Armed with powers of constitutional review, European high courts have experienced increased political significance (Reference SweetSweet 2000, Reference Sweet1992, Reference Sweet2002, Reference Tate and VallinderTate and Vallinder 1995). The judicialization of European politics has involved an expansion of the power of courts and judges in determining public policy outcomes, mainly through judgments pertaining to individual rights' protection or the limits of legislative or executive powers (Reference Hirschl, Caldeira, Kelemen and WhittingtonHirschl 2009). Consequently, it is of increasing importance to understand the factors that influence public support or opposition to court decisions.
The extent to which justices agree or disagree on the solution of a case has often been hypothesized as influencing public support.Footnote 1 Justices and judicial scholars frequently argue that whereas unanimity should boost public support for court decisions, dissent should fuel public opposition (for an overview, see Reference SalamoneSalamone 2014, Reference Zink, Spriggs and ScottZink et al. 2009). This alleged negative relationship between dissent and the courts' standing with the public has been one of the reasons why some European constitutional courts used to (or still do) prohibit the publication of dissenting opinions (Reference KelemenKelemen 2018) and that, for those courts that publish dissents, norms of consensus have tended to suppress the formalization of conflicting opinions in an effort to bolster legitimacy (e.g., Reference Hendershot, Hurwitz, Lanier and PacelleHendershot et al. 2013, Reference Walker, Epstein and DixonWalker et al. 1988).
Despite the presumed harmful effect of judicial dissent on public support, only a handful of empirical studies from the United States exist on this relationship. The results of these studies are also inconsistent. Whereas some studies show no effect of unanimity or dissent on public opinion (Reference Gibson, Caldeira and SpenceGibson et al. 2005, Reference MarshallMarshall 1987, Reference PetersonPeterson 1981), other studies suggest that unanimity does in fact bolster support (Reference Zink, Spriggs and ScottZink et al. 2009). In a recent study, however, Reference SalamoneSalamone (2014) breaks with the accepted notions in the literature and argues that dissents may help increase support of issues of higher salience among the court's policy opponents by suggesting evidence of procedural justice. Hence, the dynamics of how dissent might influence public support appear contingent upon individuals' preexisting attitudes toward the issues at stake.
This article expands upon the existing research on the relationship between judicial dissent and popular support through a series of original survey-based experiments using a nationally representative sample of Norway's population. The Norwegian Supreme Court offers a relevant and testable European case on the relationship between courts and their citizenry. Unlike several of its European counterparts, the Supreme Court has made individual justice opinions and dissents open to the public since 1863 (Reference MestadMestad 2015, Reference ØstlidØstlid 1955). Today, one in five decisions from the court involves one or more dissenting opinions. In addition, during the last few decades, the Supreme Court has emerged as a more prominent and consequential policy maker in the Norwegian judicial and political system (e.g., Reference Grendstad, Shaffer and WaltenburgGrendstad et al. 2015). We should thus expect the Court to have a greater ability to shape public opinion on issues of national interest. Understanding the potential role dissents play in shaping public opinion and support is an important aspect in this regard.
This article examines the relationship between dissent and public support through a series of vignette experiments covering three fictitious Supreme Court decisions on issues that represent a range of legal and political salience among the Norwegian public: (1) a high-salience asylum decision concerning residence permits for children of illegal immigrants, (2) a medium-salience workplace-privacy decision concerning an employer's right to read email and text messages on phones provided by the company, and (3) a low-salience decision concerning whether neighbor conflicts should be resolved by means of mediation or through litigation in the court system. The vignettes that presented the outcomes of each of the fictitious decisions were read by the participants, and information about whether the decisions were unanimous or dissensual was randomized. In addition, the experiments were designed to capture how policy supporters and policy opponents evaluate unanimity and dissent, as well as to account for an individual's diffuse support for the Supreme Court.
Breaking with accepted notions in the literature, the results of the experimental vignettes show no significant influence of unanimous decisions on public acceptance of court decisions. Instead, the results demonstrate that dissent leads to greater acceptance of the high-salience asylum decision by suggesting evidence of procedural justice. This observed effect of dissent on levels of acceptance is significant across both ex ante supporters and ex ante opponents, which is contrary to Reference SalamoneSalamone (2014) who shows that dissent primarily has a positive influence on support among ex ante opponents. However, the finding that dissent has a more general effect on public acceptance is consistent with expectations in existing literature arguing that individuals' support for policy outputs is rooted in their commitments to democratic values and processes (e.g., Reference Caldeira and GibsonCaldeira and Gibson 1992, Reference GibsonGibson 2007, Reference Gibson and NelsonGibson and Nelson 2015). Finally, the article demonstrates a substantial influence of diffuse support on public acceptance across the three court decisions, which underscores the importance for Supreme Courts to maintain a “reservoir” of diffuse support or legitimacy if they are to gain acceptance for decisions that are unpopular among segments of the population.
In the following sections, the article first discusses the widespread concerns in the literature about the public policy consequences of dissents while also making a case for why we instead should expect dissents to have a positive influence on public support for court decisions. This article then presents the experimental design and data generation process. Finally, the results of the vignette experiments are discussed, and the conclusion summarizes the article's main contributions and offers suggestions for future studies.
Judicial Dissent and Public Support for Court Decisions
When high courts review laws, justices must sometimes confront disagreements inherent in the legislative context on issues over which legislators and society may be deeply divided. However, because justices are usually appointed and not elected, they confront these issues without being electorally accountable. Consequently, scholars often argue that courts, perhaps more than the elected branches of government, require a reservoir of legitimacy and public support to gain acceptance for unpopular decisions (see, e.g., Reference Gibson and NelsonGibson and Nelson 2014, Reference Gibson, Caldeira and SpenceGibson et al. 2005).
Several studies have demonstrated the number of ways in which courts may influence public opinion and support (see, e.g., Reference Clawson, Kegler and WaltenburgClawson et al. 2001, Reference Franklin and KosakiFranklin and Kosaki 1989, Reference Gibson and CaldeiraGibson and Caldeira 2003, Reference Grosskopf and MondakGrosskopf and Mondak 1998, Reference Hoekstra and SegalHoekstra and Segal 1996, Reference Johnson and MartinJohnson and Martin 1998, Reference MondakMondak, 1990, Reference Mondak1991, Reference Mondak1992, Reference Stoutenborough, Haider-Markel and AllenStoutenborough et al. 2006). According to this literature, court decisions, particularly on issues of national interest, may have an impact on public opinion, which can manifest itself either as increased support or increased opposition to the position taken by the court. However, there has been little research on the role of judicial dissent in influencing mass opinion. This is puzzling, given that justices and scholars in the United States and elsewhere have frequently debated whether dissents are good or bad for public support.
On the one hand, judicial scholars frequently express concern about the public opinion consequences of handing down dissensual decisions (see, e.g., Reference Hettinger, Lindquist and MartinekHettinger et al. 2006, Reference SalamoneSalamone 2014, Reference Walker, Epstein and DixonWahlbeck et al. 1988, Reference Wahlbeck, Spriggs and Maltzman1999, Reference Zink, Spriggs and ScottZink et al. 2009). A common argument is that whereas the absence of dissent may promote the perception that the law is applied in a uniform and impartial manner, the occurrence of dissent may “shake public confidence in the judiciary by bringing into question the certainty of the law” (Reference Walker, Epstein and DixonWalker et al. 1988, 387). This alleged negative relationship between dissent and public confidence echoes the widely held view that the legitimacy of high courts is contingent upon the extent to which the public perceives them to be apolitical and neutral adjudicators of the law (see, e.g., Reference Gibson and NelsonGibson and Nelson 2017, Reference Scheb and LyonsScheb and Lyons 2001). According to these views, dissent should have a generally harmful effect on public support, and this should be true regardless of whether individuals hold a favorable or unfavorable attitude toward the courts' policy output.
On the other hand, dissent may help democratize a court by bringing different viewpoints to the deliberative process (see, e.g., Reference BrennanBrennan 1985, Reference DubéDubé 2000, Reference PetersonPeterson 1981). Contrary to conventional views on the harm of dissent, such a democratizing effect of dissents might lead to greater support for dissensual decisions. This argument rests on studies that have demonstrated that legitimacy and support for courts are rooted in commitment to democratic values and processes (Reference Caldeira and GibsonCaldeira and Gibson 1992, Reference GibsonGibson 2007, Reference Gibson and NelsonGibson and Nelson 2015). In addition, studies in the social psychology of law and legal compliance suggest that perceptions of procedural justice and the fairness of the resolution process are central to our evaluation of judicial outcomes (see, e.g., Reference TylerTyler 1988, Reference Tyler2006, Reference Tyler and RasinskiTyler and Rasinski 1991). Accordingly, when individuals see evidence of debate in court—as shown by dissenting opinions—they might “interpret it as the result of a fair, democratic decision-making process in which both sides were heard” (Reference SalamoneSalamone 2014, 322). This positive influence of dissent on acceptance might be particularly prevalent among the “losers” of a court decision. Whereas “winners” will ordinarily accept a court decision with which they agree (Reference Gibson and NelsonGibson and Nelson 2014), losers should be more inclined to make evaluations of procedural fairness when they decide whether to support a decision or not (Reference Lind and TylerLind and Tyler 1988, 111).
Of course, whether unanimity or dissent can influence public support is not expected to be uniform across all issues. The extent to which courts can influence individuals' acceptance of a decision by means of unanimity and dissent should depend on the salience of the issue at hand (see Reference SalamoneSalamone 2014, Reference Zink, Spriggs and ScottZink et al. 2009). For example, studies of public evaluations of U.S. Supreme Court decisions demonstrate that the Court is less able to influence opinion when a policy is of high ideological salience (Reference Brickman and PetersonBrickman and Peterson 2006, Reference Grosskopf and MondakGrosskopf and Mondak 1998, Reference Hoekstra and SegalHoekstra and Segal 1996, Reference Johnson and MartinJohnson and Martin 1998, Reference MondakMondak 1990, Reference Mondak1991, Reference Mondak1992). This finding is supported by studies in social psychology (e.g., Reference Petty, Cacioppo and BerkowitzPetty and Cacioppo 1986) and political behavior (e.g., Reference ZallerZaller 1992), demonstrating that individuals are more difficult to persuade when they hold strong (or crystallized) opinions about a topic but are persuadable when they are moderately (or to a lesser degree) engaged with an issue.
On this basis, unanimity and dissent should have little influence on individuals' evaluations of court outputs on highly salient (“hot-button”) issues on which they hold strong and crystallized opinions. However, on moderately salient issues, unanimity and dissent might invoke a reaction consistent with any of the expectations discussed above. Because of the moderate levels of salience, opinions on these issues should not be so crystallized in the minds of the public that they cannot be moved. However, at the same time, people should feel enough attachment to these issues so that it might matter for them if the decision is made unanimously in their (dis)favor or whether a dissenting minority is seen to represent their policy views. Finally, courts should quite easily be able to move opinion on issues of very low salience, as individuals should be quite willing to defer to perceived experts on matters in which they have not invested much thought. However, as people will be less politically invested in these issues, we should not expect dissent to invoke any strong feelings of procedural justice (see Reference SalamoneSalamone 2014, Reference Zink, Spriggs and ScottZink et al. 2009).
Previous Studies of the Public Opinion Consequences of Dissent
Despite widespread concerns about the public policy consequences of dissent, the first few studies on this topic found little evidence of this effect. In a review of the literature on the causes and consequences of judicial dissent in the Unites States, Reference PetersonPeterson (1981) noted that dissent might weaken a court's authority, but he argued that this was unlikely because of the limited public awareness of the workings of the courts. A few years later, Reference MarshallMarshall (1987) conducted a study of opinion polls before and after 18 U.S. Supreme Court decisions. Among several factors believed to affect public reactions to court decisions, Marshall included a measure of whether decisions were unanimous or not. His results showed no significant effect of unanimity on public opinion. However, it should be noted that only two of the decisions in Marshall's sample were unanimous.
Two decades after Marshall's study, Reference Gibson, Caldeira and SpenceGibson et al. (2005) included dissent among several other factors in an experimental vignette study about a fictional case dealing with the handling of the Florida ballots in the wake of the 2000 presidential election. Reference MarshallUnlike Marshall (1987), the authors focused on factors affecting public acquiescence to court decisions. For the specific manipulation of dissent, the participants were given vignettes characterized either as “deeply divided along partisan lines” or as “consensual and therefore not divided by party.” Again, however, the results showed little evidence for the hypothesis that sharp divisions lead to lower acceptance of the outcomes of court decisions.
Two subsequent experimental studies would yield different (though conflicting) results. In the first experiment, Reference Zink, Spriggs and ScottZink et al. (2009) presented individuals with mock newspaper articles reporting on Supreme Court decisions in which they systematically varied unanimity and dissent and the court's upholding or overturning precedents. Reference Gibson, Caldeira and SpenceUnlike Gibson et al. (2005), the authors included a manipulation for dissent independent of partisanship (i.e., “unanimous” or “minimum winning coalition”). In addition, they conducted experiments on a range of cases on different political issues of varying levels of salience. Contrasting previous studies, they found that unanimous rulings (as opposed to dissenting ones) boost support for Supreme Court decisions at all levels of salience and among all participants, including those who disagreed with the court's policy output.
In a second experiment, Reference SalamoneSalamone (2014) sought to improve on the two former studies by examining the influence of unanimity and dissent in isolation from other factors. In addition, Salamone considered the distinction between large and small dissenting majorities by separating out decisions that were unanimous, divided eight-to-one, or divided five-to-four. The results of the study showed no effect of unanimity or dissent among those who stated that they agreed with the policy outcome of the court's decisions. However, among those who disagreed with the court's policy outcome, Salamone found that unanimity and large majorities led to higher acceptance of the low-salience contract dispute resolution and, conversely, that dissensual decisions (including both five-to-four and eight-to-one decisions) led to higher acceptance among the court's opponents in a moderately salient workplace-privacy decision by suggesting evidence of procedural justice.
Overall, recent experimental studies have demonstrated that unanimity (as opposed to dissent) might foster higher levels of acceptance of Supreme Court decisions (in particular, Reference Zink, Spriggs and ScottZink et al. 2009), though this appears to be the case primarily with issues of low salience (Reference SalamoneSalamone 2014). In addition, Reference SalamoneSalamone's (2014) study suggests that we should include the possibility that dissent may foster greater levels of support for moderately salient issues by suggesting evidence of procedural justice.
The Supreme Court of Norway and the Public
Little research exists on the relationship between dissent and public opinion and support outside the U.S. context. In this regard, the Supreme Court of Norway constitutes a relevant and testable European case. The Supreme Court began to publish individual justice opinions, including those supporting dissents, in 1863 (Reference MestadMestad 2015). The decision to publish individual opinions did not come about without debate. For over four decades, the Supreme Court fought against proposals in parliament to make opinions public. A central reason for the court's resistance was that it had serious concerns about its reputation with the public (Reference ØstlidØstlid 1955).
The number of dissents in the Supreme Court has varied over time. In the decades following World War II, the annual percentage of dissensual decisions fluctuated between 10 and 20% before reaching its lowest level in the late 1980s (Reference Grendstad, Shaffer and WaltenburgGrendstad et al. 2015, 69). Then, during the last two decades, the number of dissents increased in the 1990s, reaching a record high of 29% in 1999, before stabilizing at a rate of 20% annually in the twenty-first century. The rise in dissents in the 1990s was mainly catalyzed by a reform of the criminal procedures in 1995, which secured the Court near-complete discretionary jurisdiction over its criminal caseload (Reference BentsenBentsen 2017). Since then, the court has only accepted cases for review that are expected to be instrumental to the development of Norwegian law (Reference BentsenBentsen 2017, Reference Grendstad, Shaffer and WaltenburgGrendstad et al. 2015, Reference Skiple, Grendstad, Shaffer and WaltenburgSkiple et al. 2016, Reference SundeSunde 2015).
Citizens usually rely on the media for information about court cases (Reference Clawson, Kegler and WaltenburgClawson et al. 2001, Reference Stoutenborough, Haider-Markel and AllenStoutenborough et al. 2006, Reference ZilisZilis 2015). Historically, the Norwegian media's coverage of the Supreme Court has been irregular and sporadic (Reference Grendstad, Shaffer, Waltenburg, Davis and TarasGrendstad et al. 2017). The depoliticized nature of judicial appointments to the court has been identified as a central reason for this.Footnote 2 In addition, because the Norwegian legal system is highly trusted, the Supreme Court justices have seen little need to use the media to maintain or increase the Court's standing among its citizens (Reference Grendstad, Shaffer, Waltenburg, Davis and TarasGrendstad et al. 2017). Hence, it is reasonable to assume that the public has been unaware of the daily workings and decisions of the Supreme Court.
However, in the last few decades, the Supreme Court has emerged as a more prominent and consequential policy maker in the Norwegian judicial and political system (e.g., Reference Grendstad, Shaffer and WaltenburgGrendstad et al. 2015). Consequently, it has come under greater scrutiny by the media. In particular, in their reporting on the Supreme Court, the media has given greater attention to individual justices, and this appears especially true when referencing those justices authoring or joining dissents (Reference Grendstad, Shaffer, Waltenburg, Davis and TarasGrendstad et al. 2017, 249). Therefore, we can assume that today's Supreme Court will have a greater potential to shape public opinion on issues of national interest. Understanding the role dissents play in shaping public opinion and support is an important aspect in this regard.
Experimental Design
To examine the relationship between Supreme Court dissent and public support, I use a series of survey-based vignette experiments included in Wave 7 of the Norwegian Citizen Panel (NCP) conducted from November 1, 2016, to December 2, 2016. The NCP is a research-purpose web-based panel owned by the University of Bergen. The participants in the NCP have been recruited via random sampling from the official national population registry and are representative of both the online and offline population older than 18 years of age in Norway. In total, the data for the present experiment comprises 1085 registered participants.Footnote 3
The experimental survey covered three issue areas that represent a range of legal and political salience among the Norwegian public: a high-salience asylum decision concerning the issuing of residence permits for children of illegal immigrants that are being deported, a medium-salience workplace-privacy decision concerning an employer's right to read the emails and text messages on a phone provided by the company, and a low-salience decision concerning whether neighbor conflicts should be resolved by means of mediation or through litigation in the court system. Whereas asylum issues—particularly those involving the rights of children—have been of high salience in Norwegian public debate in the aftermath of the European “refugee crisis” of 2015 (see Reference Skiple, Gudbrandsen and GrendstadSkiple et al. 2013), the other issue areas were selected based on their relative degree of public salience. This difference in issue area and salience is important in order to address the differential influence of unanimity and dissent on public opinion discussed above.
In total, the experiment comprised two parts: (1) a pretreatment questionnaire asking the participants about their policy positions on the above three issues and about their diffuse support (or legitimacy) of the Supreme Court and (2) a series of vignettes describing fictitious Supreme Court decisions, including a posttreatment question about the participants' acceptance of the decisions. Prior to the experiment, the participants were asked to read the following short text about the Supreme Court: “You will now be asked a few questions about the Norwegian Supreme Court. The Supreme Court is the highest court in the country and last court of appeal. Ordinary cases are usually decided by a panel of five justices.” This information was provided because we may assume that several of the participants in the survey were unaware of the fact that most Supreme Court decisions are made by five-justice panels. (A complete and translated overview of the pretreatment questionnaire and of the experimental vignettes is available in the Appendix).
Pretreatment Questionnaire
Prior to treatment, all participants were first asked questions about their policy positions on the three separate issues (these questions can be found in the Appendix, Section A). To avoid priming the participants, these questions were asked together with a range of other unrelated questions in the survey (this procedure is similar to Reference SalamoneSalamone 2014). That said, we are unable to fully eliminate the possibility that individuals, by virtue of already having declared their attitudes on the issues earlier in the survey, have been put in a position of reconciling their attitudes with the position of the Supreme Court and thus chose their own position over the position of the Supreme Court. However, this type of priming should not negatively influence treatment effects, given that in a situation in which individuals are fully primed to follow their ex ante positions, we should not expect unanimity or dissent to have any substantial influence on their opinions. In other words, if individuals are primed to follow their initial policy position, the experiment should constitute a stricter test of the influence of unanimity and dissent.
The answers to the ex ante policy questions were later used to create measures indicating whether the position of a participant prior to treatment was the same as that described in the vignette randomly assigned to her (or not). Accordingly, if a participant was given a vignette of a Supreme Court decision reflecting her policy positions on the issue, she is considered an ex ante supporter of the decision. If a participant was given a vignette with a Supreme Court decision opposing her policy position, she is considered an ex ante opponent.
Table 1 shows the distributions of the participants' responses to the three policy statements. As expected, we can see that the participants are highly polarized in their position on the high-salience asylum issue, while the medium-salience workplace-privacy issue and the low-salience neighbor dispute issue produce less polarization. Looking at the proportions of participants agreeing or disagreeing, 39% agree that children of illegal immigrants should be allowed to stay in Norway if their parents are evicted, while 44% disagree with this statement. On the workplace-privacy issue, 21% agree that employers should have a right to read texts and emails on company cell phones, while 72% disagree. Finally, as many as 91% of the participants agree that neighbor conflicts ought to be resolved by mediation and not in the courts, while only 4% disagree.
Table 1. Individuals' Ex Ante Policy Positions
Note: The variables were coded on a seven-point scale: 1 = strongly agree, 2 = agree, 3 = somewhat agree, 4 = neither agree nor disagree, 5 = somewhat disagree, 6 = disagree, and 7 = strongly disagree. Data: Norwegian Citizen Panel Wave 7, 2016. N = 1085.
In addition to the questions about their policy preferences, the participants were given four questions about the Supreme Court's institutional legitimacy (the questions can be found in the Appendix, Section A). These questions are designed to capture the participant's diffuse support for the institution (Reference Caldeira and GibsonCaldeira and Gibson 1992, Reference Gibson, Caldeira and SpenceGibson et al. 2003, Reference Gibson, Caldeira and Spence2005) and are commonly used in the literature on judicial support and legitimacy (see, e.g., Reference Bartels and JohnstonBartels and Johnston 2013, Reference Gibson and CaldeiraGibson and Caldeira 1995, Reference Gibson and NelsonGibson and Nelson 2015, Reference SalamoneSalamone 2014, Reference Zink, Spriggs and ScottZink et al. 2009).Footnote 4 Whereas specific support refers to the court's decisions (i.e., what they are asked to evaluate in this experiment), diffuse support refers to a reservoir of good will or favorable attitudes that help individuals tolerate outputs that they oppose (Reference EastonEaston, 1965, Reference Easton1975, Reference Gibson, Caldeira and SpenceGibson et al. 2003, Reference Gibson and NelsonGibson and Nelson 2014). Reference Caldeira and GibsonCaldeira and Gibson (1992) conceptualize diffuse support for judicial institutions as synonymous with legitimacy. This is an important control in the analysis because we should assume that those with low faith in the institution would generally have a lower likelihood of accepting its decisions (e.g., Reference Gibson, Caldeira and SpenceGibson et al. 2005) and that this should be true regardless of whether the decision is unanimous or dissensual (Reference SalamoneSalamone 2014).
Experimental Vignettes
After completing the pretreatment questionnaire, the participants were exposed to a series of vignettes about three fictitious Supreme Court decisions on the above issues (the vignettes can be found in the Appendix, Section B).Footnote 5 The participants were randomly assigned to three groups: a baseline group in which participants were given no information about unanimity or dissents; a treatment group presented with a unanimous decision; and a treatment group presented with a dissensual three-two decision (i.e., the experiment is a 2x2 with a baseline group). In addition, the experiment randomly varied the outcome of the decision in order to control for the possibility that the results are driven by the participants' policy preferences for or against the decision.
For each vignette, the participants were asked about the degree to which they find the decision acceptable (coded 1 to 5, with larger scores indicating greater acceptance). This is in line with a common research design in the literature on courts and public opinion, which usually involves presenting individuals with a court decision with which they either agree or disagree and then ascertaining whether the characteristics of the court ruling increase the probability of an individual accepting the decision or not (e.g., Reference Gibson, Caldeira and SpenceGibson et al. 2005, Reference SalamoneSalamone 2014, Reference Zink, Spriggs and ScottZink et al. 2009).Footnote 6 The wording of the experimental vignettes was as follows (the following example is for the asylum decision):
Several immigrant children have parents who are residing illegally in Norway. Imagine that the Supreme Court [no information; unanimously; by a bare majority of three against two judges] decided that these children [should be; should not be] entitled to stay in Norway if the parents are deported. In general, to what extent do you think this would be an acceptable decision by the Supreme Court?
Analysis and Results
To examine the relationship between dissent and public support, I ran two sets of logistic regression analyses of the extent to which individuals accept or reject each of the three fictitious court decisions. Whereas the first set of analyses examines the effects of the main treatment (unanimity/dissent), the second set of analyses tests whether unanimity and dissent have a differential influence on acceptance rates among ex ante supporters and ex ante opponents.
Positive Influence of Dissent on Acceptance of High-Salience Issues
The first set of analyses examine the general influence of the treatments (unanimous/dissent) when controlling for diffuse support. For each decision, the analyses compare the effect of each treatment category to the baseline referent category for which the participants received no information regarding majority size. The results of the regressions are shown in Table 2. In addition, Figure 1 presents the change in terms of the probability of acceptance of each of the court decisions for the various treatment groups compared with the untreated baseline group.
Table 2. Dissent Has a Positive Effect on Support for Asylum Decision
* p < .05 (two tailed).
Note: Effects of unanimity and dissent on support for court decisions. Effects compared to the baseline group. Data: Norwegian Citizen Panel Wave 7, 2016.
Figure 1. Change in the Probability of Acceptance Across Treatment Groups for Each of the Three Court Decisions (95% Confidence Intervals). Data: Norwegian Citizen Panel Wave 7, 2016.
Looking at the treatment effects of unanimity and dissent, there are no negative effects of dissent on levels of acceptance for any of the three decisions. Instead, there is a positive and statistically significant effect of dissent on participants' acceptance of the high-salience asylum decision. More specifically, there is a 9% increase in the levels of acceptance of the dissensual decision when compared to the levels of support in the untreated baseline group. This finding supports the alternative theoretical expectation that dissensual decisions might lead to greater levels of acceptance by suggesting evidence of procedural justice (e.g., Reference SalamoneSalamone 2014).
The effects of the treatments on the two other issues indicate further support for this finding. There is a 4% increase in acceptance of the dissensual privacy decision and a close to 4% increase in acceptance of the dissensual neighbor conflict decision. However, the differences in acceptance rates across these treatment groups fail standard statistical significance tests. Thus, it is only on the high-salience asylum issue that dissent has any positive and generalizable influence on citizens' acceptance of the court decision.
The results show some support for the hypothesis that unanimity might foster greater acceptance of the asylum decision and the neighbor conflict decision. There is a 5% increase in acceptance of the unanimous asylum decision and a 5% increase in support for the neighbor conflict decision (when compared to the untreated baseline group). However, these differences fail statistical significance tests, and we are unable to conclude whether this is in fact the case for the public at large. Nevertheless, these findings are important insofar as they suggest that individuals' evaluations of the court's policy output do not appear to increase significantly because of unanimity.
Finally, there is a positive and statistically significant influence of diffuse support on acceptance rates across the three court decisions. Hence, when citizens are asked to evaluate Supreme Court decisions, they will have a greater propensity to accept the decision if they have a high regard for the Supreme Court as an institution from the outset. This is an important finding insofar as it confirms the theoretical assumptions and findings in the previous empirical literature that high courts rely on diffuse support to gain acceptance of their decisions (e.g., Reference Gibson, Caldeira and SpenceGibson et al. 2005, Reference Zink, Spriggs and ScottZink et al. 2009).
Ex Ante Policy Views
In the second set of analyses, I include interaction terms between the treatments and the variable capturing whether individuals are either ex ante supporters or ex ante opponents of the decision. These interactions are included to test whether unanimity or dissent have a different influence on acceptance rates among ex ante supporters and ex ante opponents. It is noted, however, that individuals' support or opposition to the decisions is not randomly assigned. While we are able to randomize whether individuals are given a favorable or unfavorable decision based on their preexisting attitudes, we are naturally unable to randomize individuals' preexisting policy positions on these issues. Consequently, a number of covariates might predict the ex ante policy views that form the basis for the agreement determination. The following analysis thus includes a number of control variables related to partisanship (i.e., whether they would vote for either a left-of-center or a right-of-center party in the next election) and sociodemographic factors (i.e., the participants' gender, education, and age), which might be related to the participants' preexisting policy positions on the three issues.
The results of the second sets of regressions are presented in Table 3. First, we see that there is a statistically significant and substantial influence of ex ante policy positions on acceptance across all three decisions. Because of the added interaction term, the effects of the ex ante opponent variable are interpreted as the difference in acceptance between ex ante opponents and ex ante supporters in the baseline group: for this treatment group, ex ante opponents have a 53% lower level of acceptance for the asylum decision when compared to ex ante supporters. For the privacy and neighbor conflict decisions, the difference in acceptance between supporters and opponents is 61% and 68%.
Table 3. No Conditional Effects of Ex Ante Policy Views
* p < .05 (two tailed).
Note: Effects of unanimity and dissent on support for court decisions conditional on ex ante policy views. Effects compared to the baseline group. Data: Norwegian Citizen Panel Wave 7, 2016.
However, although there is a large difference between ex ante supporters and ex ante opponents in their posttreatment evaluations of all three court decisions, there are no statistically significant effects of the added interaction terms between the treatments (unanimous/dissent) and the participants' ex ante policy views. Hence, the extent to which unanimity or dissent influences acceptance rates across the three issues is not conditional on whether individuals are predisposed to either support or oppose the decisions they were asked to evaluate.
Again, there is statistically significant effect of diffuse support on acceptance rates across all issues. Figure 2 shows the levels of acceptance for each decision conditional on the participants' pretreatment policy views. Among both ex ante supporters and ex ante opponents, the levels of acceptance of each court decision increases as the levels of diffuse support increases. If we look specifically at ex ante opponents, the influence of high levels of diffuse support is the strongest among those who were given the high-salience asylum decision. For this decision, the acceptance levels are almost 50% higher among those who hold the Court in the highest regard when compared to those who hold it in very low regard. Consequently, when citizens are asked to evaluate unfavorable and politically salient Supreme Court decisions, they will have a greater propensity to accept the decision if from the onset they have a high regard for the Supreme Court as an institution.
Figure 2. Effects of Diffuse Support on Ex Ante supporters' and Ex Ante opponents' Acceptance of the Asylum Decision (Upper Left Graph), the Privacy Decision (Upper Right Graph), and the Neighbor Conflict Decision (Lower Left Graph) (95% Confidence Intervals). Data: Norwegian Citizen Panel Wave 7, 2016.
Discussions and Conclusion
As European courts are experiencing growing political power (e.g., Reference SweetHirschl 2004, Sweet 1992, Reference Sweet2000, Reference Sweet2002, Reference Tate and VallinderTate and Vallinder 1995), it is of increasing importance to understand what factors influence popular support or opposition to court decisions. One important aspect in this regard concerns the public's evaluations of unanimity and dissent. However, despite widespread concerns about the public policy consequences of dissents, the literature on the relationship between dissent and public opinion is sparse and limited to the U.S. public's evaluations of Supreme Court decisions.
This article offers the first study of public reactions to judicial dissent outside the U.S. context by conducting a series of survey-based vignette experiments on a nationally representative sample of Norway's population. The vignettes covered three fictitious Supreme Court decisions on issues that represent a range of legal and political salience among the Norwegian public. In addition, the experiment was designed to capture how policy supporters and policy opponents evaluate dissensual decisions in order to test the alternative expectation that policy opponents might show greater support for a dissensual decision because they see their policy views represented in the outcome.
Breaking with accepted notions in the literature, the experimental results show no significant influence of unanimous decisions on public acceptance of court decisions. In addition, there is no evidence in the experimental data suggesting that dissent has a negative influence on people's support for court decisions. Instead, the results demonstrate that dissent leads to greater acceptance of the high-salience asylum decision by suggesting evidence of procedural justice. This observed effect of dissent on the levels of acceptance is significant across both ex ante supporters and ex ante opponents, which is contrary to the expectation that opponents (or “losers”) of a court decision should be more inclined than supporters (“winners”) to make evaluations of procedural justices when they decide to oppose a decision or not. This is also contrary to the experimental results by Reference SalamoneSalamone (2014), which shows that dissent primarily has a positive influence on support among ex ante opponents. However, that dissent has a more general effect on public acceptance is consistent with expectations that individuals' support for policy outputs is rooted in their commitments to democratic values and processes (e.g., Reference Caldeira and GibsonCaldeira and Gibson 1992, Reference GibsonGibson 2007, Reference Gibson and NelsonGibson and Nelson 2015).
Dissent only moves opinion on the high-salience asylum issue, which is contrary to the theoretical expectation that courts should have a difficult time influencing public opinion on “hot-button” issues on which individuals hold strong and crystallized opinions. There are two potential explanations for this result. First, we might assume that when the public is highly polarized on an issue, which is arguably the case concerning the rights of immigrant children in Norway, they might expect justices to be polarized on the issue as well. Evidence that both sides of the issue have been heard—as shown by the dissenting opinion—should therefore matter more when public opinion is polarized. To put it more succinctly, procedural justice might play a greater role in the evaluations of court decisions on issues on which the public is polarized, which might explain why dissent leads to greater levels of acceptance of the high-salience asylum decision. Future studies could address the relationship between issue polarization and procedural justice more directly.
The second explanation regards the relative salience of the asylum issue. Although questions about the rights of immigrant children have been of high salience in Norwegian public debate in the wake of the 2015 European “immigration crisis,” it might be considered less salient that those “hot-button” issues included in existing experimental studies in the United States (e.g., gay marriage, Florida ballots, abortion). This is especially true if we consider the decisive role the U.S. Supreme Court plays in deciding these issues. Hence, the observed effect of dissent on individuals' acceptance of the asylum issue might instead be regarded as evidence that courts should have an easier time moving public opinion on moderately salient issues on which opinions are not fully crystallized in the minds of the public.
In addition to the effects of dissent on acceptance rates, there is a substantial influence of diffuse support on public acceptance across the three court decisions. Most importantly, there is a relatively strong influence of high levels of diffuse support on acceptance rates among ex ante opponents of the high-salience asylum decisions. This is an important finding insofar as it confirms the theoretical assumptions and findings in the previous empirical literature that high courts rely on diffuse support to gain acceptance of unpopular decisions (e.g., Reference Gibson, Caldeira and SpenceGibson et al. 2005, Reference Zink, Spriggs and ScottZink et al. 2009). Moreover, this result underscores the importance for Supreme Courts to maintain a “reservoir” of diffuse support or legitimacy if they are to gain acceptance for decisions that are unpopular among segments of the population.
Finally, as this is the first empirical study outside the U.S. context to suggest that dissent may bolster public acceptance of court decisions, the constraints of this one-time experiment still leave unanswered questions about the long-term effects of unanimity and dissents. While one possible implication of this study is that courts in the long term might bolster their legitimacy by means of publishing dissents, more research on such a long-term view would need to be considered if we are to reach this conclusion.
Appendix A. Pretreatment Questionnaire
For each question, the participants were asked to what extent they agree or disagree with the statement. Their answers were measured on a scale from 1 (strongly agree) to 7 (strongly disagree).
Ex Ante Policy Positions
In general, to what extent do you disagree or agree with the following statements?
1. Immigrant children whose parents are illegal residents in Norway should be entitled to stay in the country if their parents are deported.
2. If an employer gives you a mobile phone to be used for work, the employer should be allowed to read all text messages and emails sent and received on this phone.
3. It should not be possible to bring neighbor conflicts (e.g., about the height of trees and fences) to the court without first having tried to resolve the conflict through mediation.
Diffuse Support for the Supreme Court
In general, to what extent do you disagree or agree with the following statements?
1. The Supreme Court of Norway can usually be trusted to make the right decisions.
2. The right of the Supreme Court of Norway to decide certain types of controversial issues should be reduced.
3. The decisions of the Supreme Court of Norway consistently favor some groups more than others.
4. If the Supreme Court of Norway started making many decisions most people disagreed with, it might be better to do away with the court altogether.
Appendix B. Experimental Vignettes
For each vignette, the participants were asked about their acceptance of the fictitious court decision. This was measured on a scale from 1 to 5, where higher values indicate greater levels of acceptance.
Vignette for High-Salience Decision
Several immigrant children have parents who are residing illegally in Norway. Imagine that the Supreme Court [no information, unanimously, or by a bare majority of three against two judges] decided that these children [should be or should not be] entitled to stay in Norway if their parents are deported. In general, to what extent do you think this would be an acceptable decision by the Supreme Court?
Vignette for Medium-Salience Decision
Several employers give their employees cell phones to use for work. Imagine that the Supreme Court [no information, unanimously, or by a bare majority of three against two judges] decided that employers [should be or should not be] allowed to read all the text messages and emails contained on the telephones issued to the employees by the company. In general, to what extent do you think this would be an acceptable decision by the Supreme Court?
Vignette for Low-Salience Decision
Thousands of neighbor conflicts arise in Norway every year. Imagine that the Supreme Court [no information, unanimously, or by a bare majority of three against two judges] decided that people involved in neighbor conflicts (e.g., about the height of trees and fences) [should be or should not be] allowed to go directly to the courts without first having tried to resolve the conflict through mediation. In general, to what extent do you think this would be an acceptable decision by the Supreme Court?
