Conformity

by Cass R Sunstein

  Or consider the difference between individual behavior and team behavior in the Dictator Game, used by social scientists to study selfishness and altruism.28 In this game, a subject is told that she can allocate a sum of money, say $10, between herself and some stranger. The standard economic prediction is that most subjects will keep all or almost all of the money for themselves; why should we share money with strangers? But the standard prediction is wrong. Most people choose to keep somewhere between $6 and $8 and to share the rest.29 That is interesting enough, but the question here is how individual behavior is affected once people are placed in teams.

  The answer is that team members choose still more equal divisions.30 This result seems best explained by reference to a rhetorical advantage disfavoring selfish behavior, even within a group that stands to benefit from selfishness. Apparently, people do not want to appear to be greedy in front of fellow group members. Of course this outcome, and the effect of group influence, would change if the team in the Dictator Game had some reason to be hostile to the beneficiaries of their generosity. We can easily imagine a variation of the Dictator Game in which, for example, people of a relatively poor and embattled religious group were deciding how much to allocate to another religious group that was thought to be both hostile and far wealthier. In this variation, the rhetorical advantage might favor greater selfishness.

  What produces a rhetorical advantage? The simplest answer points to prevailing social norms, which of course vary across time and place. Among most Americans, current norms make it easier to argue, other things equal, for higher penalties against corporations for egregious misconduct. But it is possible to imagine subcommunities (corporate headquarters?) in which the rhetorical advantage runs the other way. In any case it is easy to envisage many other contexts in which one or another side has an automatic rhetorical advantage.

  Consider, as possible examples, debates over whether there should be higher penalties for those convicted of drug offenses, whether more refugees should be allowed into one’s country, whether more money should be spent on national defense, or whether tax rates should be reduced. In modern political debates, those favoring higher penalties and lower taxes often have the upper hand. Of course there are limits on the feasible level of change. But when a rhetorical advantage is involved, group deliberation will produce significant shifts in individual judgments. Undoubtedly legislative behavior is affected by mechanisms of this sort, and it is likely that many movements within judicial panels can be explained in similar terms.

  Are rhetorical advantages unhelpful or damaging? In the abstract, this is an impossible question to answer, because shifts have to be evaluated on their merits. Perhaps the higher punitive awards that follow deliberation are simply better. So too, perhaps, for the movements by doctors, taking more heroic measures, and by groups deciding how equally to spread funds. The only point is that such advantages exist, and it would be most surprising if they were always benign.

  More Extremism, Less Extremism

  Group polarization is not a social constant. It can be increased or decreased, and even eliminated, by certain features of group members or their situation.

  First, extremists are especially prone to polarization. It is more probable that they will shift, and it is probable that they will shift more. When they start out at an extreme point and are placed in a group of like-minded people, they are likely to go especially far in the direction with which they started.31 There is a lesson here about the sources of terrorism and political violence in general. And because there is a link between confidence and extremism, the confidence of particular members also plays an important role; confident people are both more influential and more prone to polarization.32

  Second, if members of the group think they have a shared identity and a high degree of solidarity, there will be heightened polarization.33 One reason is that if people feel united by some factor (for example, politics or religious convictions), dissent will be dampened. If individual members tend to perceive one another as friendly, likable, and similar to them, the size and likelihood of the shift will increase.34 The existence of affective ties reduces the number of diverse arguments and also intensifies social influences on choice. One implication, noted above, is that mistakes are likely to be increased when group members are united mostly through bonds of affection and not through concentration on a particular task; it is in the former case that alternative views will be less likely to find expression. Hence people are less likely to shift if the direction advocated is being pushed by unlikable or unfriendly group members.35 A sense of “group belongingness” affects the extent of polarization.36 In the same vein, physical spacing tends to reduce polarization; a sense of common fate and intragroup similarity tend to increase it, as does the introduction of a rival “outgroup.”37

  An interesting experiment attempted to investigate the effects of group identification.38 Some subjects were given instructions in which group membership was made salient (the “group immersion” condition), whereas others were not (the “individual” condition). For example, subjects in the group immersion condition were told that their group consisted solely of first-year psychology students and that they were being tested as group members rather than as individuals. The relevant issues involved affirmative action, government subsidies for the theater, privatization of nationalized industries, and the phasing out of nuclear power plants.

  The results were striking. Polarization generally occurred. But there was the least polarization in the individual condition; polarization was far greater in the group immersion condition, when group identity was emphasized. This experiment strongly suggests that polarization is highly likely to occur, and to be most extreme, when group membership is made salient. Political activists of all kinds are often aware of the fact; so are many entrepreneurs.

  Third, over time, group polarization can be fortified because of “exit,” as members leave the group because they reject the direction in which things are heading.39 If exit is pervasive, the tendency to extremism will be greatly aggravated. The group will end up smaller, but its members will be both more like-minded and more willing to take extreme measures. That very fact will mean that internal discussions will produce more extremism still. If the strongest loyalists are the only people who stay, the group’s median member will be more extreme, and deliberation will produce increasingly extreme movements.

  Fourth, when one or more people in a group know the right answer to a factual question, the group is likely to shift in the direction of accuracy.40 If the question is how many people were on the earth in 1940, or where the Olympics were held in 2004, or the distance between Berlin and Paris, and if one or a few people know the right answer, the group is likely not to polarize but to converge on that answer. The reason is simple: the person who knows the answer will speak with confidence and authority and is likely to be convincing for that very reason.

  Of course this is not inevitable. Solomon Asch’s conformity experiments show that social pressures can lead to errors even with respect to simple factual claims. But in many cases, group members who are ignorant will be tentative, and members who are informed will speak confidently. This is enough to ensure convergence on truth rather than polarization. Here there is a link between what prevents polarization and what shatters cascades: a person who knows, and is known to know, the truth.

  In this light, it becomes easier to understand the outcomes of experiments that show a potential advantage of groups over individuals.41 One set of experiments involved two analytic tasks. The first involved a statistical problem, requiring subjects to guess the composition of an urn containing blue balls and red balls. (This experiment involved team decision-making and was not a test for cascade effects.) The other involved a problem in monetary policy, asking participants to manipulate the interest rate to steer the economy.

  People were asked to perform as individuals and in groups. The basic results for the two experiments were similar. Groups significantly o
utperformed individuals (and they did not, on balance, take longer to make decisions). Perhaps most surprisingly, there were no differences between group decisions made with a unanimity requirement and group decisions made by majority rule.

  How can these results be explained? The experimenters do not have a complete account. An obvious possibility is that each group contained one or more strong analysts, who were able to move the group in the right direction. But a series of regressions, comparing the performance of the best individual players, offers only mixed support for this hypothesis.42 It seems that in these experiments, group results were driven by the best points and arguments, which would be spread among the various individual players. Here we find a tribute to the widespread belief that groups can do much better than individuals.

  Fifth, depolarization might be found when the relevant group consists of individuals drawn equally from two extremes.43 Thus if five people who initially favor caution are put together with five people who initially favor risk-taking, the group judgment might move toward the middle. Consider a study consisting of six-member groups specifically designed to contain two subgroups (of three persons each) initially committed to opposed extremes; the effect of discussion was to produce movement toward the center.44 One reason may be the existence of information and persuasive arguments in both directions.45

  Interestingly, this study of equally opposed subgroups found the greatest depolarization with obscure matters of fact (e.g., the population of the United States in 1900)—and the least depolarization with highly visible public questions (e.g., whether capital punishment is justified). Matters of personal taste depolarized a moderate amount (e.g., preference for basketball or football, or for colors for painting a room).46 Hence “familiar and long-debated issues do not depolarize easily.”47 With respect to such issues, people are simply less likely to shift at all, in part “because the total pool of arguments has long been familiar to all,”48 and nothing new will emerge from discussion.

  These findings suggest a separate point: group members might not shift at all when they begin with strong convictions. If you put together a group of people who love Brexit with a group of people who hate Brexit, they might end up just where they started.

  Consider in this regard an experiment designed to see how group polarization might be dampened.49 The experiment involved the creation of four-person groups, which, on the basis of pretesting, were known to include equal numbers of persons on two sides of political issues (whether smoking should be banned in public places, whether sex discrimination is a thing of the past, and whether censorship of material for adults infringes on human liberties). Judgments were registered on a scale running from +4 (strong agreement) to 0 (neutral) to –4 (strong disagreement).

  In half of the cases (the “uncategorized condition”), subjects were not made aware that the group consisted of equally divided subgroups in pretests. In the other half (the “categorized condition”), subjects were told that they would find a sharp division in their group, which consisted of equally divided subgroups. They were also informed who was in which group and told that they should sit around the table so that one subgroup was on one side facing the other group. In the uncategorized condition, discussion generally led to a dramatic reduction in the mean gap between the two sides, thus producing a convergence of opinion toward the middle of the two opposing positions (a mean of 3.40 scale points, on the scale of +4 to –4).

  But things were very different in the categorized condition. Here the shift toward the median was much less pronounced. Frequently there was barely any shift at all (a mean of 1.68 scale points). In short, calling attention to group membership made people far less likely to shift in directions urged by people from different groups.

  My discussion of group influences—of conformity, cascades, and polarization—is now complete. I have emphasized many findings from social science, but I have tried at the same time to give a sense of how those findings bear on issues in law and politics. It should be clear that there is a long list of potential applications, and any set of selections from that list is inevitably arbitrary. In the discussion that follows, I emphasize four areas in which an understanding of group influences helps to illuminate legal problems.

  The first involves law’s expressive function—the circumstances in which a mere statement, made by the law, is likely to affect people’s behavior. I draw a link among legal pronouncements, Milgram’s experimenter, and Asch’s unanimous confederates. The second involves the institutions of the U.S. Constitution, based on the founding enthusiasm for the expression of diverse and dissenting views. I suggest that the U.S. Constitution creates a deliberative democracy of a distinctive kind—a deliberative democracy that prizes heterogeneity.

  The third area involves the value of dissent in a place not always thought to benefit from it: the federal judiciary. Because judges are subject to conformity and cascade effects as well as group polarization, it is exceedingly important to promote ideological diversity within the federal courts. The fourth and final area involves affirmative action in higher education. Focusing generally on the significance of cognitive diversity, I offer an ambivalent lesson, suggesting that racial diversity is, in some domains, unimportant for the exchange of (relevant) ideas, but that it can be important in other domains, sometimes in both undergraduate and law school education.

  Chapter 4

  Law and Institutions

  Many people have been interested in law’s expressive function—in the role of law in “making statements,” as opposed to regulating conduct directly through actual punishments for violations.1 In this chapter, I make three suggestions. First, we can better understand the expressive function of law if we see certain legal enactments as offering signals about good behavior and about what other people think is good behavior. Second, a legal expression is most likely to be effective if violations of the law are highly visible; visibility matters because people do not want to incur the wrath of others. Third, a legal expression is less likely to be effective if violators are part of a deviant subcommunity that rewards, or at least does not punish, noncompliance. In such cases, behavior within the subcommunity can counter-act the effects of law.

  Each of these points can be closely connected with an understanding of conformity, cascades, and group polarization. We can thus use that understanding to see when government might bring about compliance without relying on public enforcement—and also when enforcement is likely to be indispensable.

  Law as Signal

  Sometimes law is infrequently enforced, but there is automatic or near-automatic compliance.2 It is in this sense that law seems to have an expressive function, making statements and having effects merely by virtue of those statements. When such effects occur, it is because the law offers signals on both the informational and the reputational sides. If law is made by sensible people, and if it bans certain conduct, there is a good reason to presume that the conduct should be banned. And when law bans certain conduct, there is good reason to presume that other people think the conduct should be banned. In either case, sensible people have fresh reason to do what the law asks them to do.

  Of course the presumptions can be rebutted. Informed people might know that the law is asking people to do something senseless or not to do something sensible. They might also know that most people, or most relevant people, actually reject the law. But if these cases are the exception rather than the rule, we can have a better understanding of why law will produce movement even if no one is enforcing it.

  Consider, for example, an empirical study of bans on smoking in public places.3 The simplest lesson is that people comply with those bans even though they are hardly ever enforced. The study finds that in three cities in California—Berkeley, Richmond, and Oakland—the authorities heard very few complaints about violations. In Berkeley, the responsible health department officials found it unnecessary to issue even a single formal citation, and no cases were referred for prosecution. In restaurants in Richmond,
compliance was nearly 100 percent, with workplace compliance hovering between 75 and 85 percent. The level of compliance was also extremely high in Oakland, with the exception of “certain restaurants in the Asian community where nearly all the patrons are smokers.”4 High levels of compliance were also found in workplaces, high schools, and fast-food restaurants. Other studies, conducted in Cambridge, Massachusetts, and Winnipeg, Manitoba, similarly find that bans on public smoking are almost entirely self-enforcing.5

  This evidence suggests that a legal pronouncement can have the same effect as Solomon Asch’s unanimous confederates. When a law bans smoking in public places, the pronouncement carries information to the effect that it is wrong, all things considered, to smoke in public places. Equally important, the law suggests that most people believe it is wrong to smoke in public places. And if most people think it is wrong to smoke in public places, would-be smokers are less likely to smoke, in part because they do not want to be criticized or reprimanded. Importantly, would-be smokers also know that those who would reprimand them would have the law on their side. They would not merely be confronted by people who dislike being around smokers. Those who confront them would be able to say that smokers are violating the law.

  It follows that when law is effective but unenforced, an important reason is the possibility of private enforcement. If violations have a high degree of visibility, and if violators risk the wrath of private enforcers, compliance is likely to become widespread. “In contrast to violations of laws against driving and drinking, narcotics use, and tax evasion, infractions of no-smoking rules in public places are relatively visible . . . to an almost omnipresent army of self-interested, highly motivated private enforcement agents—nonsmokers who resist exposure to tobacco smoke.”6 In some cases, the law might even be equivalent to Stanley Milgram’s experimenter, with a significant degree of authority even if no sanctions will be imposed. To the extent that the experimenter’s authority comes from a perception of knowledge and expertise, the law is closely analogous.