially negative response in public opinion. Although institutional reforms happen to all types of political institutions, and public evaluation of court curbing has generated a vibrant stream of recent research, we have little empirical evidence to document the extent to which the costs of reforming courts differ from those incurred by politicians who seek to curb other institutions. This is a critical baseline to our understanding of court curbing, providing key insights into the question of whether the public regards courts as “unique.”

Given that public support supposedly shields institutions from interbranch aggression and courts typically enjoy higher levels of public support than other types of political institutions, we expect that citizens will withhold support for incumbents for court curbing proposals at a greater level than proposals which seek to alter another branch of government. If it is the case that courts hold a privileged position in the public eye, the penalties for court curbing should be greater than for similar reforms proposals to other institutions.

We further expect the size of the penalty to vary based according to the partisanship of the proposer (Bartels and Johnston, 2020; Driscoll and Nelson, 2023; Nelson and Driscoll, 2023). A long line of research has demonstrated that the public is more likely to approve of policies when they are proposed by copartisans, especially when the issues are low salience (Mummolo et al., 2019). Indeed, rather than punishing incumbents for holding disappointing policy stances, many voters will just adopt a copartisan’s issue position as their own, following the cues of well-placed opinion leaders. Recent scholarship suggests this logic applies to issues of institutional integrity. Experimental research documents that subjects were more favorable to court reforms when proposed favorite candidates (Armaly, 2018), and that the public is willing to accept some attacks on courts when they approve of the attacker (Clark and Kastellec, 2015; Nelson and Gibson, 2019). Similarly, Badas (2019) draws on historical survey evidence to show that supporters of the New Deal suite of policies favored Roosevelt’s proposed expansion of the Supreme Court and curtailing of the Court’s jurisdiction, so as to limit interference with the New Deal’s passage and implementation. Thus, we expect that the costs of court curbing are greater for out-partisan, rather than copartisan, legislators.

Research design

To test these hypotheses, we conducted an experiment using a survey of 1,586 respondents, selected to mirror the composition of the American people on the basis of gender, race, ethnicity, and age and fielded by Qualtrics from 29 July–2 August 2019. The vignette varied on two dimensions: (1) the content of the legislative proposal, and (2) the proposer’s partisanship. Our hypothetical senator proposed to reduce the U.S. Supreme Court’s ability to provide protections to LGBT students in public schools. ¹ We contrast this treatment with two other types of legislative proposals: an analogous proposal to strip the U.S. Department of Education’s rulemaking ability regarding LGBT protections in public schools, and a benign proposal to rename a local U.S. post office in honor of Benjamin Franklin. These two treatments allow us to compare the effect of court curbing to a similar proposal that targets another type of political institution and against a common-place congressional proposal. Each of these proposals was said to have been supported by one of three types of incumbents—not stated, Copartisan, or Out-partisan. ² Respondents were therefore randomly assigned into one of nine equally sized categories (3 proposals × 3 incumbent partisanship).

An example treatment read as follows:

Imagine that an incumbent Republican Senator from a nearby state who is seeking reelection in November, 2020 has introduced a bill that would restrict the ability of the U.S. Supreme Court to decide cases involving LGBT accommodations in public schools. ⁴

In the other two treatments, respondents were told either that the incumbent “...has introduced a bill that would restrict the ability of the Department of Education to regulate issues involving LGBT accommodations in public schools” or “...has introduced a bill that would rename a post office in the state capitol in honor of Benjamin Franklin.” We focus on this note in the Court-Agency comparison because the two proposals are most similar, providing the cleanest test of the consequences of court curbing rather than analogous institutional reforms. The Appendix discusses the Court-Post Office comparison, which enables us to compare court reform to a pedestrian congressional action.

We followed the experiment with four outcome variables. ³ Our first two outcome variables gauge the approval of both the incumbent’s job performance and the proposal itself. 51% of respondents had a positive assessment of the incumbent’s job performance, and a slight majority of respondents approved of the proposal. The third outcome measures the possibility that court curbing proposals might galvanize supporters, probing whether they would be willing to contribute to their campaign, attend a campaign rally, wear a campaign sticker or try to persuade a friend to vote on the incumbent’s behalf. 38% of respondents were at least “somewhat likely” to be mobilized by the incumbent. Finally, we measured vote intention, asking whether the respondent would vote for the incumbent at the ballot box. 22% answered affirmatively. These four items form a reliable scale, with a Cronbach’s alpha value of 0.83. Factor analysis reveals a unidimensional structure with an average factor loading of 0.74 on the first dimension. Given these stong psychometric properties, we combine these four variables into an interval measure of overall Incumbent Support, and rescale that variable to range from 0 to 1 (larger values indicate more support), and use it as our outcome variable. In the Appendix, we provide results using the individual indicators.

Importantly, to the extent that differences in support for institutions have their root in beliefs about impartiality, comparing responses to judicial and administrative jurisdiction stripping presents a critical test of our theory. Both courts and agencies purportedly operate through a routinized decision-making process that emphasizes impartiality, and both federal judges and bureaucrats in the United States typically enjoy strong job protections provided by life tenure or civil service laws. The data provide support for the assumption underlying our analysis. In response to a standard job performance question, 65% of respondents said the U.S. Supreme Court was doing a “great” or “pretty good” job. That percentage is higher than the analogous congressional (39%), presidential (45%), and Department of Education (45%) figures. Thus, even though we selected two institutions that might benefit from the rosy benefits of purported independence and impartiality among the public, our sample, expressed more positive affect for the judiciary other institutions.

Results

We begin by looking for differences across the three different types of proposals: one to strip the jurisdiction of the U.S. Supreme Court, a second to strip the jurisdiction of the U.S. Department of Education, and a proposal to rename of a post office. We report the unconditioned average treatment effects, though our results do not change when we include controls for respondents’ political and demographic characteristics, as shown in the Appendix.

These direct effects (and accompanying 95% confidence intervals) are plotted in the left-hand panel of Figure 1. Respondents are most supportive of incumbents renaming a post office, scoring an average of 0.52 on our summary support score. This support drops by 0.07 for incumbents who aimed to strip the jurisdiction of the U.S. Department of Education and by another 0.06 for incumbents who seek to undermine the Supreme Court. These results provide some support for our hypothesis that the public regards courts as unique, and is less supportive of incumbents who seek to curb courts than other political institutions.

Figure 1.

Incumbent support across all treatments. The first two panels display the direct effects of the experiment. The right-hand panel displays the mean differences between the Supreme Court and Department of Education treatments. The point estimates represent the dependent variable is scaled on the 0–1 interval. In all panels, the whiskers represent 95% confidence intervals. Higher values of the outcome variable indicate more support for the incumbent.

The middle panel of Figure 1 displays the average treatment effects by partisanship. Public responses to court curbing illicit the strongest negative reaction across our treatments and appear to be most heavily conditioned by shared copartisanship. Respondents were more supportive of a copartisan rather than an out-partisan who sought to strip the Supreme Court of jurisdiction, although support for a copartisan who attacks the court is no different than support for an incumbent whose partisanship was not stated. This suggests that the public may punish incumbents for attempts to undermine the high court, but the extent to which this is true is conditioned by copartisanship.

On the other hand, we observe no differences according to partisanship in the Department of Education conditions: respondents are no more willing to punish an out-partisan legislator who seeks to restrict the authority of that agency than a copartisan with the same goal. Likewise, we find no evidence that partisanship influences the public’s responses to the benign post office proposal.

Our hypotheses concerned the differential cost of attacking a court rather than an agency and how that effect varies by copartisanship. The right-hand panel of Figure 1 presents these results. The first point in the figure shows the overall cost, averaging across the three partisanship conditions. There is a greater cost to court curbing than stripping the jurisdiction of an agency; the difference is −0.06 (p < .01), accounting for about a one-quarter standard deviation shift in the outcome.

As the remaining points in the right-hand panel of Figure 1 demonstrate, this effect is driven by responses to out-partisan incumbents. For respondents who learned of a proposal by a copartisan or an incumbent whose partisanship was not provided, there is no statistically discernible difference in the level of incumbent support between attacked institution. For the generic (no party stated) incumbent, the estimated difference is −0.05 (p = .14); for copartisan incumbents, the estimate is −0.03 (p = .39). By comparison, there is a statistically distinguishable difference in the cost of attacking a court rather than an agency for out-partisan incumbents. Here, the difference is over three times as large as the effect for copartisan incumbents (−0.10) and is statistically significant (p < .01), a difference that amounts to about half of a standard deviation of the outcome.

Discussion and conclusion

We sought to examine the extent to which the public is more protective of courts than other types of political institutions. The U.S. courts have long benefited from a political socialization process that emphasizes their purported independence and impartiality, scholars have for decades reported that judicial institutions were “special” in the eyes of the mass public (Caldeira and Gibson, 1992; Ura, 2014). We took this logic to design an experiment to contribute to a vibrant stream of research on public reaction to court curbing: research has investigated public reaction to attacks on courts, but only ever in isolation, and never with reference to a comparable institutional reform (Bartels and Johnston, 2020; Driscoll and Nelson, 2023).

We therefore expected that penalties associated with attacking courts would be more severe than attacking other institutions that would seek to enact similar policies. We tested our argument in an experiment that compared the costs of curbing a court and an administrative agency—two institutions that are distant from Hibbing and Theiss-Morse’s “Washington system” but which have similar jurisdictional involvement in the implementation of policy. We acknowledge that this single experiment on a single issue gives a narrow view of the public’s esteem, but this specificity was by design: it allowed us to examine respondents’ evaluations of realistic institutional measures in a salient policy realm that is directly comparable across institutions.

Our results both corroborate and challenge the established wisdom that “courts are special” in the public eye. On the whole, we report that the public was less supportive of incumbents who proposed stripping the jurisdiction of a court rather than an agency, a finding that is consistent with conventional wisdom. Yet, once we account for partisanship, we see that the consequences for institutional reform are strongly conditioned by partisanship; the observed differences are driven by a greater willingness to punish out-partisan incumbents who aim to undermine the judiciary than one who seeks to limit the power of an administrative agency. When the legislator is an out-partisan or has ambiguous partisanship, we observe no evidence that the costs of curbing a court differ from those of attacking an agency. And, as we show in the Appendix, when we focus our analysis on vote choice—the most electorally consequential indicator—we see no difference in respondents’ willingness to punish even out-partisan incumbents who seek to undermine court authority. Importantly, although we uncovered no electoral benefit for incumbents to introduce these proposals, they may exist. After all, many candidates campaign on the prospect of “reforming” political institutions. In the context of our experiment, we can imagine that some subset of the public regards the rapid expansion of rights for the LGBT community—as well as the Court’s role in entrenching those rights—as an illegitimate overstep of judicial authority. From this vantage, incumbent efforts to limit the court’s unjustified activism may foster admiration among citizens or mobilize a base of electoral support (Nelson and Driscoll, 2023). Yet, our results show an overall, albeit modest, support cost to these proposals. Future work should probe these results further, investigating when (if ever) these types of proposals may benefit candidates, and evaluate a broader portfolio of possible costs the public might inflict (Driscoll and Nelson, 2023).

There is much that remains to untangle. The costs of court curbing are likely to exceed those faced by incumbents who seek to weaken legislators or executives, institutions that do not benefit from the halo of supposed impartiality among the public. And, among institutions that are associated with rank politicization and heightened partisanship, the partisan differential in the costs of attacking these institutions will likely be more pronounced than what we observe here. Further, because the control condition in our experiment—the post office condition—was a benign institutional reform, we are unable to test an obvious individual-level hypothesis underlying our theory: that those who hold an institution in higher esteem are more likely to punish an incumbent. Thus, the research agenda moving forward is both clear and important.

Supplemental Material

Supplemental Material - Are courts “different?” Experimental evidence on the unique costs of attacking courts

Supplemental Material for Are courts “different?” Experimental evidence on the unique costs of attacking courts by Amanda Driscoll and Michael J. Nelson in Research & Politics