Individual Programmatic Assessment: Exploring a Classic Study in Social Psychology

CHAPTER 13 SOCIAL INFLUENCE: DOING WHAT OTHERS DO AND SAY

· Rob a bank.

· Run naked in public.

· Cheat on an exam.

· Voice some of my weirder thoughts to strangers.

· Not pay my bills.

· Scream in a crowded place.

· Kill someone.

· Stop working.

· Rescue children from parents who yell at them in public.

What would you do if you could do anything at all and get away with it, with no social repercussions? My students over the years have offered various answers, some of which I've just paraphrased. Some forbidden impulses are aggressive, immoral, or illegal; some are harmless but weird; and some are even prosocial in intent. What they all have in common is the chance to deviate from social norms, to ignore social conventions. In actual practice, when my colleagues and I have assigned social psychology students actually to do something deviant (excluding the dangerous, harmful, or illegal), they decided to:

· Face the back of a crowded elevator.

· Stage a screaming personal argument in the middle of class.

· Wear their clothes backward.

· Drape themselves with holiday lights (plugged-in) during class.

Professors fantasize about doing counternormative behavior also. One professor (Meiss, 1997) suggested activities for teaching the first day of class that included donning a feather boa, or wearing a hood and gurgling, or announcing a CPR dummy as the TA, or screaming group chants.

Most of the time, of course, professors do not wear feather boas, scream at their students, or make gurgling noises, and students do not wear holiday lights to class. Why? For the same reasons that explain the example in  Chapter 1 , when students in my classes make their syllabus into a paper airplane on the first day: simply because an authority asks and peers comply.

In the first chapter, we defined social psychology essentially as the influence of people on other people. As such, social influence research is the hallmark of social psychology. Not surprisingly, all the core social motives enter prominently into social influence. Nevertheless, we will see that social influence research has been more specific than that. After defining social influence conceptually and operationally, we will address how the core motives relate to social influence. The main part of the chapter examines classic topics in social influence: conformity with peers, obedience to authorities, and compliance with requests. But each area brings in more recent approaches, such as the roles of group identity and minority influence in conformity, the role of power in obedience, and the role of ingroups in compliance. In every case, as well, culture matters to who influences whom, when, and how. The focus, however, is always on social relationships and social context, and we will see that group identity most often provides the relevant context, which is why this chapter comes after the groups chapter. You will also see that this chapter on social influence recaps many ideas covered in previous chapters, precisely because social influence caps the study of social psychology, serving as our penultimate chapter in this journey.

WHAT IS SOCIAL INFLUENCE?

Conceptual Definitions

Social influence broadly encompasses any changes in beliefs, attitudes, or behavior that result from interpersonal interaction. More narrowly, social influence focuses on changes mainly in behavior resulting directly from interpersonal interaction. Social influence typically refers to these interpersonal processes that change other people's behavior (Chaiken, Wood, & Eagly, 1996; Cialdini & Goldstein, 2004; Turner, 1995). As such, social influence processes focus on norms and roles in rich, interactive settings. Social influence is so basic that many social species display it, for example in the socially transmitted local feeding norms of wild vervet monkeys (van de Waal, Borgeaud, & Whiten, 2013) and humpback whales (Allen, Weinrich, Hoppitt, & Rendell, 2013).

In contrast to human attitude change via persuasion, the influence agent may or may not intend to influence the other. Notice also the emphasis on behavior and on direct interaction, which sets social influence apart from attitude change. This type of influence focuses more on behavior than on private thoughts and feelings. In contrast, persuasion theories of attitude change ( Chapter 6 ) focus on deliberate advocacy to change opinions, focusing on intentional messages.

Some key concepts aid our analysis of social influence. First is norms, which we've encountered in many previous chapters, as the unwritten rules for behavior. Second, the sections of this chapter separately address conformity, the influence of the majority on individual behavior; obedience, the influence of authority demands on subordinates; and compliance, the influence of a peer's request.

Operational Definitions

CONFORMITY

In studying conformity, social psychologists build on work by Muzafer Sherif (1936) on the autokinetic effect and Solomon Asch (1955) on line judgments. Recall from  Chapter 1  that Sherif's groups developed norms to interpret the optical illusion of an isolated point of light seeming to move in a dark space. Recall also from  Chapter 1  that Asch's participants sometimes went along with the group's erroneous judgment of line lengths, despite the evidence of their eyes. Conformity is such a basic process that measures can include even brain signatures for socially conforming memories of an event (Edelson, Sharot, Dolan, & Dudai, 2011).

Conformity research has long used group settings, sometimes with confederates giving prearranged answers. It typically has measured agreement with an erroneous or arbitrary norm (e.g., Abrams, Wetherell, Cochrane, Hogg, & Turner, 1990). Clearly, conformity is more obvious when people under group pressure answer in ways they would not ordinarily. Control groups demonstrate this difference, the effect of group pressure, by having confederates absent or not giving wrong answers. In some field studies, the confederates are absent, but evidence of other people's behavior appears, for example, in the amount of litter left behind (recall Cialdini et al.'s studies from  Chapter 9 ). This stretches the conceptual definition of social influence as behavior change resulting from social interaction, until you remember the actual, imagined, or implied presence of others defined as “social.”

OBEDIENCE

How much do you agree with this statement, attributed to Karl Marx? “Those who hold and those who are without property have ever formed two distinct classes.” Early research on the influence of authorities began with the work on prestige suggestion, in which the same statement is judged differently, depending on whether it is attributed to a derogated authority (for Americans in the 1940s, Karl Marx) or to its rightful author, John Adams (Asch, 1948).

Research on reactions to authorities as obedience per se began with Milgram's classic (1964) study, probably the most famous study in social psychology. As  Chapter 10  implied, the aggression machine that participants use in their role as teachers to punish the errors of the student (confederate) originated simultaneously in aggression and obedience research. We will describe the Milgram study in more detail later.

COMPLIANCE

Compliance research focuses on people's attempt to get someone else to do something, and an early example includes the Freedman and Fraser (1966) foot-in-the-door technique. In this case, some people were asked to comply with a trivial request (place a tiny be-a-safe-driver sign in their front window). These people later complied with a large unreasonable request (install a DRIVE CAREFULLY billboard on their front lawn), more than people who had not acceded to the seemingly harmless initial request. In dozens of subsequent studies, the key ingredient was face-to-face interaction and a small initial request.

All told, the operational definitions of social influence rank as some of people's favorite studies in social psychology because they involve realistic interpersonal situations, apparently tiny manipulations of social context, and meaningful changes in behavior.

Core Social Motives: Belonging, Understanding, Controlling, Self-enhancing, and Trusting

Much of social influence fits a framework of maintaining one's group identity and harmony within the group. As such, social influence results significantly from the core social motive of belonging. However, this is a relatively new approach to social influence.

For many years, social psychologists assumed that the primary impetus for social influence was to avoid the uncertainty brought about by one's apparent disagreement with others. Researchers differentiated two kinds of uncertainty-reducing motivation for social influence. Informational influence was described as trying to get at objective reality (Deutsch & Gerard, 1955). As such, it would fit the core motive to develop a shared understanding. The other traditional kind of motivation was normative influence, sometimes described as trying to elicit approval from others (Turner, 1995). As such, it could fit the idea of gaining social rewards and avoiding social punishments, consistent with the core social motive to control one's outcomes. But normative influence could be broader; it could also fit the need to enhance the self, maintain self-esteem, and avoid anxiety or guilt (Chaiken et al., 1996). Or it could be viewed as a version of belonging, when people go along with the group norms in order to fit in. As we will see, the mix of motives probably is no longer so usefully split between informational and normative influences, but the background is useful to understand the framing of the original research problems as “objective reality” versus “social reality.”

Finally, I would argue for the importance of the core social motive to trust others in the ingroup as benevolent. Allowing oneself to be influenced by others in the ingroup, assuming their influence to be benign, is separate from merely trying to maintain harmony through belonging. Instead, it implies that others are a trustworthy source of influence.

Sometimes social influence is broken down into three motives that overlap with the ones noted: mere compliance in the service of gaining rewards and avoiding punishments (like controlling); conformity with valued reference groups (like belonging); and internalization of valid information about reality (like understanding) (Kelman, 1958; note that the use of compliance and conformity retains the flavor of our earlier definitions but uses the terms more narrowly). Mentioning this social influence triad serves to remind us that social psychologists have long been aware of people's varied motivations for going along with social influence.

Summary of Definitions and Motives

Social influence could be all of social psychology (as we defined it: how the thoughts, feelings, and behaviors of individuals are influenced by the actual, imagined, or implied presence of others). More often, it is changes in behavior that result from social interaction. Operationally, the research tends to fall into methods for studying conformity with peers, obedience to authorities, and compliance with direct requests. All the core social motives are relevant to social influence, as we will see.

CONFORMITY: BELONGING AND UNDERSTANDING BY DOING WHAT OTHERS DO

Conformity, going along with others, reflects adhering to group norms for behavior. We start with the classic studies, some already described, address more recent research on the processes involved, and end the section by taking seriously the role of group belonging in social influence. We also discuss contemporary myths and legends, in the context of conformity to shared understanding of group norms.

Classic Studies: Sherif and Asch

Participants sat in a darkened room, along with others, voicing their judgments about an apparently moving point of light. Sherif (1935) wanted to demonstrate the impact of norms on individual judgment, so he used an ambiguous task in which all answers (other than zero) were arbitrary and wrong. In later research, other participants reacted to statements such as the one about social classes and property, noted earlier, or this one about rebellion, attributed to Lenin or Jefferson: “I hold that a little rebellion, now and then, is a good thing, and as necessary in the political world as storms are in the physical.” Asch (1948) found that readers changed the meaning of “rebellion” to “agitation” when the author was Jefferson and “rebellion” to “revolution” when the alleged author was Lenin. They then rated their agreement with the statement accordingly. In keeping with his Gestalt approach (see  Chapter 4 ), Asch (1940, p. 458) famously noted that the difference was a “change in the object of judgment” (agitation or revolution), rather than a “change in the judgment of the object” (rebellion as intrinsically good or bad).

Subsequently, still other participants sat, along with others, voicing their judgments about the comparative length of three lines to one standard line (see  Figure 1.1 ). Asch (1956) wanted to show the limitations on conformity, so he chose a less ambiguous judgment task than either illusory movement or ambiguous words, expecting little or no social influence. Nevertheless, participants made errors about a third of the time (compared with controls judging alone, who made practically no errors).

This simple task set up a cottage industry of replications and extensions (R. Bond & Smith, 1996). As of 1994, about 40 years later, 94 studies had used Asch's paradigm and another 39 had used a modified paradigm in which participants sit in booths, hearing false feedback about others' unanimous but wrong responses (Crutchfield, 1955). That's an impressive rate of more than three new studies per year for four decades. In 133 studies, the average participant conformed 29% of the time, according to the meta-analysis. Conformity was greater when researchers (a) most closely replicated Asch's original paradigm, (b) used ambiguous stimuli (as reflected by a high error rate), and (c) used ingroup members as the influential minority, all for obvious reasons. Furthermore, although conformity occurs across all 17 cultures in which it has been tested, it does occur (d) more strongly in relatively collectivist cultures. This seems to be a function of individualist cultures' emphasis on individual autonomy, both affective and intellectual, which fits other patterns of cultural variation. For example, East Asians tend to encounter social influence attempts that focus on conformity, portrayed as harmony and connectedness, in advertising and consumer preferences. Americans, in contrast, tend to encounter social influence attempts that focus on independence, freedom, and uniqueness, again as reflected in advertising and consumer preferences (Kim & Markus, 1999). These cultural tendencies for social influence make sense, given everything we have seen so far.

Less obvious are the reasons for larger conformity effects according to two unrelated factors (R. Bond & Smith, 1996): The first anomaly is more conformity in studies that included women as participants. Other research has found that both women and men each are more open to social influence in judgments that do not reflect their own gender's expertise (Eagly & Carli, 1981); perhaps the female participants in the Asch paradigm were more likely to mistrust their own perception of comparative line measurement because they viewed it as outside their own expertise. Own perceived competence does correlate with resisting conformity (Allen, 1965).

The second anomaly is more conformity in those studies that took place at an earlier date. Apart from stereotypes about the conforming 1950s, the nonconforming 1960s to 1970s, and the individualistic 1980s, it is hard to know how to interpret this result. The more important point is how robust the fundamental conformity effect really is across a variety of settings.

Arguably, Asch's work focused on a stripped-down form of social influence without real interaction. His method took an individualist view of groups, rather than a more social interactionist view of groups. That is, he focused on the participants as individuals within a group rather than on group process per se. None of his groups actually interacted, and he focused on individual independence rather than group members' interdependence (Leyens & Corneille, 1999). Asch believed that people maintain independence mainly because they want the group to achieve a correct consensus, so they cooperatively abandon their own position if they personally seem to be wrong (Levine, 1999). One of the reasons for viewing the conformity studies as the proper topic of this chapter on social influence, rather than the previous one on groups, is that Asch's groups were not very groupy.

Nevertheless, Asch (1940) recognized the difference between the influence of a congenial group (which people accept) and an antagonistic group (which people reject). Note also, for future reference, how much the conformity effect depends on ingroup members as the influential majority. What's more, this holds in all cultures (M. H. Bond & Smith, 1996). An early review of the conformity studies noted that the participant's attraction and similarity to, as well as interdependence with, the erroneous majority readily predicted how much people conformed (Allen, 1965). This hints at the crucial role of group identity.

Asch believed that groups help people to construe the meaning of ambiguous phenomena, and then they conform accordingly (Levine, 1999). This suggests a causal sequence whereby groups affect initial perception. However, sometimes people conform and then change the meaning of the object in order to justify their conformity (Buehler & Griffin, 1994; Griffin & Buehler, 1993). This suggests a causal sequence whereby groups affect judgment, which then affects reinterpretation of the stimulus. What's the evidence: Does influence occur during initial perception, automatically, or only post hoc, after conformity, perhaps with more thought?

Conformity Processes: Fairly Automatic

Both processes of conformity apparently occur. In a distinction reminiscent of dual-process models of social cognition ( Chapter 4 ), persuasion ( Chapter 6 ), and bias ( Chapter 11 ), one process is relatively automatic and the other more controlled kinds of processes. Consider the following example of automatic conformity to social norms.

Participants viewed photographs of a library reading room or a train station interior after learning they would have to visit the setting afterward (Aarts & Dijksterhuis, 2003). Still other participants viewed the library without the goal of going there. Presumably, when the goal of going to a social environment is active, the relevant norms are activated to cue conformity. And, indeed, only those people primed with the goal to visit the library did more rapidly recognize words related to silence (silent, quiet, still, whisper) than unrelated control words. In a subsequent study, participants primed in the same way also spoke more softly, even on an unrelated task. Parallel effects occurred for participants primed with the goal of visiting an exclusive restaurant and behaving in a well-mannered way. Primed participants ate crackers during unrelated experimental tasks and actually tidied up more often if they had been primed. This was especially true for people who personally had demonstrated a previously strong associative strength between the exclusive restaurant and being well mannered.

Relying on norms may be a rapid (if not exactly automatic) heuristic strategy that people use when they do not have time, capacity, or inclination to use more effortful processes (Wood, 2000). Otherwise unmotivated participants conform to the consensus indicated by opinion poll results, regardless of the poll's reliability, demonstrating a heuristic form of conformity (Darke, Chaiken, Bohner, Einwiller, Erb, & Hazlewood, 1998). Similarly, people will conform to match the positivity of another person's self-presentation, apparently unaware that they have done so (Vorauer & Miller, 1997), consistent with a relatively automatic process. Even if participants are motivated (by incentives for accuracy), they still use heuristic processing when the task becomes too difficult. In this case, people conform at high levels (i.e., even more than when the task is not difficult; Baron, Vandello, & Brunsman, 1996). People can be automatically influenced, too, simply by the friends they keep; the interdependence of two friends' experiences means that if one person likes, for example, a restaurant, then the other person will come along to try it as well, creating a spillover of the first person's preference, even apart from persuasion processes, simply because we oversample our friends' favorite experiences (Denrell & Le Mens, 2007).

Indirectly, people's relatively automatic tendency to think others agree with them may reflect a kind of unthinking conformity. That is, the false consensus effect (Ross, Greene, & House, 1977) shows that people tend to think that other people would do as they have done. The most famous example is the original study in which some students agreed to walk around campus wearing a sandwich board that said “Eat at Joe's.” Students who agreed thought that others also would agree, whereas students who declined thought others would decline. Many years earlier, students who admitted to cheating on exams similarly expected others to cheat (Katz & Allport, 1931). The false consensus effect is well established, according to meta-analysis (Mullen et al., 1985). Most people think they are in the majority (Krueger & Clement, 1997), so even minorities overestimate the degree of consensus with their position. Although mostly believing they are in the majority, majorities sometimes underestimate the degree of consensus (Gross & Miller, 1997). People in general assume that most others share their opinions, a phenomenon also called social projection, which appears to be an unthinking or even automatic process (Krueger, 1998). The false consensus and social projection effects do have limits: People rapidly infer that only the ingroup will agree with them, not that the outgroup will. This reinforces the importance of conformity to ingroups.

Minority Influence: Another Process

As suggested here and in  Chapter 12 , people conform to the ingroup and assume that the ingroup agrees with them. Moscovici viewed the emphasis on conformity to the ingroup as a tyranny of the majority (Martin & Hewstone, 2001; Wood, 2000). In good contrarian fashion, he set out to demonstrate the conditions under which people actually do conform to an outgroup. He emphasized not only an outgroup but a minority outgroup (Moscovici, 1980, 1985). As introduced in  Chapter 12 , minority influence—the impact of a smaller subgroup that disagrees with the majority—comes through conflict. The minority's opposing viewpoint confronts the majority viewpoint and creates conflict within the individual and between individuals in the two subgroups.

Minority viewpoints create influence by becoming salient and demanding attention because of conflict. Recall that persistent minorities provoke others to think more creatively about their positions (Nemeth, 1986), a process called divergent thinking. Majority influence (conformity) may be relatively automatic and unthinking, as suggested earlier, but minority influence in contrast may require more thought and even a reflective conversion.

Minorities influence only when they behave consistently, according to meta-analysis (as noted in  Chapter 12 ; Wood, Lundgren, Ouellette, Busceme, & Blackstone, 1994). The importance of consistency stems from the Kelley attribution theory ( Chapter 3 ). When someone consistently adopts a unique response, that indicates low consensus. The response focuses on a specific attitude object (high distinctiveness). And when the person does so consistently, observers infer a dispositional cause, that is, a sincere belief.

A sincere minority belief that opposes the norm requires some creative thought in order to resolve its contradiction of prevailing views (Maass, West, & Cialdini, 1987). Some would argue that minority and extreme positions will be more thoughtful and sophisticated because people have to defend them more often (Sidanius, 1988; Zdaniuk & Levine, 1996). Others argue that ingroup minorities in particular receive more attention. For example, in one study (Alvaro & Crano, 1996), participants read an antigay message attributed to an obscure group at their own university (ingroup minority), the same group at another university (outgroup minority), or their own student body (ingroup majority). Of these three sources, when the message was counterattitudinal for a particular recipient, the ingroup minority's message received the most positive thoughts, the most accurate memory, and the least negative evaluations (see  Table 13.1 ). In effect, the ingroup minority receives the benefit of the doubt—a leniency advantage (Crano, 2000), as noted in  Chapter 12 —but one that entails thoughtful consideration of the ingroup minority position.

TABLE 13.1 Leniency for Ingroup Minority Opinion in Counterattitudinal Communications

	Message Source
Measure	Majority	Ingroup Minority	Outgroup Minority
Positive thoughts	1.16	2.18	.82
Memory errors	2.05	.89	2.13
Negative evaluations of source	16.90	13.61	13.83

Source: From Alvaro & Crano, 1996. Copyright © British Psychological Society.

In contrast to the dual-process idea that conformity to the majority happens unreflectively, but that minority influence requires creative thinking, some theorists have proposed a single mechanism for both kinds of social influence. For example, social impact theory (Latané, 1981; introduced in  Chapter 12 ) argues that any kind of influence, whether majority conformity or minority influence, results from the same factors. In the immediate social force field (a concept borrowed from field founder Kurt Lewin;  Chapter 1 ), influence results from the number, strength, and immediacy of influence sources. Whereas the majority may have greater numbers, they may not have greater strength of opinion or immediacy. A related social influence model (Tanford & Penrod, 1984) argues that influence results similarly from the number of influence sources but adds source consistency, as well as the number of influence targets, to predict conformity and minority influence.

A third single-mode approach, the judgmental process model (Kruglanski & Mackie, 1990), argues for one process, based on minorities and majorities showing many similar moderators of influence. Both increase their influence in similar ways, with increases in, for example: credible sources, ingroup sources, consistent sources, target bias toward compatible conclusions, similarity to target position, and focus on source versus message. But the model also notes that minority and majority influence processes typically differ by extremity of position, target's need for closure, congruence with the prevailing view, status rebellion, respectability, salience, accessibility, and conflict. All these typical differences in minority and majority influence suggest that, while they share some common features of social influence, they also often operate by different processes.

Minority influence occurs most in private, that is, on measures that protect the converted majority individuals from appearing publicly to abandon their majority position (Wood et al., 1994). For the same reason, influence is often indirect, emerging on issues merely related to the contentious issues, or delayed beyond the immediate context (Crano, 2000). Thus, majorities can be converted by minorities, but majority individuals do not admit it to others, and perhaps not to themselves, thereby avoiding public identification with the unpopular minority position. And indeed, when people lose majority status in a group, they are less attracted to the group and expect less positive interaction (Prislin, Limbert, & Bauer, 2000), as  Chapter 12  noted.

Group identity is key. When people identify strongly with the majority, they do indeed conform to it and resist minority influence. But when they identify with the minority, they will shift to align with them instead (Wood, Pool, Leck, & Purvis, 1996).

Self-categorization Theory: Conforming to Social Reality

We have already seen several instances of group identification shaping social influence by both minorities and majorities (i.e., conformity). The importance of the ingroup in social influence reaches its apex in the self-categorization theory interpretations of social influence. Implied in the traditional work on conformity is relatively automatic, mindless error. However, people may thoughtfully conform to the social reality that is accurate, at least from the most relevant perspective, namely, their ingroup.

Instead of automatically doing what others do, perhaps people know what to do by knowing who they are (Abrams et al., 1990). In three studies, each used an established paradigm but also manipulated ingroup identity. Ingroup identity mattered every time. In the Sherif autokinetic paradigm, the impact of the confederates on norm formation decreased as their outgroup membership became more salient. In the Asch conformity paradigm, ingroup (but not outgroup) confederates increased conformity pressures. In a group polarization paradigm, distinct subgroups (i.e., mutual outgroups) inhibited opinion convergence. In each case, the ingroup defined reality, whether in perceiving the movement of light, the length of lines, or the shift in opinions. As we saw in  Chapter 12 , groups punish deviance from the ingroup norm. But groups also reward pronorm deviance from the modal group opinion—that is, someone who conforms to the ideal group prototype, which is not necessarily the actual group average (Abrams et al., 2000; Hogg, 2001).

Self-categorization theory defines social influence this way: the processes by which people agree about appropriate behavior, as defined by ingroup norms (Turner, 1991). Given that social norms guide people's actual behavior, then in this view, social influence phenomena such as conformity and minority influence entail changing behavior to fit ingroup consensus about what is appropriate. As such, error is an irrelevant concept because norms describe actual similarities among ingroup members. Norms reduce subjective uncertainty about ambiguous situations, creating subjective validity—confidence in what is appropriate, correct, and desirable. Subjective validity clearly applies to conditions described earlier as normative influence.

But does subjective validity describe informational influence? How can social reality be objective reality? Informational influence researchers always assumed objective, physical reality testing as the gold standard. Gradations in accuracy presumably involved valid arguments about objective reality, assessed via in-depth processing (as in the central, systematic, deliberate processing of dual-mode attitude theories,  Chapter 6 ). In contrast, a social identity approach argues that normative versus informational influence constitutes a false dichotomy. Individuals test reality by validating cognitive, perceptual, and behavioral information. This reality testing provides subjectivity validity only if the individual believes that relevant others would have achieved the same results of the reality test. If the individual is the only one who can obtain the test result, then the individual is unique, deviant, and idiosyncratic, and the person's social standing is threatened. Moreover, the person's sense of self is threatened (why am I the only one who can see that six-foot hairy spider coming down the street?). Information is intrinsically social, in the social identity theory (SIT)/self-categorization theory (SCT) view. This interpretation also fits our use of the core motive for a socially shared understanding, as an offshoot of the basic motive to belong.

In this view, people will be more open to influence the more they identify with and see themselves as similar to the group. Correspondingly, the group will have more influence the more it is cohesive, consensual, consistent, and distinctive. As noted, ingroups are more influential, as are members who exemplify its prototype (Abrams et al., 1990, 2000; Allen, 1965; Clark & Maass, 1988a, 1988b; Mackie, 1986, 1987; Mackie & Cooper, 1984; Martin, 1988a, 1988b, 1988c; van Knippenberg & Wilke, 1988; Wood et al., 1996). When people identify with the ingroup, they especially conform to norms that define the group (Reicher, 1984a, 1984b; Spears, Lea, & Lee, 1990; Wilder & Shapiro, 1984). What's more, the group identity as a context and the individual behavior influence each other, so it's a two-way street (Postmes, Haslam, & Swaab, 2005). All these principles fit an SCT/SIT account of conformity processes, and they broaden the context for interpreting these effects. The framework provides new vitality to this classic area of research.

Memes, Modern Myths, Rumors, and Gossip

The research covered in this conformity section all pertains to norms, but a closely related term has surfaced recently. Some researchers have adopted the term meme (Dawkins, 1976), which is supposed to be a social analog to the gene but instead pertains to units of culture. For example, urban legends (modern myths) such as disgusting stories about pets in microwaves or razor blades in Halloween candy—never shown to have actually occurred—nonetheless circulate, and the more disgusting, the more popular the legend (Heath, Bell, & Sternberg, 2001). Rumor and gossip both reflect similar transmission of social information, important to group identity. But they both tend to be verbal, whereas norms tend to be behavioral, at least as most often studied so far. Still, contemporary legends can control behavior, as when parents either forbid their children to go trick-or-treating or scrutinize the candy they bring home.

A rumor—defined as a word-of-mouth belief without secure evidence—shows an important form of peer social influence. Older research on the psychology of rumor (Allport & Postman, 1946; Festinger et al., 1948; Knapp, 1944) noted that rumors surface under ambiguous circumstances when people seek meaning to explain emotionally important topics. Later work more precisely indicated that rumors are generated and transmitted when people are anxious, uncertain, credulous, and involved (Rosnow, 1980, 1991). As rumors spread, Allport and Postman posited that they level (eliminating some details), sharpen(emphasizing others), and assimilate. As anyone knows who has played “telephone,” distortions occur at each stage of verbal transmission, and phonetic research supports this experience (Tiffany & Bennett, 1968). Initial tellers of information make less extreme judgments than listeners who receive the information secondhand (Gilovich, 1987; Inman, Reichl, & Baron, 1993), suggesting a process similar to Allport and Postman's leveling and sharpening. This seems to occur in part when listeners do not pay careful attention, so they miss important mitigating information (Baron, David, Brunsman, & Inman, 1997). As we saw in  Chapter 3 , when people first judge another person's behavior, without correcting for mitigating circumstances, they are more likely to make an extreme dispositional attribution. Rumors and gossip (rumors specifically describing people) are justifiably notorious for being merciless with regard to other people's reputations.

Summary of Conformity

Conformity, the influence of the majority on individual behavior, began with studies of perceptual suggestion, group norms, and group pressure. Although people show much independence, it was the degree of conformity in the face of apparently objective reality that captured the attention of researchers. People do conform to the ingroup fairly automatically, to the extent that they identify with it, feel similar, and are attracted to it. This occurs in all cultures, although more so in collectivist ones. When minorities do influence people, they do it by being persistent and consistent, and the influence appears most on private or indirect responses. People are open to minority influence—as with majority influence—when they identify with the minority. Self-categorization and social identity theories explain conformity as a process of reducing uncertainty with the aid of appropriate social reality. Some of the more unusual cases of reducing uncertainty occur in the social transmission of urban legends or rumors. The joint roles of understanding and belonging are clear throughout this work.

OBEDIENCE: BELONGING, CONTROLLING, TRUSTING, AND UNDERSTANDING BY DOING WHAT OTHERS SAY

Earlier, we defined obedience as the influence of authority demands on subordinates. In discussing obedience, we examine the notorious Milgram studies, which illustrate social forces influencing the individual, and power in general, which especially illustrates controlling resources and maintaining relationships as methods of influence.

Social Forces: Milgram

The single most famous study of obedience—and arguably the best-known study in social psychology—remains hugely impactful. Following his mentor, Solomon Asch, Milgram designed a simple but elegant paradigm for studying obedience to malevolent authority (Elms, 1995; Milgram, 1963, 1965). He used the very structure of the experiment to create an authority who could demand obedience. The experimenter welcomed two community participants to a laboratory at Yale University. The experimenter described a learning experiment in which the teacher would punish the learner for errors by using an electric shock machine (described earlier as the Buss aggression machine,  Chapter 10 ). The participants, typically both middle-aged men, drew slips of paper to determine their roles, and the teacher's first job was to strap the learner's wrists to the chair arms, so the electrodes would not accidentally fall off. The teacher received a sample shock—an unpleasant 45 volts—and then went to the adjoining room to begin teaching the learner a series of word pairs. The learner started out well but progressively made more mistakes, each one requiring the teacher progressively to administer 15 volts more shock. From his room, the learner began to protest the shocks, kicking the wall in apparently intolerable pain and eventually ceasing to answer. In response to any hesitation on the teacher's part, the experimenter, dressed in a gray technician's coat and holding an authoritative clipboard, insisted as follows, in an escalating series of commands, at each succeeding sign of resistance:

· “Please continue.” or “Please go on.”

· “The experiment requires that you continue.”

· “It is absolutely essential that you continue.”

· “You have no other choice; you must go on.”

In that elaborate context, participants obeyed. The core findings showed that 65% of participants progressed up the shock generator past Danger: Severe Shock to 450 volts labeled simply XXX. Only 35% resisted.

Participants clearly suffered in the process of obeying, displaying their tension by nervous laughter, hand-wringing, and facial distress. They reported moderate to extreme tension and nervousness, when asked. Fortunately, they were debriefed carefully about the matters on which they were not fully informed: The study concerned obedience, not learning and memory; the learner was a confederate; no shocks were ever delivered, except the sample to the teacher; and the experimenter's prods, as well as the learner's protests, were carefully scripted.

The results were appalling, of course. Most variations of the original paradigm (silent, remote learner) failed to depress obedience by much (Milgram, 1974); people still obeyed if the learner screamed and protested a heart condition (screaming, remote learner; 62%). But fewer people did obey when the learner sat in the same room as the teacher (proximate; 40%) or when they personally had to place the learner's hand on the shock plate (touching; 30%). The maximum shock decreased steadily from the farthest to most immediate conditions: from silent remote to screaming remote to proximate to touching. Besides the victim's immediacy, the authority's immediacy also mattered. Obedience dropped sharply as the experimenter moved from a few feet away to a telephoned intercom to a mere tape recording. And conducting the experiment in a rundown commercial suite decreased obedience to 48%. When teachers were supported by disobedient peers, only 10% obeyed.

Milgram explicitly understood the conditions of obedience to reflect the social force field operating on the participant. Going back again to Kurt Lewin's concept of life space ( Chapter 1 ), people operate in a social force field, influenced by social pressures that vary in strength, which is determined by their number, direction, and proximity. Social impact theory, described earlier in this chapter, adopts a similar framework. When the authority is proximate, the authority's commands carry more force. When the victim is immediate, the victim's suffering carries more force. Disobedient peers and a less impressive setting also weaken the experimenter's force field.

Social forces know no cultural or temporal boundaries. Subsequent research has demonstrated the robustness of the obedience effects (Blass, 1999; Miller, Collins, & Brief, 1995). Men and women show the same rates of obedience, although women react with greater tension and nervousness. Over the decades since the original studies, obedience rates have stayed the same. Similar rates of obedience occur in Europe, Jordan, South Africa, and Australia.

In one conceptual replication—that is, obedience to authority using a different paradigm—participants had to criticize and derogate a job applicant, disturbing him so that he failed the test and remained unemployed (Meeus & Raaijmakers, 1986, 1995). Although the consequences for the applicant getting the job were allegedly real, the experimenter supposedly wanted to disrupt the applicant in order to study stress. Despite their discomfort with the task, 90% obeyed. However, in numerous variations, as with the original Milgram studies, experimenter absence and peer rebellion both reduced obedience, as did information indicating the participant's own legal liability.

The obedience effects are robust, across a variety of people and many situations, but individual differences nonetheless do matter (Blass, 1991). Sometimes authoritarianism matters; this fits its definition ( Chapter 11 ) as the tendency toward conventionality (conformity to traditional values), authoritarian submission (obeying powerful leaders), and authoritarian aggression (sanctioned aggression against deviants). All these factors relate to obedience. Other factors sometimes matter: Dispositional hostility, trust in authorities, moral judgment, external locus of control, and belief in divine control can predict obedience.

In general, personality variables for ceding control tend to predict obedience, which makes sense. Our core motive of control emerges here. The control contingency appears to take this form: If I do this, I will not be punished. Also relevant are a core motive for shared understanding. To understand the situation the way the authority defines it (“the experiment requires that you continue”), participants apparently resist having to contradict the authority's assertion that they have no choice. Disobedience would disrupt a powerful, socially shared understanding. Finally, some researchers have proposed that trusting the authority's expertise and legitimacy is also relevant, although the evidence for this core social motive is mixed.

In keeping with the control motive interpretation, when obedience does occur, lay observers tend to view the participants as having relinquished control to a legitimate, expert authority (Blass, 1995, 1996a; Blass & Schmitt, 2001). The more authoritarian the observers, the more they absolve the participant of responsibility, probably because they endorse authoritarian submission. Also, the more observers view obedience as common, the more they absolve the participant (Blass, 1996b; Tyler & Devinitz, 1981).

Ultimately, responsibility rests with the researcher for the participant's suffering. Replications of the Milgram studies have ceased in the United States because of regulations protecting human research participants from procedures now viewed as harmful. Nevertheless, some important questions remain for conceptual replications of obedience studies. First, the exact features of the situational variants that maintain or reduce obedience are not well specified (Blass, 1991). That is, what exactly is it about the experimenter's presence or absence that might matter? What is it about Yale versus the storefront that might matter?

Second, some other situational findings (victim proximity) also might not be reliable (Blass, 1991); boundary conditions need to be clarified. One likely candidate for boundary condition is the victim's response. A meta-analysis of the original Milgram studies shows a critical decision point at 150 volts, the point at which the victim first asks to be released (Packer, 2008). The maximum disobedience occurred here and predicts the overall rate of disobedience. Apparently, participants disobeyed when they saw the victim's rights as overriding the experimenter's commands.

Third, participants are not always assigned randomly to condition, so the original Milgram studies and some follow-ups are not true experiments, and confounding, for example, time with condition, is possible. As noted in  Chapter 2 , other famous studies (the Zimbardo prison study, the Langer-Rodin nursing home study) were not true experiments but still have had enormous impact as demonstrations of the power of the situation. What's more, the “control group” in this case is people's expectations about what would happen, and observers consistently underestimate the amount of obedience that will occur.

Milgram explicitly undertook his research to explain the obedience of seemingly ordinary people following Nazi orders to destroy millions of Jewish people during the Holocaust. He certainly succeeded in demonstrating the extent of obedience to one form of destructive authority. But does this fully capture the potential evils of social influence? Evil is defined as action that constitutes an intentional, egregious violation of moral norms, especially when the actor is fully responsible but gains little, compared with the suffering of the victims (Baumeister, 1997; Berkowitz, 1999; Darley, 1992). The Milgram studies are a paradigm for understanding evil, but opinions differ as to whether the actual participants were necessarily behaving in an evil manner.

Power: Control Resources and Maintain Belonging

To jaywalk or not to jaywalk? Stand at a stoplight and watch who follows whom to cross against the light. Meta-analysis indicates that people normally jaywalk about 24% of the time, but if someone else obeys the light, they jaywalk only 16%, whereas if someone else jaywalks, they do too, about 44% of the time. A high-status model (i.e., a middle-aged white male in a business suit) produces more obedience to the light by obeying than does a high-status model who jaywalks (Mullen, Copper, & Driskell, 1990). Status is defined by having a high position in a hierarchy, and it often correlates with power. The jaywalking example illustrates the everyday roles of status and power in social influence.

Classic definitions of power rely on the role of social influence. According to this view, power is the ability to exert influence. In field theory terms, Lewin (1941, 1944) defined power as the amount of force one person can induce on another. Force does not necessarily imply change, for the target can resist. “Power is potential influence, and influence is power in action” (Ng, 1980, p. 157). Social psychologists realized early that power comes in many forms (French & Raven, 1959):

· Reward power = target's perception that the other person controls benefits

· Coercive power = target's perception that the other controls punishments

· Referent power = target identifying with the other

· Expert power = target perceiving the other to be knowledgeable

· Legitimate power = target's perception of the other's right to influence

Notice that all these forms of power presuppose that the target perceives the powerful other in certain ways. Thus, if one person actually controls resources, is appointed leader, or possesses relevant knowledge, that is not necessarily power. Power operates via perception. In this view, power thus is ceded by the powerless, an irony that researchers often ignore (Dépret & Fiske, 1993; Fiske & Dépret, 1996).

An alternative definition that both recognizes the role of the powerless and relies less on the success of influence but more on the potential for influence would say that power is control over valued resources (Fiske, 1993; Fiske, 2010; Fiske & Berdahl, 2007; Keltner, Gruenfeld, & Anderson, 2003), which may include rewards, punishments, status, identity, and knowledge. In this power-as-control view, control is central to understanding power, although other motives are relevant, as we shall see.

We address research on each basis of power, in turn. In reviewing the French-Raven bases of power, we return to many of the leadership principles described in  Chapter 12 , but the emphasis here contrasts the different sources of influence, rather than the role of the leader in the group.

CONTROLLING REWARDS

Leaders who offer rewards contingent upon performance are indeed effective, according to meta-analysis (Lowe, Kroeck, & Sivisubramaniam, 1996). This kind of contingency motivates subordinates to perform because they thereby know how they can control their outcomes. This text has argued that control is a core social motive; people do want to know the contingency between their own behavior and their outcomes. From this, it follows that those who lack control over their outcomes (the powerless) will attend to those who do control those outcomes (the powerful). The less powerful attend upward, seeking understanding and potential control (e.g., Fiske, 1993). Those who hold control will be less motivated to attend to their subordinates, who retain less control over resources (Fiske, 2010). A series of studies have found precisely those complementary patterns: People attend to others on whom their outcomes depend and interpret their behavior in dispositional terms, in an effort to predict their behavior (Dépret & Fiske, 1999; Erber & Fiske, 1984; Guinote, Brown, & Fiske, 2006; Neuberg & Fiske, 1987; Ruscher & Fiske, 1990; Stevens & Fiske, 2000). Attention to diagnostic information allows individuated impressions of the powerful. People expect the powerful to be free of constraint, so they attribute their behavior to more dispositional causes (Overbeck, Tiedens, & Brion, 2006). All this careful interpretive processing comes at a cost: Being powerless impairs executive function that allows people to maintain a focus on their main goals (Smith, Jostmann, Galinsky, & van Dijk, 2008).

Conversely, powerful people attend less to those who depend on them (Goodwin, Gubin, Fiske, & Yzerbyt, 2000; Operario & Fiske, 2001a) and in general the powerful are free to prioritize central, goal-consistent, focal features of any given situation—expectancies, goals, sensations—making them more goal directed but also changeable and unpredictable (Guinote, 2007a, 2007b, 2008; Guinote, Judd, & Brauer, 2002; Weick & Guinote, 2008). Their attentional neglect of seemingly peripheral features (and people) allows stereotyping of subordinates, as we saw in  Chapter 11 . Powerful parties to a conflict do indeed judge their opponents less accurately, whereas they are judged accurately by others (Ebenbach & Keltner, 1998; Keltner & Robinson, 1996, 1997). Powerful people are not good at taking others' perspectives (Galinsky, Magee, Inesi, & Gruenfeld, 2006). Powerful people's focus on the crucial features relevant to pursuing their goals means that they can view others in instrumental ways, that is, approaching them only if they are useful, thereby objectifying them (Gruenfeld, Inesi, Magee, & Galinsky, 2008).

This neglect-of-useless-subordinates effect occurs mainly for powerful people who are exchange-oriented and focused on organizational, rather than relational and interpersonal concerns (Chen, Lee-Chai, & Bargh, 2001; Overbeck & Park, 2001). When powerful people feel interdependent with an opponent, they are more generous, but if they are an interdependent team, they are less generous toward an opponent team (Howard, Gardner, & Thompson, 2007). Powerful people, who already control rewards, by definition, also are free to follow the rewards in the situation.

A related view of power (Keltner, Gruenfeld, & Anderson, 2003) associates power with rewards and freedom for the powerful person. Because of this, people in positions of power should be more oriented toward approach tendencies: cheerful, focused on rewards, behaviorally unconstrained, and cognitively automatic. People with less relative power should orient more toward inhibition tendencies: emotionally negative, focused on threat and on others' interests, behaviorally constrained, and cognitively controlled (see  Table 13.2 ). This approach-inhibition theory of power integrates a variety of responses (affect, cognition, behavior) by analyzing the interpersonal power of people who control rewards (and punishments). For example, being powerful sparks positive emotions and behavioral approach; being powerless creates negative emotions and behavioral inhibition (Anderson & Berdahl, 2002; Langner & Keltner, 2008). Powerful people are more likely to take risks (Anderson & Galinsky, 2006), to fix an unpleasant situation (Galinsky, Gruenfeld, & Magee, 2003), and to move first to gain a competitive advantage (Magee, Galinsky, & Gruenfeld, 2007). But all this holds only when both sides perceive their power to be legitimate (Lammers, Galinsky, Gordijn, & Otten, 2008). If legitimate, power is clearly a preferable state because one can act on one's goals. But power is not without its costs.

TABLE 13.2 Power, Approach, and Inhibition

	Power
Reaction	High	Low
Affect	Positive	Negative
Attentional focus	Rewards	Threat, punishment
	Others as means to own ends	Self as means to others' ends
Information-processing	Automatic	Controlled
Social behavior	Disinhibited	Inhibited

Source: From description by Keltner et al., 2003.

A more cynical metamorphic effects theory suggests that power corrupts relationships (Kipnis, 1976). In this view, people choose influence strategies based on their degree of dominance over others, their expectations of compliance, and their goals (Kipnis, 1984). Powerful people use rewards to influence subordinates who ingratiate (present themselves as likable and flatter the powerholder) (Kipnis & Vanderveer, 1971). Dominant people do like ingratiators, but they may not respect them (Operario & Fiske, 2001a). Powerful people in general devalue their subordinates because they believe that they themselves controlled their subordinate's performance, so therefore the subordinate does not deserve any credit (Kipnis, 1972; Kipnis, Castell, Gergen, & Mauch, 1976; Kipnis, Schmidt, Price, & Stitt, 1981). Ironically, this does not enhance the powerholder's own self-esteem. A cynical view would argue the adaptive value of the potentially powerful developing strategies for exploiting the potentially powerless through “deception, manipulation, coercion, intimidation, terrorization, and force”; evidence includes perceiving who's muggable, cheatable, abusable, assaultable, or killable (Buss & Duntley, 2008). Grim thoughts.

On the other hand, reward power does have beneficial effects, as noted, on performance. In one study ( Figure 13.1 ), reward power apparently encouraged perceiving the powerholder as having both expert and referent power. These perceptions in turn led to more constructive behavior and improved performance (Rahim, Antonioni, & Psenicka, 2001). And to be fair, powerful-subordinate relations function best when each expects the other to behave in complementary fashion (Tiedens, Unzueta, & Young, 2007). The more people care about a task, the more they perceive their partner to behave in the complementary power position (Tiedens & Fragale, 2003). Organizations arguably function better as hierarchies (Gruenfeld & Tiedens, 2010).

Figure 13.1  Relationships among Different Kinds of Power, Conflict Styles, and Job Performance

Source: From Rahim et al., 2001. Copyright © Information Age Publishing. Adapted with permission

CONTROLLING PUNISHMENTS

In many frameworks, whether the powerholder can dispense benefits or punishments would seem like flip sides of the same coin. But controlling punishments is more coercive than controlling rewards, as the French and Raven label implies. Power does carry overtones of control, more than mere influence does, and coercive power in particular carries those negative overtones (Hollander, 1985).

Coercive power works in the short run but not the long run. Subordinates may comply under coercion, but they disrespect and dislike the supervisor. They also become avoidant (withdrawn) and not happy, open, problem-solving, or obliging (Carson, Carson, Roe, Birkenmeir, & Phillips, 1999; Rahim, Antonioni, & Psenicka, 2001; Shaw & Condelli, 1986). Reciprocally, the more a powerholder uses coercive power, the less favorably the powerholder views the subordinate and the whole process (O'Neal, Kipnis, & Craig, 1994). Discipline problems evoke the use of punishments (Goodstadt & Kipnis, 1970), so maybe coercive power (punishment) operates in a power relationship already contaminated by poor authority relations. Being in the role of an authority may give people the perceived right to punish, but the effects soon evaporate, and continued obedience requires surveillance.

REFERENT POWER

When people identify with the authority, the source of power is relational or referent. According to meta-analyses, people like leaders who show them consideration (Lowe et al., 1996; Wofford & Liska, 1993) and establish good relations (Gerstner & Day, 1997). Often, they are more productive as well under good leader-subordinate relationships.  Chapter 12  noted that perceptions of just treatment depend in part on feeling respected by the group (Tyler & Lind, 1992). Referent power predicts a constructive, problem-solving orientation by subordinates and avoids a difficult, dominating, unobliging orientation (Rahim et al., 2001; see  Figure 13.1 ). Not surprisingly, referent power predicts commitment, attitudinal compliance, behavioral compliance, and satisfaction, according to meta-analysis (Carson, Carson, & Roe, 1993; e.g., Rahim, 1989).

Perceived expertise and legitimacy help to confer referent power (Rahim et al., 2001). We saw in  Chapter 12  that leaders who embody the group prototype elicit social identification (i.e., referent power), in keeping with social identity and self-categorization theory. Indeed, members who leave the group and return, although they may be highly involved, creative, knowledgeable, and apparent candidates for leader, instead are viewed as less loyal and valuable; therefore, they have less influence, because their identity is questioned (Gruenfeld, Martorana, & Fan, 2000).

Leaders oriented toward the group's social relationships are most effective in settings of mid-range difficulty for the group, defined in terms of control over task, authority, and morale (Peters, Hartke, & Pohlmann, 1985; Schriesheim, Tepper, & Tetrault, 1994; Strube & Garcia, 1981; see Richard, Bond, & Stokes-Zoota, 2003, report of these meta-analytic effects). This tends to fit Fiedler's (1964) contingency theory of leadership, introduced in  Chapter 12 , which says that the effectiveness of a task versus social leadership style depends on the favorability of the situation, although the support is uneven.

EXPERT POWER

Expert power, authority stemming from knowledge, also tends to fit with contingency theory, namely, the effectiveness of task-oriented leaders when situational control is either very low (structure is needed; Graen, Alvres, Orris, & Martella, 1970; Peters et al., 1985; Schriesheim et al., 1994; Strube & Garcia, 1981) or very high (efficiency is fine; Peters et al., 1985; Schriesheim et al., 1994; Strube & Garcia, 1981). Being task-oriented under these circumstances apparently confers credibility on the leader.

The role of intelligence in anointing authorities also suggests the importance of expert power (Lord et al., 1986; Mann, 1959; Stogdill, 1948). When subordinates are inept, expert power intervenes to correct poor performance (Goodstadt & Kipnis, 1970). Expert power builds subordinate commitment, compliance, satisfaction, and attitude change (Rahim, 1989; Rahim & Mainuddin, 1993). These effects on persuasion also fit the discussion in  Chapter 6  on source credibility effects. What's more, high status confers perceived competence, especially for perceivers high on belief in a just world or social dominance orientation (Oldmeadow & Fiske, 2007), so perceived power and expertise reinforce each other.

Expert power might seem to rely on the difference between self and the powerful expert (Ng, 1980). But SIT/SCT interprets expertise as the person who most represents the group consensus and therefore its prototype (Turner, 1991). The powerful expert differs from self only in better expressing the group's shared identity. Thus, any difference occurs in a pronormative direction. This interpretation ties expert power to referent power and to legitimacy.

LEGITIMATE POWER

Being in the role of an authority gives the opportunity for exercising various kinds of power (Rahim et al., 2001). So, who then is viewed as an appropriate leader? In  Chapter 12 , we saw that leaders have good relationships with others (essentially referent power), being close, central, satisfied, and respected, according to meta-analysis (Bass, 1954; Mullen, 1991). Legitimacy and referent power indeed are closely linked (Rahim et al., 2001).

Certain individual characteristics predict who is perceived to be a legitimate leader: Authorities are typically expected to be male, as we saw in  Chapter 12 , but also older (Stogdill, 1948). Expectation states theory, introduced in  Chapter 12  (Berger, Webster, Ridgeway, & Rosenholtz, 1986; Ridgeway, 2001), predicts that people develop expectations about the competence of other group members through status characteristics (such as gender, age) and culturally valued abilities that may or may not be relevant to the task at hand. Given the traditional divisions of labor between men and women, according to social role theory, women are expected to be less competent leaders (Eagly & Karau, 2002). Although a masculine style predicts leadership (Lord et al., 1986), women are penalized for behaving in a masculine fashion, so they are less likely to be viewed as authorities in traditionally masculine settings (Eagly et al., 1992), as noted in the last chapter.

Other personality characteristics predict who will be perceived as having legitimate power, that is, the right to be a leader or an authority: intelligence and dominance (Lord et al., 1986; Mann, 1959; Stogdill, 1948). As noted in the last chapter, a benign view of leader personality holds across cultures (Den Hartog et al., 1999), and it is consistent with notions of charisma: visionary, inspirational, team-integrating, trustworthy, diplomatic, and not self-centered, malevolent, inexplicit, or autocratic. Consistent with this interpretation of charisma, charismatic leaders indeed are effective, according to meta-analyses (Fuller, Patterson, Hester, & Stringer, 1996; Lowe et al., 1996).

And of course, legitimate power is conferred by culturally accepted means: election, moral mandate, or appointment. People may even seek power to resolve injustice (Foster & Rusbult, 1999). Democratically elected leaders, for example, like and respect their subordinates (Kipnis et al., 1981), and their subordinates like them as well, according to meta-analysis (Gastil, 1994). The importance of the processes that select leaders and determine group decisions can hardly be overstated. Process sometimes matters more than the outcomes, a point in  Chapter 12  introduced as procedural justice (Tyler & Lind, 1992; Tyler & Smith, 1998). Legitimacy is the subjective sense that an authority (or institution) is appropriate, proper, and just (Tyler, 2006). Legitimacy has a strong relational component, whereby legitimate leaders elicit voluntary deference to the extent people feel valued in the group and authorities maintain good relations (Tyler, 1997; Tyler et al., 1996). Powerholders who are more relationship-oriented—as opposed to more instrumental and exchange-oriented—behave in more socially responsible ways (Chen et al., 2001), and powerholders can be constrained to behave in more interpersonally concerned ways (Overbeck & Park, 2001). Obtaining voluntary acquiescence creates the most effective leadership, even when times are hard (Tyler, 2006).

People prefer democratic, elected, internal leaders, especially when they identify strongly with the group (Foels, Driskell, Mullen, & Salas, 2000; Gastil, 1994; Van Vugt & De Cremer, 1999). Legitimacy may be conferred by best representing the group ideal or consensus (Hogg, Hains, & Mason, 1998). People talk in terms of “leaders” when they identify with the otherwise legitimate authority as having referent power, but they tend to talk in terms of “power” per se when the authority controls resources (rewards and punishments), especially if they do not identify with the authority (Turner, 1991). Belonging motives thus are critical to legitimate leadership.

Summary of Obedience

Obedience—the social influence of an authority or a powerholder on a subordinate—operates by social forces, coming from the authority, peers, and potential victims. Milgram showed that people are surprisingly obedient to destructive but otherwise legitimate authority, and the obedience effect holds up over time, culture, gender, personality, and many situations. Some people (low authoritarians) and some situations (rebellious peers, authority distance, victim proximity) do undermine obedience, but the role of obedience is surprisingly strong compared to common sense. On the other hand, people could not survive long in groups if the members did not typically obey the authorities. A motive for socially shared understanding between authority and subordinate matters here, as does trust, to some extent.

Authority stems from various bases of power: instrumental, controlling rewards and punishments; relational, conferring referent power through identification; expertise, through knowledge or consensus; and legitimacy through normative beliefs. Control motives (resources) and belonging motives (relational identity) are relevant here.

COMPLIANCE: STRATEGIES TO UNDERSTAND SELF, MAINTAIN BELONGING, AND CONTROL RESOURCES

We defined compliance as social influence from a peer's request. Several strategies are used by real-world influence agents (Cialdini, 1987, 1993). The core motives of the targets make these strategies effective, as they variously promote the target's motives for socially shared understanding, belonging, and controlling outcomes.

Understanding Self as Consistent

A single consistency explanation for compliance appears in various guises. Consistency (introduced in  Chapters 5  and  6 ) refers to people's strong tendency to view themselves as behaving in ways that fit their earlier behavior. Self-perception (introduced in  Chapters 3 ,  5 , and  6 ) refers to Bem's theory that people learn about themselves by observing their own behavior. Learning by doing refers to steps along the continuum of social influence, starting with a small, innocuous step and culminating in a major step, sometimes along a path of destructive behavior (Staub, 1989, 1999). What they all share in common is a person behaving in ways consistent with earlier, related behavior but ending up in ways that might otherwise be undesirable when considered in isolation. That is, they describe the slippery slope to behavior people normally want to avoid. Compliance works in many of the same ways.

The foot-in-the-door technique (FITD; Freedman & Fraser, 1966) describes this phenomenon as a strategy for eliciting compliance: A small favor (sure to be granted) is followed by a large, related favor. In one of the original studies, housewives agreed to a short phone survey about household products, followed a few days later by agreeing to have several men search their home for two hours. Similarly, as noted earlier, placing a small “drive carefully” sign in one's window increases compliance with a request to erect an enormous lawn sign several days later. Normally, the FITD strategy requires a delay between the initial small request and the later large one. If the requests come too close together, people feel exploited, and they refuse, perhaps due to reactance, that is, an unwillingness to let other people control them (Brehm, 1966; Brehm & Cole, 1966; Brehm & Sensenig, 1966). Otherwise, as we will see, the FITD effect depends on how it's done (Beaman, Cole, Preston, Klentz, & Steblay, 1983), but it works under the right circumstances.

The self-perception explanation for FITD is that people infer their attitude based on their response to the first request and then act on the larger request in response to their newly inferred attitude. This process suggests three hypotheses, all supported by meta-analysis (Burger, 1999):

· A stronger FITD effect when people are involved in the initial request, because then they will recall it when they encounter the second one

· A stronger FITD effect when people actually perform the first request, rather then merely agreeing to it, because action creates more commitment

· A weaker FITD effect when the initial request is so big that people refuse

All these results also fit the consistency notion. But consistency goes one step farther. In keeping with the consistency explanation, people with a high individual need for consistency show bigger FITD effects (Cialdini, Trost, & Newsom, 1995). Individual differences in need for consistency can override even the normal requirement for a delay between the initial small request and the subsequent large one, if the person's prior compliance is salient (Guadagno, Asher, Demaine, & Cialdini, 2001).

Closely related to the idea of consistency is commitment. As long as people commit themselves to the initial request, many of the same compliance processes are set in motion. For example, if people answer “just a few questions” and then are requested to answer continued questions (i.e., continuing same request; Burger, 1999), they are more likely to comply.

Also, in processes similar to self-perception, attributions can be key in compliance processes. One's attribution for agreeing to the initial request, for why one did it, encourages similar future behavior. If the initially compliant person is labeled as cooperative or helpful, compliance with the second request is more likely because it fits the new self-attribution (Burger, 1999). In a similar vein, attributions of intrinsic motivation—and continued compliance—are more likely if the requester avoids extrinsic rewards, also in line with self-perception theory.

Using the FITD technique has negative consequences for the person using it. Influence agents who use the FITD strategy enjoy it less and like the target less (O'Neal et al., 1994). Perhaps their own self-perception changes, as they learn by doing to manipulate or exploit another person.

All these processes hinge on people's motivation to understand themselves as maintaining a consistent and coherent self. Learning about their own history of compliance (that is, invoking the consistency-commitment principles) especially affects Americans or those with a highly individualist orientation. In contrast, a more belonging type of motivation—social proof (their peers' history of compliance)—can affect Poles and those with more collectivist orientation (Cialdini, Wosinska, Barrett, Butner, & Gornik-Durose, 1999). This brings us to the second core motive underlying compliance.

Belonging via Reciprocity, Liking, and Approval

People feel obligated to comply, in return for a favor, gift, invitation, and so on. This, of course, is the principle of “free gifts” used in marketing and fund-raising. People feel more obligated than they otherwise would. Survey researchers have discovered that attaching to their mailed questionnaires a small incentive, monetary or not, increases the return rate by nearly 20%. Payment contingent on completing the survey does not have the same effect. It's the unsolicited gift that does the trick (Tourangeau, 2004).

Reciprocity, the exchange of benefits for like benefits, appears in relationships ( Chapter 8 ), helping ( Chapter 9 ), and groups ( Chapter 12 ). Meta-analysis (Burger, 1999) supports reciprocity as a compliance principle for FITD. Reciprocity requires that each person avoid making repeated requests, which break the rules of fair exchange. When the same person makes both the small and large request with no intervening delay, people refuse to comply, as apart from all other combinations of delay (short/long) and requestor (same/different). This supports the idea that people react to perceived exploitation: Why is this person asking so much and giving so little? It also fits reactance, as noted earlier. For people to maintain their relationships, they must abide by the rules of reciprocity (Cialdini & Trost, 1998).

Sometimes, the reciprocity norm allows influence agents to obtain compliance by starting big and then moderating to a more reasonable request. The door-in-the-face technique trades on the influence agent appearing to step back, which makes the target feel guilty and perhaps need to reciprocate by agreeing to the more moderate request. Whatever the mechanism, it works (O'Keefe & Hale, 2001).

People also comply to maintain mutual liking. Influence agents often try to appear likeable, and we saw in  Chapter 6  that persuasion occurs in part when sources are likeable and attractive. Liked friends sell more Tupperware. Their likeability and social ties, viewed as degree of social capital, are measured by the strength of the buyer-seller tie and buyer indebtedness to the seller (reciprocity again). Social capital significantly predicts the likelihood of purchasing little plastic food storage dishes and other Tupperware products (Frenzen & Davis, 1990).

As we know, attractive people receive favorable attributions of being successful, socially skilled, and somewhat more competent (Eagly et al., 1991; Feingold, 1992b). Accordingly, they sell products, win votes, raise funds, and shorten their jail sentences, all reflecting their success as compliance agents. This builds on the belonging principle of people going along with those they like.

Perceived similarity also predicts compliance, further reinforcing the belonging motive. A fundamental principle of attraction is perceived similarity ( Chapters 7 ,  11 , and  12 ). People comply more often when similar others have complied, which maintains their sense of social identity. People compare themselves with similar others to see how they are doing ( Chapter 5 ), which functions to help them understand, enhance, and improve themselves. But it also reflects their sense of belonging with similar others.

From belonging with similar others also follows people's need for social approval. People comply with implied group norms, as we have seen in conformity. But norms also reinforce compliance with direct requests. For example, the FITD effect increases with relevant norms (Burger, 1999). In one study (DeJong, 1981), people who signed a petition (small initial request) were told that almost everyone else also agreed or that they were the only one so far. When told their behavior fit the norm, they were much more likely to help a second person, as befits complying with the perceived norm to be helpful.

Thus, a number of compliance principles fit the belonging motive: reciprocity, liking, attraction, social comparison, and adhering to norms. We turn now to another core social motive, control over one's outcomes.

Controlling Resources by Valuing Freedom and Scarcity

People also comply to control their own outcomes, attaining valued resources. We have already touched on reactance, which is the clearest indicator of this motive. People will comply in order to maintain their sense of freedom to choose. The principle of scarcity reflects people's tendency to value rare commodities, so they will try to obtain items that are going out of stock, about to become unavailable, or otherwise almost out of reach (Cialdini & Trost, 1998). College students valued even their cafeteria food when they learned that it was no longer available because of a fire in the building (West, 1975).

Items or information that are exclusive or hard to obtain are viewed as more valuable, so people use scarcity as a heuristic for value (Cialdini, 1993). Information, too, is more persuasive when it is a rare commodity (Brock & Brannon, 1992); censorship increases the value of banned communications (Worchel, 1992). And even love affairs are more appealing when they are secret (Wegner, Lane, & Dimitri, 1994). We saw earlier that obstacles can increase passion for another person ( Chapter 8 ). Apparent scarcity can increase the value of an object. Hence, scarcity is an effective compliance technique that relies on people's motive to control valued resources.

Summary of Compliance

People comply with direct requests partly in order to maintain a view of themselves as consistent. The foot-in-the-door technique describes people's compliance with a big request, because it fits their previous compliance with a related, though smaller, request. Consistency, commitment, self-perception, and learning-by-doing all explain related compliance phenomena that maintain coherent self-understanding.

People also comply in order to maintain their belonging with valued others. People adhere to the rules of reciprocity, returning benefit for benefit received. They comply with people who are likeable and attractive. And they comply with the norms of their groups in order to maintain social approval.

People attempt to control scarce resources because rarity often indicates value. But also people want to maintain their freedom to choose. Reactance names people's resistance to restrictions on their own freedom.

CHAPTER SUMMARY

Social influence is change in behavior as a result of interpersonal interaction. Because it focuses on actual behavior change and on social interaction, the research and theory remain distinct from persuasion and attitude change. Moreover, not all social influence is deliberate, whereas persuasion is. Operationally, social influence paradigms include (a) conformity to group norms that are arbitrary or erroneous (thereby showing the effects of group pressure); (b) obedience to destructive authority, enacted with the framework of the experimenter as authority over a participant; and (c) compliance with direct requests, mostly in field studies.

All five core social motives bear on social influence, which makes sense given its central role in social psychology. Early descriptions of social influence discussed its role in reducing uncertainty. Researchers differentiated between informational influence (about seemingly objective reality) and normative influence (about supposedly subjective social reality). Similarly, researchers described compliance in the service of controlling outcomes, identification in the service of group belonging, and internalization in the service of objective understanding.

However, these distinctions break down because informational and normative influence both depend on social consensus and multiple motives. The same is true of their respective parallels as internalization (due to information) and identification (due to social ties). The overlap between social and objective reality is clear, upon reflection. People would not, for example, assume that they would be the only ones to see objective reality but instead assume that the social consensus would agree with them because others would see what they see. Similarly, people rarely admit that their social reality is not objectively true. Belonging and understanding thus intermingle here. Controlling resources, trusting ingroup others, and enhancing self also matter.

The first conformity studies used ambiguous stimuli (Sherif's point of light, Asch's abstract words) to show conformity. Seeking to show independence, Asch introduced a perceptual task on which the majority was clearly in error; still, people conformed a surprising amount (one-third of the time). Scores of replications have shown stronger conformity under conditions that replicate the original studies, use ambiguous stimuli, and employ ingroup confederates as the majority viewpoint. Collectivist cultures, people operating outside their own spheres of expertise, and earlier eras also show stronger conformity effects.

Conformity can be relatively automatic, or at least heuristic, as when people assume that others share their opinions (false consensus effect) and assume they are conforming with what would be the prevailing norm. Majority influence, in particular, conveys this kind of unthinking agreement. In contrast, a consistently disagreeing minority viewpoint creates conflict, potentially resulting in divergent, creative thinking and attitude change on private or indirect dimensions, as a result of having to think about the minority perspective. Social impact theories describe the number, strength, and immediacy of influence agents as producing change, by whatever process, automatic or controlled.

Social categorization clearly matters, as people are open to social influence by ingroup others. People conform because their ingroup provides subjective validity, especially given the more similar and attracted they feel to the ingroup. People view the most representative perspective as the prototype of the group and conform to it as a function of identity. The more cohesive, consistent, and coherent the group perspective, the more people conform, all in keeping with social identity approaches.

Salient forms of conformity include memes (units of culture that get passed along), rumors (verbally conveyed beliefs, without benefit of evidence), and gossip (rumors about people). People's belief in these intangible cultural artifacts reflects compliance as a result of group belonging and socially shared understanding.

Obedience to authority was most dramatically demonstrated by Milgram's original studies, reflecting the operation of social forces. The high rates of destructive obedience replicate across gender, place, and decade, although personality differences do matter somewhat. And contextual manipulations that reflect the proximity of social forces (the authority, the victim, peers) also matter a lot.

Power is the potential to influence others, through control over resources and relationships. Coercive power (punishment) often backfires, whereas reward power leads to more constructive results, although it requires surveillance. Referent power extends our understanding to the arena of group belonging, as leaders represent the prototypic ideal or group consensus and have power that way. This kind of power depends on maintaining group relationships. Expert power depends on perceived knowledge, but can also represent the group's consensus. From all sources of power and contributing to all of them is legitimacy, that is, the authority having the perceived right to lead and to demand obedience. Control over resources and belonging to group relationships both are important motives, along with the socially shared understanding of the situation and trust in ingroup authorities.

Compliance results from strategies that allow people to understand the self as consistent and committed. People also seek to maintain belonging through reciprocity, liking, and social approval. They attempt to control resources by complying with requests that protect their freedom and allow them to obtain scarce commodities.

SUGGESTIONS FOR FURTHER READING

1. Cialdini, R. B. & Goldstein, N. J. (2004). Social influence: Compliance and conformity. Annual Review of Psychology, 55, 591–621.

2. Crano, W. D. (2000). Milestones in the psychological analysis of social influence. Group Dynamics, 4, 68–80.

3. Fiske, S. T. (2010). Interpersonal stratification: Status, power, and subordination. In S. T. Fiske, D. T., Gilbert, & G. Lindzey (Eds.), Handbook of social psychology (5th ed., Vol. 2). New York: Wiley.

4. Hogg, M. A. (2001). Social categorization, depersonalization, and group behavior. In M. A. Hogg & R. S. Tindale (Eds.), Blackwell handbook of social psychology: Group processes (pp. 56–85). Malden, MA: Blackwell.

5. Hogg, M. (2010). Influence and leadership. In S. T. Fiske, D. T., Gilbert, & G. Lindzey (Eds.), Handbook of social psychology (5th ed., Vol. 2). New York: Wiley.

6. Martin, R., & Hewstone, M. (2001). Conformity and independence in groups: Majorities and minorities. In M. A. Hogg & R. S. Tindale (Eds.), Blackwell handbook of social psychology: Group processes (pp. 209–210). Malden, MA: Blackwell.

7. Turner, J. C. (1991). Social influence. Pacific Grove, CA: Brooks/Cole.

8. Wood, W. (2000). Attitude change: Persuasion and social influence. Annual Review of Psychology, 51, 539–570.