A substantial amount of judicial behavior research is devoted to the impact of judges’ personal preferences and attributes on their decisions. Some scholars maintain that judges’ decisions are guided more or less exclusively by legal considerations (Reference DworkinDworkin 1978; Reference GillmanGillman 2001), while others have argued that judges’ decisions are heavily influenced by a variety of other factors, including their political ideology, race, gender, age, religious orientation, social connections, and even birth order (Reference BaumBaum 2006; Reference BlakeBlake 2012; Reference Boyd, Epstein and MartinBoyd, Epstein, and Martin 2010; Reference Collins and MoyerCollins and Moyer 2008; Reference Farhang and WawroFarhang and Wawro 2004; Reference Heise and SiskHeise and Sisk 1999; Reference McGuireMcGuire 2013; Reference Segal and SpaethSegal and Spaeth 1998; Reference Sisk, Heise and MorrissSisk, Heise, and Morriss 1998; Reference Songer, Davis and HaireSonger, Davis, and Haire 1994; Reference YarnoldYarnold 2000). Although this debate is important in understanding and predicting judicial decisions, both sides overlook a crucial aspect of judicial decision making: the perceptions of judges’ audiences. In addition to understanding how judges make decisions, it is just as important to understand how the public thinks judges make decisions. The public's perceptions of courts’ decisions need not correspond to reality. Judges’ decisions are conveyed and interpreted by politicians and pundits who further obscure the already opaque judicial process. As a result, the question of whether a judge's personal biases actually play a role in the decision may have little bearing on whether they are perceived as biasing the decision.
The public's perceptions of decision making on the court can also impact how decisions are made in the first place. Evidence has shown that perceived bias in judicial decision making can weaken the legitimacy of courts and legal institutions (Reference GibsonGibson 1989; Reference RamirezRamirez 2008). If legitimacy reaches a low enough level, the court may have concerns about the non-implementation of its decisions. Because courts control neither the purse nor the sword, judges have an incentive to maintain a certain degree of legitimacy, which may motivate them to act strategically to protect the legitimacy of their decisions and the institution (e.g., Reference Epstein and KnightEpstein and Knight 1998). Among other methods of protecting legitimacy, avoiding the appearance of bias is an important consideration of the court in maintaining their stature and ensuring implementation of their decisions.
In this article, we examine the determinants of perceived bias in courts’ decision making. In particular, we focus on the effect of judges’ personal attributes on perceptions of bias and legitimacy. Previous research has shown that personal attributes such as the racial composition of courts can influence perceptions of legitimacy for racial in-group members (Reference Scherer and CurryScherer and Curry 2010). We continue in this vein by measuring the direct effect of judicial group attributes on perceptions of bias and legitimacy in a given court case.
While a variety of factors may determine whether the public perceives a court's decision as biased, this article focuses on the impact of judges’ membership in a societal out-group. Specifically, we examine how the decisions of atheists, a religious out-group, are perceived differently than those of Christians, a religious in-group. The United States has become increasingly secularized over the past few decades: in contrast to the small fraction of Americans that were religiously unaffiliated in 1950, nearly one-fifth of Americans today are unaffiliated with a religion (Reference Putnam and CampbellPutnam and Campbell 2010). As portions of the American population become more expressly nonreligious, it is important to understand how their role in governing institutions affects perceptions of those institutions. Although atheist identification has become more common, atheists are still one of the most disliked groups in American society (Reference Gervais, Shariff and NorenzayanGervais, Shariff, and Norenzayan 2011; Reference HarperHarper 2007; Reference Putnam and CampbellPutnam and Campbell 2010). Since most Americans view religion as a precondition for morality (Reference Gervais, Shariff and NorenzayanGervais, Shariff, and Norenzayan 2011), they may be especially likely to see atheist leaders as dishonest or immoral. Moreover, the majority of Americans prefer religion to play a relatively prominent role in public life (Pew Research Center 2012). Due to the public's simultaneous distrust of atheists and desire for religion in public deliberations, we hypothesize that the decision making of atheists in public office is less likely to be viewed as fair or legitimate.
Judges’ personal religious attributes are perceived as particularly important factors in their decision making in cases involving religious interests. For example, in the aftermath of the Supreme Court's 2014 decision in Burwell v. Hobby Lobby to exempt certain corporations from the Affordable Care Act's requirement that employers cover contraceptives for their female employees, observers suggested that the religious makeup of the Court may have influenced the outcome of the case (e.g., Reference FreedmanFreedman 2014). The five justices comprising the majority coalition were Roman Catholic, a faith which prohibits the use of artificial birth control. Additionally, the U.S. Council of Catholic Bishops filed an amicus curiae brief in favor of the Hobby Lobby owners’ right to refuse coverage on religious grounds. This high-profile case highlights the importance of religious characteristics to questions of both actual and perceived bias in jurisprudence.
In the proceeding sections, we test the effect of judges’ affiliation with particular religious groups on perceptions of their decision making in a hypothetical case dealing with the constitutionality of a publicly displayed Nativity scene, a seasonal exhibit representing the birth of Jesus Christ in a manger. To strengthen our causal claims, we utilize a survey experiment to show that respondents are significantly less likely to trust a court that decides to strike down the display of a Nativity scene on public property when the judge is identified as an atheist. We observe no corresponding effect for a Christian judge handing down a pro-manger decision, however, suggesting that perceptions of bias and legitimacy may be conditional on the out-group status of judges. These results demonstrate the need for additional research on the factors driving the public's evaluations of governmental institutions.
Procedural Justice and Perceptions of Legitimacy
Scholars have long noted the popularity of the Supreme Court and the judicial system, despite its counter-majoritarian function and relative removal from public oversight (Reference CaldeiraCaldeira 1986; Reference Hibbing and Theiss-MorseHibbing and Theiss-Morse 1995, Reference Hibbing and Theiss-Morse2002; Reference Mondak and SmitheyMondak and Smithey 1997).Footnote 1 Indeed, some have suggested that its operation outside of the public purview actually increases support for the Court (Reference Hibbing and Theiss-MorseHibbing and Theiss-Morse 1995, Reference Hibbing and Theiss-Morse2002). Unlike Congress, which performs the messy job of being responsive to constituents, courts can exercise decision making in ways that do not allow citizens to “see the sausage being made,” as the old adage suggests. In fact, most citizens care little about the formation of public policy, and are willing to let others make even important political decisions on their behalf (Reference Hibbing and Theiss-MorseHibbing and Theiss-Morse 1995, Reference Hibbing and Theiss-Morse2002).
This is not to suggest that citizens do not care at all about the processes of government. As public approval ratings of Congress indicate, citizens are unwilling to cede total control of governing to elected or unelected officials. Rather, citizens tend to care more that the processes of government are functioning as they should: that officials are fulfilling the duties of their jobs rather than taking advantage of their positions for personal gain or to instill their personal values into law (Reference BairdBaird 2001; Reference BeneshBenesh 2006; Reference GanglGangl 2003; Reference Hibbing and Theiss-MorseHibbing and Theiss-Morse 1995, Reference Hibbing and Theiss-Morse2002; Reference RamirezRamirez 2008; however, see Reference Gibson and CaldeiraGibson and Caldeira 2011). This has been termed “procedural justice theory,” where the public is generally more concerned with procedure than with substantive policy outcomes. Applied to the judicial system, this would suggest that citizens are willing to accept most rulings by a court unless they perceive the decision to be rendered in a biased fashion. In sum, the public cares about whether courts are actively following commonly accepted legal guidelines and norms.
Perceptions of whether a court is functioning in such a manner are directly tied to the public's willingness to grant legitimacy to that court (Reference BeneshBenesh 2006; Reference GanglGangl 2003; Reference GibsonGibson 1989; Reference PetrickPetrick 1968; Reference RamirezRamirez 2008; however, see Reference GibsonGibson 1991; Reference MondakMondak 1993). Legitimacy is often defined as diffuse support for the court, or a “reservoir of goodwill” to be tapped when courts make unpopular decisions (Reference EastonEaston 1965; Reference GibsonGibson 1989; Reference Gibson, Caldeira and SpenceGibson, Caldeira, and Spence 2005; Reference MondakMondak 1990; Reference Zink, Spriggs and ScottZink, Spriggs, and Scott 2009). While legitimacy has been measured in varying ways, the essential components include willingness to support, trust, maintain, and accept the authority of courts in the face of decisions of which one disapproves. Thus, mere approval of the court or its decision does not fully capture the concept of legitimacy. Legitimacy requires trust in the court's long-term authority and general ability to make decisions even when issuing unpopular rulings.
For the judicial branch, legitimacy is an especially important tool to maintain. The courts control neither the purse nor the sword, so they rely on others to carry out legitimate policy decisions (Reference CaldeiraCaldeira 1986). Legitimacy may also contribute to one's propensity to acquiesce to or comply with unpopular court decisions (Reference Epstein and KnightEpstein and Knight 1998; Reference GibsonGibson 1989; Reference Gibson, Caldeira and SpenceGibson, Caldeira, and Spence 2005; Reference Tyler and RasinskiTyler and Rasinski 1991). These components are essential for courts to carry out their institutional functions. Therefore, it is not surprising that judges do have an incentive to be concerned with legitimacy (Reference ClarkClark 2009; Reference Epstein and KnightEpstein and Knight 1998; Reference FallonFallon 2005).
Given this link between procedural justice and the legitimacy of courts, we can identify some basic expectations of the public for a decision (and the court) to be viewed as legitimate. First, the public expects judges to rely on legal precedent and to follow legal norms (Reference Baird and GanglBaird and Gangl 2006; Reference RamirezRamirez 2008; Reference Scheb and LyonsScheb and Lyons 2001; Reference Zink, Spriggs and ScottZink, Spriggs, and Scott 2009). Courts are supposed to be “above politics,” and thus any descent into partisan or political interests is likely to be viewed as procedurally unjust by the public. Relatedly, the public expects judges to be able to put aside their personal interests in a particular case. If a conflict of interest is perceived, a judge is expected to recuse him or herself to maintain the legitimacy of the court (Reference Gibson and CaldeiraGibson and Caldeira 2013). Other scholars have also suggested that the public is less supportive of courts deferring to other branches of government, as it may indicate that the courts are not doing their jobs independent of outside interference (Reference Scheb and LyonsScheb and Lyons 2001). Additionally, the public tends to view cases as being fairly decided when they are issued with consensus and support on the court for the majority coalition (Reference Zink, Spriggs and ScottZink, Spriggs, and Scott 2009).
Lastly, an important component to the theory of procedural justice is the trust citizens are willing to place in judges. Trust in government leaders is significantly related to perceptions of how well they are doing their jobs (Reference HetheringtonHetherington 1988). For a case outcome to be viewed as procedurally just, citizens need to trust the judge making the decision (Reference GanglGangl 2003; Reference PetrickPetrick 1968; Reference RamirezRamirez 2008). Thus, trusting judges to make sound decisions has the potential to impact the legitimacy of the court as a whole.
Religious Out-Groups and Procedural “Injustice”
Although the public is concerned with governmental institutions following the prescribed procedure, Americans are also generally inattentive to governmental processes and the decision making of public officials. Public opinion scholars have frequently and consistently documented Americans’ low knowledge of politics and government (e.g., Reference Converse and ApterConverse 1964; Reference Delli Carpini and KeeterDelli Carpini and Keeter 1996). However, the average citizen does not need to be an expert about every subject to form opinions; instead, they can rely on heuristics to fill in the gaps in their political knowledge (Reference LupiaLupia 1994; Reference Page and ShapiroPage and Shapiro 1992; Reference PopkinPopkin 1991; Reference Sniderman, Brody and TetlockSniderman, Brody, and Tetlock 1991). One way heuristics can be employed is to form conclusions about political candidates. Using readily available political information, such as partisan identification or endorsements, and the candidates’ personal characteristics, citizens form opinions about who is suitable to hold public office, likely to follow the procedures ascribed to the office, and whether they are likely to agree with them on substantive policy issues (e.g., Reference DownsDowns 1957; Reference McDermottMcDermott 2006). Among personal characteristics influencing these decisions are race (Reference McDermottMcDermott 1997), gender (Reference KochKoch 2000; Reference McDermottMcDermott 1997), and importantly for this study, religion (Reference Campbell, Green and LaymanCampbell, Green, and Layman 2011; Reference McDermottMcDermott 2009).
Unlike elected officials, federal judges are nonpartisan and appointed, so the public is even less likely to hold informed opinions about judges and the details of their decisions (Reference CaldeiraCaldeira 1986; Reference Gibson and CaldeiraGibson and Caldeira 2009). Judges may be more or less likely to adhere to the expectations of procedural justice, so the public must rely on media portrayals and other background information to assess procedural fairness (Reference Baird and GanglBaird and Gangl 2006; Reference RamirezRamirez 2008). As with elected candidates, we argue that one personal characteristic of judges that may impact these assessments is their religious affiliation, or lack thereof. Because religion is often crucial to the development of fundamental values and opinions, it can serve as a useful heuristic to translate information about how a judge is expected to behave. Religion is also at least partially a social activity, raising the possibility for it to cause other real or perceived conflicts of interest in the judge's decisions. Furthermore, previous studies have indicated that religion can play a role in determining a judge's ruling, yet the impact of religion on public perceptions of these rulings has still been largely overlooked (Reference BlakeBlake 2012; Reference Heise and SiskHeise and Sisk 2013; Reference MyersMyers 1988; Reference YarnoldYarnold 2000). In this article, we introduce a judge's religious affiliation and disaffiliation as potential heuristics that the public may use in evaluating the legitimacy of a particular judge's decision making. We specifically assess whether judges’ religious (dis)affiliations affect perceptions of bias in their decisions related to prominent, divisive religious issues.Footnote 2
Judges’ religious beliefs are not likely to be viewed equally as suspect, however. Indeed, religion is also notably a source of intergroup suspicion and conflict, meaning religiously based heuristics may carry loaded assumptions about the behavior of the judge in question. Among other polarizing aspects of religion, ethnocentric divisions along religious lines have been blamed for fueling “culture wars” and other political and social clashes (e.g., Reference GreenGreen et al. 1996; Reference HunterHunter 1991; Reference LaymanLayman 2001). In these battles, religious identities cause individuals to sort others into members of two groups: in-groups (those who share common religious values), and out-groups (those who have dissimilar values). Importantly, this sorting impacts the way others are viewed. “Members of in-groups (until they prove otherwise) are assumed to be virtuous: friendly, cooperative, trustworthy, safe, and more. Members of out-groups (until they prove otherwise) are assumed to be the opposite: unfriendly, uncooperative, unworthy of trust, dangerous, and more” (Reference Kinder and KamKinder and Kam 2009: 8).Footnote 3
Among the religious groups most likely to translate information and conflict to the public are atheists. Studies have consistently shown atheists to be one of or the single most disliked group in society (Reference Gervais, Shariff and NorenzayanGervais, Shariff, and Norenzayan 2011; Reference HarperHarper 2007; Reference Putnam and CampbellPutnam and Campbell 2010). As Figure 1 demonstrates, atheists are viewed as unfavorable by a majority of the American public, and are the least accepted of all religious groups. According to a 2014 Pew Research Center poll, around 53 percent of Americans would be less likely to vote for a presidential candidate if he or she were identified as an atheist (Pew Research Center 2014a). Of the traits surveyed, atheism received the most negative responses: more than previous drug use, an extramarital affair, or having no prior experience in office. This lack of support for nonreligious candidates is further reinforced by another 2014 Pew Research Center survey, which found that atheists were one of the least liked religious categories in a thermometer rating (Pew Research Center 2014b). On a scale from 0 (coldest) to 100 (warmest), atheists received an average rating of 41. In contrast, Jews, Catholics, and evangelical Christians are viewed warmly by the American public (63 for Jews, 62 for Catholics, and 61 for evangelical Christians). Even among Democrats, the party widely perceived as less religious, the average rating for atheists was only 46.
Figure 1. Unfavorability of Groups in Society.
This low favorability rating is driven primarily by a mistrust of atheists, as most of the public views religion as a precondition for morality (Reference Gervais, Shariff and NorenzayanGervais, Shariff, and Norenzayan 2011). Prior research has shown that this distrust of atheists exists even among atheists themselves and people from more liberal, secular populations (Reference GervaisGervais 2014; Reference Gervais, Shariff and NorenzayanGervais, Shariff, and Norenzayan 2011). As we suggest above, trust is a significant component of perceptions of procedural justice. Therefore, atheist judges, due to their mistrusted “out-group” status, may be less likely to be viewed as upholding the requirements of procedural justice and more likely to issue policy decisions viewed as illegitimate.
Given this dislike for atheists generally, we expect to find that perceived violations of procedural justice will be intensified when originating from this societal out-group. The public will likely place higher burdens on atheist judges to exhibit procedural fairness, especially as atheists are perceived as untrustworthy and amoral. Therefore, in the proceeding analysis, we expect to observe a larger conditional effect for procedural justice concerns when the judge is an atheist, a member of an out-group, in which case even minor perceived violations of judicial norms can diminish a court's legitimacy.
Conversely, the public may be more willing to trust Christian judges, who they see as members of a societal in-group, to make procedurally fair decisions. The American public is generally much more favorable toward religious public figures than nonreligious figures. According to data collected by the Pew Research Center in 2012, the majority of Americans prefer religion to play a relatively prominent role in public life and say they are comfortable with politicians talking about their religious faith and beliefs. Furthermore, most Americans are concerned about what they see as religion's declining influence. Two-thirds of the public believe religion is losing its influence in American life, and 49 percent see this as a bad thing. Moreover, a majority of those who think religion's influence is on the rise think this is a good thing. A majority also believe that some politicians have gone too far trying to keep religion out of the government. Since Christians are more trusted by the public, citizens may be less skeptical of Christians and their ability to follow procedure and make proper decisions.
Given these findings, we expect Americans to apply different standards in their evaluations of judges who are members of societal out-groups (atheists) than societal in-groups (Christians). In instances where a judge can potentially be perceived as biased, we expect Americans to be more critical and distrusting of nonreligious judges. Thus, it is not simply the manner in which judges make decisions that affects perceived legitimacy, but also the judge's own background, especially if the judge is part of an out-group.
Data and Experimental Design
To examine the impact of judges’ personal religious characteristics on legitimacy and perceptions of judicial decision making, we employ data collected in an online survey completed by 326 respondents on November 3, 2013.Footnote 4 Subjects were recruited from the pool of paid task completers on Amazon.com's Mechanical Turk (MTurk). On average, the survey took fewer than five minutes to complete, and subjects were compensated $0.20.Footnote 5 Only adult users in the United States were permitted to participate. While MTurk does not provide a nationally representative sample of U.S. residents, it is used regularly in other disciplines (Reference Buhrmester, Kwang and GoslingBuhrmester, Kwang, and Gosling 2011) and occasionally in political science (Reference ArceneauxArceneaux 2012) to conduct survey experiments. Previous analyses have also shown MTurk users to be more diverse than those in most college student panels and other online panels (Reference Berinksy, Huber and LenzBerinksy, Huber, and Lenz 2012; Reference Buhrmester, Kwang and GoslingBuhrmester, Kwang, and Gosling 2011). Moreover, while an MTurk sample may differ from the general population, the fact that it differs is less important than the ways in which it differs. As we discuss below, the unrepresentative composition of our sample may actually provide a more conservative test of our theory.
Our survey included an embedded experiment in which respondents were randomly assigned to read an excerpt from one of four fabricated news articles describing the outcome of an Establishment Clause case in a fictional circuit court. Because the scenario is fictitious, we did not use the Supreme Court, as respondents were more likely to have knowledge of this Court. The contested issue in all four cases was the display of a Nativity scene on public property. In two of the four cases the deciding vote in the case, Judge Thomas Jones, upheld the display. In the other two cases the judge struck down the display. Additionally, the judge was identified as a Christian in one of the pro-manger cases and as an atheist in one of the anti-manger cases.Footnote 6 In one of the excerpts reporting the decision to uphold the Nativity display and in one of the excerpts reporting the decision to remove the display, the judge's religious affiliation is not mentioned. These two excerpts serve as controls without any implied religious bias, allowing us to compare perceptions of a potentially biased decision with an unbiased one.
Since our primary aim is to compare perceptions of potentially biased court decisions with unbiased decisions, we have not included excerpts in which judges rule against their own religious predisposition. That is, participants were not exposed to decisions in which an atheist upheld a Nativity display or a Christian struck down a Nativity display. These cases run counter to the stereotypical preferences of Christians and atheists, and would thus represent an additional example of unbiased decision making. As we already have an “unbiased” control group for each type of decision, these additional modified decisions would be redundant for our purposes. Furthermore, the textual changes necessary to make these excerpts realistic would render them incomparable to the other treatments in their decision group.Footnote 7
The excerpt for the pro-manger decision was as follows (religious orientation treatment in brackets):
“Ahead of the Christmas holiday, the Twelfth Circuit Court of Appeals is receiving criticism for its decision upholding the display of a Nativity scene on public property. The court, citing the First Amendment, said that the display was constitutional and did not violate the separation of church and state. In response, outraged groups are threatening to appeal the ruling to the United States Supreme Court. Judge Thomas Jones, the deciding vote, is drawing particular criticism [because he is a committed Christian].”
The excerpt for the anti-manger decision was nearly identical (differences in bold-face):
“Ahead of the Christmas holiday, the Twelfth Circuit Court of Appeals is receiving criticism for its decision striking down the display of a Nativity scene on public property. The court, citing the First Amendment, said that the display was unconstitutional and violated the separation of church and state. In response, outraged groups are threatening to appeal the ruling to the United States Supreme Court. Judge Thomas Jones, the deciding vote, is drawing particular criticism [because he is a committed atheist].”
Of the 326 participants, 84 subjects were randomly assigned to read the pro-manger control decision, 79 subjects to read the pro-manger Christian decisions, 85 subjects to read the anti-manger control decision, and 78 subjects to read the anti-manger atheist decision.
After reading the excerpts, subjects were instructed to complete a battery of questions.Footnote 8 These included four questions about their perceptions of judicial decision making. Previous studies have varied widely in their attempts to measure perceptions of court legitimacy and judicial decision making. Some have relied on the use of a single survey item, while others have utilized composite scores from a larger battery of questions (Reference Caldeira and GibsonCaldeira and Gibson 1992, Reference Caldeira and Gibson1995; Reference Scherer and CurryScherer and Curry 2010). While we conduct our analyses on responses to individual items rather than averages across a battery of questions, we also look at multiple items to distinguish between support for the particular court's decision in a specific case, support for the particular court in general, and generalized support for the judiciary more broadly.
The first question asked respondents: “To what extent do you agree or disagree with the Court's decision in this case?” (Agreement). The three subsequent questions began by asking respondents: “To what extent do you agree or disagree with the following statement?” The three statements were: “This court can generally be trusted to make decisions that are right for the country as a whole” (Trust); “Judges can put aside their personal beliefs to make objective decisions based on the law” (Objectivity); and “People should obey a court decision even if they do not agree with it” (Obedience). While previous studies have used Likert scales to gauge agreement, we employ 100-point, end-defined scales to detect differences in agreement beyond categories that are predetermined for the respondent. Each question allowed respondents to express their agreement on a 0–100 scale, with 100 representing the most positive possible evaluation on each item.
We include this range of dependent variables to distinguish between respondents’ perceptions of the specific outcome of the case (Agreement), the particular court's decision making in general (Trust), courts and judges’ decision making in general (Objectivity), and the legal process (Obedience). We ultimately expect the treatment to have no effect on respondents’ agreement with the decision, while diminishing their trust in the atheist judge's decision making. That is, we do not anticipate that individuals will change their stance on the well-debated issue of public displays of religion due to our survey treatment, but we do expect that the suggestion of biased decision making by a religious out-group judge will influence their views of that judge's competency and willingness to follow legal procedure. Although we predict that the treatment is less likely to affect generalized perceptions of courts, the latter two variables are included to test whether the treatment effect extends to these cases. In line with the theory outlined above, we thus expect respondents exposed to the anti-manger atheist decision to perceive this specific court as less trustworthy than those exposed to an anti-manger decision in which the judge is not identified as an atheist. Perceptions of bias should also be conditional on in-group/out-group status. Conversely, we do not expect individuals to respond to the implied bias in the pro-manger Christian decision. Therefore, we expect to observe no difference between the respondents exposed to the two pro-manger decisions.
In addition to our experimental treatment, we also collected a range of information about respondents’ personal characteristics via a battery of demographic questions.Footnote 9 In general, our sample is more male, more white, more educated, and less conservative than the general population. This is consistent with what previous studies have found regarding the characteristics of MTurk workers (Reference Berinksy, Huber and LenzBerinksy, Huber, and Lenz 2012). Most importantly, however, our sample is much less religious than the general population. Figure 2 shows the religious orientations of our sample. Approximately 63 percent of respondents identified with no religion or identified as atheists. As discussed above, the greatest concern with the composition of an MTurk sample is how it affects the results of our experiment. Because our hypothesis is that respondents will perceive an atheist's decision as less legitimate but not a Christian's, the sample actually provides a more conservative test. Our disproportionately nonreligious sample should bias our experiment in the opposite direction of the results we expect, because we would expect the nonreligious to hold more positive beliefs about atheists than religious respondents. For example, previous research has shown that nonreligious Americans are less likely than other religious groups to say atheists should not be allowed to teach in public schools (Reference Putnam and CampbellPutnam and Campbell 2010: 485).
Figure 2. Religious Orientations of the Sample.
To the degree that our sample affects our findings, it should cause us to underestimate the effect of judges' unpopular religious beliefs on perceptions of their decision making. Thus, while there is very little risk of a type I error (i.e., a false positive), the nonreligious bias in our sample may create a danger of a type II error (i.e., a false negative). To partially circumvent this risk, we also separately examine the treatment effects for religious respondents. Figure 3 shows the frequency of worship attendance among respondents in our sample. Consistent with their religious orientations, the majority of respondents never attend worship services. However, approximately 19 percent of respondents attend religious services several times or more per year. Because they are less likely to identify with the atheist out-group, our expectation is that this subsample of religious respondents will be even more likely to exhibit bias against out-group members and thus have lower levels of trust in the court.
Figure 3. Religious Attendance of the Sample.
Results
Table 1 displays the effect of each treatment for the overall sample. As we hypothesized, identifying a judge as an atheist or a Christian has no observable effect on agreement with either the pro-manger or anti-manger court decision (Agreement); as a whole, respondents are no more or less likely to agree with a decision in this specific case based on the judge's affiliation with a religious out-group. Consistent with our expectations, however, the anti-manger treatment does have a negative effect on respondents’ trust in the particular court referenced in the excerpt to make decisions (Trust). Conversely, the implication of bias by a Christian in the pro-manger treatment results in no decrease in trust. Thus, the results confirm our suspicion that out-groups are more likely to be perceived as untrustworthy or biased decision makers.
Table 1. Effect of Treatment on Perceptions of Decision
Note: Entries are OLS regression coefficients. Standard errors are in parentheses.
* p < 0.05, two-tailed.
Notably, however, the treatment in both decisions has no impact on respondents’ generalized belief in the objectivity of judges (Objectivity) or commitment to obeying unpopular legal decisions (Obedience). This encouragingly suggests that the impact is not far-reaching enough to erode generalized trust in courts or the legal system as a whole. This is somewhat expected, as the court in question is a single circuit court with limited visibility and influence. While perceived bias in a higher court or multiple courts might produce a broader effect on public perceptions, our results indicate that the presence of atheists—even when viewed as untrustworthy—on the bench is insufficient cause to doubt the entire legal system.
Nevertheless, the trust placed in a particular court to make decisions is central to a court's long-term legitimacy. The U.S. legal system necessitates a certain level of public faith in its courts. Therefore, reductions in public trust hurt courts’ ability to serve as adjudicating mechanisms. Figure 4 shows the mean level of trust in the court for each group with confidence intervals around the values.Footnote 10 Although a difference of nine points is a relatively mild effect, previous research has observed a relatively stable, minimum level of legitimacy that is resistant to perceived conflicts of interests (Reference Gibson and CaldeiraGibson and Caldeira 2013). Given the stability of belief in judicial legitimacy, any effect is relatively strong evidence in support of our hypothesis and might serve to erode faith in the court.
Figure 4. Predicted Trust in the Court.
Because Agreement does not differ between the control and treatment groups, we are able to compare members of the treatment group with members of the control group of a similar level of agreement. Interestingly, within the anti-manger control and treatment groups, identifying the judge as an atheist negatively affects Trust even for those in the highest quartile range of Agreement. That is, even those who strongly agree with the judge's decision to strike down the display of a Nativity scene on public property are less likely to trust the judge to make decisions when he is identified as an atheist (by a difference of about 12 points, p < 0.05).Footnote 11 In contrast, for those in the highest quartile of Agreement within the pro-manger treatment and control groups, there is no significant difference in Trust. This suggests that out-group status can elicit distrust even among those who agree with a judge's decision making.
Additionally, it is important to recall the over-representation of the nonreligious in our general sample. That we observe an anti-atheist effect with such an overwhelmingly nonreligious sample is compelling evidence in support of our hypothesis. Similarly, we might expect the nonreligious to be most likely to respond negatively to Christian judges, yet we see no negative effect. This is somewhat consistent with previous findings that the nonreligious are relatively tolerant of religious fundamentalists compared with the intolerance expressed by the highly religious toward atheists (Reference Putnam and CampbellPutnam and Campbell 2010: 485).
However, the presence of numerous nonreligious respondents in our main sample is likely causing us to underestimate our treatment effects, given national demographics of religious affiliation. Therefore, Table 2 and Figure 5 show the treatment effects for religious respondents (i.e., those respondents who attend worship services several times or more per year). As expected, we observe much larger effects among the subset of religious respondents than for the overall sample. Comparing within the respondents who read excerpts covering anti-manger decisions, religious respondents were approximately 20 points less likely to trust the court with an atheist judge to make decisions that are right for the country as a whole (Trust). As expected, the religious respondents express no additional distrust toward the potential biased Christian judge in the pro-manger decision.
Figure 5. Predicted Trust Among Church Attendees.
Table 2. Effect of Treatment on Church Attendees
Note: Entries are OLS regression coefficients. Standard errors are in parentheses.
* p < 0.05, two-tailed.
Conclusion
While our findings are restricted to court decision making on the issue of public Nativity displays, the importance of this issue should not be minimized. Despite Supreme Court rulings in Lynch v. Donnelly (1984) and County of Allegheny v. ACLU (1989), public Nativity displays remain controversial. On November 19, 2012, a United States district judge ruled against a Christian group that sought to install a Nativity display on public land in the city of Santa Monica, CA during the month leading up to the Christmas holiday. “The atheists won and they will always win,” William Becker, the attorney for the Christian group, opined in response to the decision (Reference CurryCurry 2012). Becker went on to compare the judge to Pontius Pilate, further illustrating the potential impact of judicial decisions on their perceived legitimacy (Reference GrovesGroves 2012).
Furthermore, we have shown that respondents do critically evaluate decision making on this issue. When exposed to anti-manger decisions attributed to an atheist judge, respondents express significantly less faith in the court than when the judge's religious affiliation is not identified. Importantly, however, when a Christian judge rules in favor of a Nativity display, the public is no more critical. Respondents react negatively only when they perceive a systematic bias in favor of atheists, not when the bias favors Christians. This is the case despite a largely nonreligious sample, which suggests that a judge's identification with a societal out-group influences the perceptions of even those who are not personally opposed to the judge's views (although we do observe larger effects for the religious, suggesting that personal views also matter). This means that secularization may offer little in the way of a solution. As both political institutions and the general population become increasingly nonreligious, biases against atheists in the nonreligious public may persist.
Although the public's perceptions of courts do not typically receive a great deal of focus from mass public opinion scholars, we argue here that the public's attitudes toward these unelected decision makers also warrant further research. The role courts play in resolving deeply divisive political issues occasionally places them under substantial public scrutiny, and this scrutiny is prone to many of the same biases that infect the public's thinking across an array of other political issues. Because much of a court's authority lies in the public's deference, it is important to understand the factors that contribute to the erosion of the public's faith in the court.
Moreover, our findings have important implications for the study of judicial politics. Substantial attention has been given to the potential impact of judges’ personal preferences and attributes on their decisions, but few have examined how those attributes affect the public's perceptions of judges’ decision making. We have shown that judges who are part of a societal out-group may face greater public scrutiny. As a component of procedural justice theory, trust of individual judges is highly important for viewing the court as fair and unbiased. Given this expectation, atheist judges, and judges from other out-groups in society, are likely to face continued bias against their perceived competency and decision making. We show that the public is concerned not only about appropriate legal procedure, but also about the personal beliefs and attributes of those making the legal decisions. Therefore, as judges face elections or nomination processes, out-group biases may serve to further disadvantage unpopular minority groups. As the pool of nominees for public office becomes less religious (mirroring secularization trends in the general population), it will become increasingly important to understand how nonreligious biases influence perceptions of law and legal decision making.
Although our results strongly suggest that members of the public demonstrate a bias toward religious out-groups, it remains to be seen whether these findings generalize to other kinds of religious out-groups (e.g., Muslims), religious out-groups in cases not dealing with religious issues, and other kinds of out-groups (e.g., based on race, class, sexual orientation, or political views). While we are unable to test each of these cases with existing data, the next stage of this research should further delineate the effects of case type: whether the religious identity of judges will impact perceptions of decision making in cases lacking a religious component. Moreover, as we observe no effect for the Christian treatment, it is also possible that these results are limited only to certain societal out-groups. While we suspect that out-group status has at least a minimal effect on perceptions of court decision making in other instances, it is possible that the effect for atheists—a particularly unpopular religious out-group—is unique. This warrants further investigation. Future studies may also benefit from experimentally varying the in-group to include different denominations of Christianity or matching to the respondent's own religious identity.
A further limitation of this study is the fact that judges’ religious affiliations are not frequently publicized. Indeed, procedural justice concerns in cases of religion may be mitigated by the fact that religious affiliation is not physically identifiable as other characteristics may be, such as race or gender. Conversely, this information is occasionally available on judge's Web sites, and may become prominent in highly politicized cases (such as the Burwell v. Hobby Lobby case discussed previously). Just as the public has more or less accurate perceptions of how judges make decisions, members of the public will also likely make assumptions about the religious affiliation of judges. Previous research has shown that the public dramatically underestimates the percentage of the population that is Protestant in the United States, while overestimating the percentage of Mormons, Muslims, atheists, and the nonreligious (Grey Matter Research 2012). This suggests that the public may also overestimate the number of atheists serving as judges and perceive procedural justice violations.
Although we examine perceptions of judicial decision making, further research may well establish that out-group biases also apply to the perceived legitimacy of how other officials and nongovernmental employees make decisions. Given the public's hesitance to vote for atheists for elected offices (Reference Gervais, Shariff and NorenzayanGervais, Shariff, and Norenzayan 2011), it seems likely that religion is used as a heuristic for trust in one's abilities to fairly carry out the procedures required of the position. In particular, tying how the media frames decision making to characteristics of decision makers seems to be a fruitful avenue for further research on the legitimacy of the institutions of government (Reference RamirezRamirez 2008).
Ultimately, however, these results demonstrate that a judge's religious preferences can have tangible implications for his or her decision-making process and case outcomes. Because judges have incentives to care about the legitimacy of their decisions (Reference ClarkClark 2009; Reference Epstein and KnightEpstein and Knight 1998; Reference FallonFallon 2005), it is possible that judges are willing to strategically shift their behavior to avoid the appearance of bias. In particular, members of out-groups may have an incentive to either not reveal their true preferences or to actually vote counter to their preferences for the sake of legitimacy (Reference Heise and SiskHeise and Sisk 2013; Reference SunsteinSunstein 2007; Reference YarnoldYarnold 2000). Conversely, using religion as a tool for judicial decision making may itself be problematic. Therefore, empirical research on courts should take greater account of the role of religion as both a component of judicial decision making and a factor in the public's evaluation of courts.
Appendix
Table A1 Demographic Characteristics of Groups
Table A2 Effect of Religious Attributes on Trust for Highest Agreement Quartile
Note: Entries are OLS regression coefficients. Standard errors are in parentheses.
* p < 0.05, two-tailed.
