Causes and effects of groupthink bias on judges of international courts
by Philipp Günther
The discussion surrounding the recent judgment of the German Federal Constitutional Court on the Public Sector Purchase Programme (PSPP), in which it declared the decision of the Court of Justice of the European Union to be ultra vires, demonstrates that when judges hand down a controversial decision, commentators will examine not only the merits of the judgment but also the ulterior political motives of the individual judges that might have influenced the decision at hand. However, the unidimensional focus on the judges’ ideological preferences overshadows a myriad of other factors which may have a more significant effect on the decision-making process of judges, such as group composition or conformity effects.
One of the most common assumptions about decision-making is that groups usually enhance the quality of a discussion and the subsequent outcome. If group members talk to each other and exchange competing views, they are able to learn and make better decisions, thereby avoiding the mistakes one single member would typically make. Does this suggest, however, that groups are always better at making decisions because the pool of available information and arguments is enlarged? Research in social psychology has shown for decades that groups often fail to live up to their potential, and even act foolishly in high-stakes situations, because they can enhance mistakes made by the individual group members. One possible explanation for the missed potential of collective intelligence is the so-called groupthink theory. The theory of groupthink has its roots in social psychology and was first described by Irving Janis in 1972. Janis defined groupthink as “a mode of thinking that people engage in when they are deeply involved in a cohesive in-group, when the members’ strivings for unanimity override their motivation to realistically appraise alternative courses of action.” The theory posits that the supposedly neutral process of decision-making in structured groups can be influenced by mental shortcuts, cognitive illusions and internal group dynamics. Those extraneous factors often lead to excessive concurrence-seeking, which in turn may result in self-censorship and faulty decision-making. This insight is especially relevant in the context of international courts and tribunals, since their decision-making processes almost always involves group choices. However, it is important to distinguish groupthink bias from consensual decision-making. While a consensus requires a conscious decision by the group members, groupthink occurs subconsciously. Hence, individuals are not responsible for groupthink bias, but rather it results from the context in which they form the decision. The following article investigates two main questions: (1) To what extent different international courts are prone to groupthink bias, and (2) if their decision-making process might be influenced by groupthink bias.
Measuring groupthink behaviour
While Janis laid the theoretical groundwork, his original model garnered substantial criticism because the proposed causes of groupthink behaviour were challenging to replicate under laboratory conditions. Hence, psychological scholars have tried to expand and modify Janis’ original model in such a way that the proposed causes of groupthink behaviour correspond with empirical observations. One of the most significant evolutions in groupthink theory has been the general group problem solving (GGPS) model developed by Ramon Fuller and Sally Aldag, who broadened the narrow focus of the original model. My research adopts Fuller and Aldag’s model, which examines three underlying root causes or so-called antecedents of group decision-making, and supplements their model by adding a fourth antecedent proposed by Baron. The sum score model that I utilize, thus, employs the following four antecedents: (1) group structure, (2) decision-making context, (3) decision characteristics, and (4) salient norms (i.e., emerging or dominant group norms that are produced or revealed through group interaction). These antecedents, in turn, are based on 18 variables, which are scored on a modified Likert scale from one to five. Thus, the higher an international court scores on a given variable in the model, the more likely it is exhibiting symptoms of groupthink biased behaviour.
To test the model, five of the most prominent international courts and tribunals were selected: the International Court of Justice (ICJ), the International Tribunal for the Law of the Sea (ITLOS), the Appellate Body of the World Trade Organization (WTO AB), the Grand Chamber of the European Court of Human Rights (ECtHR GC), and the Appeals Chamber of the International Criminal Court (ICC AC). The results of the GGPS analysis indicate that the WTO AB and ICC AC are most likely susceptible to groupthink bias while the other courts – especially the ICJ and ITLOS – are relatively resistant. The WTO AB obtained the highest scores in the model by far, meaning that its decision-making process appear to be influenced by groupthink bias.
Fuller and Aldag stress that the antecedents of groupthink behaviour do not necessarily lead to faulty decision-making. According to the GGPS model, a group must not only be susceptible to groupthink bias, but it also must exhibit one or more emergent group processes before it can be assumed that groupthink behaviour actually occurred. It is important to note that the employed research design is inherently limited due to the subject of study. Fuller and Aldag propose a plethora of different symptoms which should become observable in a groupthink biased decision-making group that is susceptible to groupthink bias. However, since the decision-making process of international courts is not directly observable, the present model is only able to rely on certain empirical observations to deduce that group processes have indeed occurred. In addition, it is almost impossible to derive the presence of groupthink behaviour from the language of their rulings directly. Nevertheless, there is one proxy capable of capturing the emergent group processes related to treatment of dissenters, self-censorship, and use of mindguards: the occurrence of separate or dissenting opinions. Judges of those courts are conscious of the associated social costs of dissenting and are incentivised to remain silent, although their opinion diverges from the majority. However, it is imperative to note that while the lack of dissents is treated as a groupthink symptom, this does not automatically make the occurrence rate of dissents the dependent variable. The occurrence rate of separate opinions is used as a proxy since the actual symptoms that occur during deliberations are not directly measurable.
At each of the five tribunals studied, the possibility to dissent is provided by the respective procedural rules. Consequently, if a tribunal received a high sum score in the application of the GGPS model, this may be viewed as a predictor that judges are less likely to write an individual opinion that deviates from the majority view. The empirical data on separate opinions back up this prediction. For instance, while the occurrence rate of separate opinions at the WTO AB – which scored highest in the model – is extremely low, decisions of the ICJ or the ITLOS – which scored lowest – almost always come with a dissenting opinion. These empirical results demonstrate an inverse correlation between the presence (or absence) of antecedents and the presence (or absence) of dissenting opinions. Since the WTO AB has received a particular high sum score under the GPPS model, and emergent group processes are indirectly observable through the absence of dissenting opinions, it is plausible to assume that groupthink behaviour could influence the respective judges as well as the decision-making process in general.
The Appellate Body of the WTO: Paragon of cooperative decision-making?
The results of the GGPS analysis suggest that the WTO AB is the judicial body that is most prone to groupthink biased behaviour. But what makes the decision-making process at the WTO AB so particular? In comparison to other international courts, groupthink behaviour seems to have a significant effect on influencing the decision-making characteristics of the WTO AB, because it reinforces and is reinforced by the search for unanimity prescribed by existing procedural norms. James Bacchus, one of the founding members of the AB who later became chairperson, also shared this overall sense of consensuality: “Whatever our individual role may be in any particular appeal, each of us strives always to reach a ‘consensus’ in every appeal. We are not required to do so. The treaty does not prohibit dissents.” He goes on to mention that “time after time, around our table, we have, after exhaustive mutual effort, made seven minds into one.” This sentiment ultimately allowed groupthink effects to impact the decision-making process by limiting the development of alternative proposals during the deliberative process. If there is no preferred solution during much of the deliberation, groupthink behaviour will most likely not occur, since alternative opinions are given sufficient space. Before each appeal, the AB Secretariat prepares a detailed memo that outlines the merits of the case and proposes desirable courses of action. WTO AB Members are more likely to agree with the Secretariat’s opinion because the relevant arguments are already laid out for them, and the associated social costs of dissenting may be considerable. As Meredith Lewis recounts, “Members continue debating until the last holdout either comes to agree with the majority or decides to drop his or her opposition in the interests of obtaining a unanimous decision.”
This mode of deliberation stands in stark contrast to the decision-making process of those courts that were least prone to groupthink bias: the ICJ and the ITLOS. Both have formal internal judicial procedures which regulate their deliberative process in meticulous detail. More importantly, judges will usually have to prepare a so-called “note” individually, wherein they express their preferred solution regarding the issue at hand. In the case of the ICJ, only after the circulation of those notes to the other judges, do they meet in conference and discuss the salient issues. The preparation of individual notes under Article 4 of the ICJ Resolution is akin to the decision-making process of the modern U.S. Supreme Court, where the different judges act more like competing law firms than a collective that repeatedly talks through the issue until they reach a common solution. Since judges have already spent a considerable amount of time to draft their favoured opinion, they are less likely to agree with alternative views that emerge during deliberations. Former ICJ President Stephen Schwebel further elaborated on this tendency: “Each judge must deliver; he has to write his note and say what he thinks as to the issues of the case. The down side of this procedure is that once a judge has concentrated all of his [sic] efforts for some weeks in figuring out how the questions of the case should be answered, too often he proves reluctant to rethink his thoughts when he is confronted with those of his colleagues.” Schwebel implies that judges will not easily change their minds during the decision-making process because the process of preparing notes has already entrenched their views. In turn, the generation of contingent courses of action cannot be stifled during deliberations, because judges can rely on their well-prepared notes during arguments. In contrast, such a formalised process of deliberation and decision-making was not foreseen for the WTO AB, where members of the assigned division as well as the other AB members are involved from the beginning of the appeal. Under these conditions, it is easier to convince members during the decision-making, since they have not prepared an individual note. Unlike ICJ judges, the AB members meet and discuss until all controversial issues have been resolved. However, those continuous meetings enhance groupthink effects. As a result, AB members are incentivised to refrain from offering new alternatives during deliberations.
The signal versus noise in groupthink biased behaviour
Does groupthink have a significant impact on the decision-making process of international courts and tribunals? Or does it merely describe certain phenomena of decision-making that accompany the antecedent conditions? In short, does the groupthink model confuse significant signals with irrelevant background noise that occurs during the deliberative process?
Research in the context of corporate actors has shown that groupthink behaviour may be moderated by emphasizing specific procedural rules that regulate the decision-making process. Such findings suggest that it might be possible to reduce groupthink bias in the context of decision-making by courts and tribunals, since these bodies are subject to a high level of detailed rules on deliberation and voting procedures. These characteristics, in turn, prompt the question whether groupthink can even develop in such an environment. One might be tempted to argue that the effects of groupthink should be minimal in international courts or tribunals, if they exist at all. Nevertheless, it would be premature to ignore all of the insights from decades of social psychology research showing that groupthink behaviour occurs among groups with structured decision-making processes. The crux of the matter is that unlike other groups, the deliberative process of international judges will most likely always be cloaked in secrecy. Furthermore, judges will generally be unaware of the subconscious effects of groupthink behaviour. Thus, the potential for groupthink bias is always present in a context where its effects remain opaque to the decision-makers. Judges will consciously pursue different goals during the decision-making process while sub-consciously being influenced by cognitive heuristics such as groupthink bias. A member of the bench contemplating a decision would hardly notice that his or her decision was in part due to groupthink. In fact, much like any interpretative method can be used by a judge to justify a preferred outcome, every conscious consideration can, in turn, be utilized to rationalise subconscious biases.
Moreover, groupthink theory in general, and the GGPS model in particular, are capable of elucidating certain empirical phenomena that other theories of judicial decision-making fail to explain adequately. For instance, the present analysis provides a more compelling explanation for the relative absence of AB dissents than the institutional “life-cycle hypothesis” theory, which predicts that more mature courts have accumulated legitimacy and are therefore willing to display disunity and issue separate opinions more frequently, compared to a relatively nascent tribunal. In the case of the AB, the “life-cycle hypothesis” theory posits that the acceptance and legitimacy of the institution should increase over time and that, in turn, the issuance rate of dissents should rise. However, the facts do not support this argument. On the contrary, the rate of separate opinions issued by AB members has been mostly stagnant since the inception of the institution, and a slight decrease in the occurrence rate is evident during the last decade. The GGPS model anticipates provides a better explanation for this result, since it argues that groupthink effects occur within each AB generation regardless of the members’ personal incentives or the increased legitimacy of the institution.
Even though it is arduous to capture groupthink behaviour reliably, the present analysis has shown that it is possible to draw some conclusions based on the results of the GGPS analysis. Groupthink, like other cognitive biases, most likely occurs in the decision-making process of the examined international courts and tribunals to some extent. Some courts are likely to be more affected by those heuristics (WTO AB, ICC AC), while others prove to be generally resistant (ICJ, ITLOS, ECtHR GC). While there may be some short-term benefits of groupthink such as increased legitimacy for the institution, this positive effect is outweighed by the negative consequences such as increased peer pressure and the limited input of opposing opinions. It is difficult to ascertain precisely how groupthink bias affects the deliberative process and which decisions have been affected by it. As long as it is impossible to conduct laboratory examinations of the actual deliberations of international courts, whereby all other independent variables are controlled, it will be futile to declare some behaviour as an obvious example of groupthink bias. While this outcome may seem unsatisfactory, it fits neatly within the decades of social psychology research that asserts that motivations are of a fundamentally ambiguous nature and cannot be retraced to a single instigating factor.
Groupthink theory alone is not adequate to provide a comprehensive framework for judicial decision-making. The conclusions that the theory provides are small jigsaw pieces in the bigger picture of the decision-making process of judges. While it contests some established assumptions, it does not purport to have an explanatory monopoly. In the context of international adjudication, groupthink theory is more valuable in furnishing additional perspectives by applying a model to groups which were traditionally thought to be largely unaffected by cognitive biases. The insights provided by groupthink theory should thus help to illuminate that judges make poor decisions under certain circumstances, just like any other group.
Philipp Günther, LL.M. (Univ. Amsterdam) is a law student at the Free University of Berlin and works as a student assistant at Michael Wrase’s research group at the WZB Berlin.
 Irving L. Janis, Victims of Groupthink: A psychological study of foreign-policy decisions and fiascoes (Houghton Mifflin 1972), p. 9.
 Ramon Aldag and Sally Fuller, ‘Beyond Fiasco: A Reappraisal of the Groupthink Phenomenon and a New Model of Group Decision Process’ (1993) 113 Psychological Bulletin, p. 533.
 Robert Baron, ‘So Right It’s Wrong: Groupthink and the Ubiquitous Nature of Polarized Group Decision-Making’ in Mark Zanna (ed), Advances in Experimental Social Psychology, Volume 37 (Elsevier Academic Press 2005).
 James Bacchus, ‘Table Talk: Around the Table of the Appellate Body of the World Trade Organization (2002) 35 Vanderbilt Journal of Transnational Law, pp. 1021, 1029.
 Meredith Kolsky Lewis, ‘The Lack of Dissent in WTO Dispute Settlement’ (2006) 9 Journal of International Economic Law, pp. 895, 917.
 Stephen Schwebel, ‘The Docket and Decisionmaking Process of the International Court of Justice’ (1990) 13 Suffolk Transnational Law Journal 543
 Christoph Engel, ‘The Behaviour of Corporate Actors: How much Can We Learn from the Experimental Literature?’ (2010) 6 Journal of Institutional Economics 445, 453–455.
 Yuval Shany, Assessing the Effectiveness of International Courts (OUP 2014) 145–147.