Social influence - Social approach

Simply Psychology - Michael W. Eysenck 2022


Social influence
Social approach

What we say and how we behave are influenced by other people. We want to be liked and to fit into society, which often leads us to hide what we really think and to behave in ways that will meet the approval of others. Social psychologists have discovered we are influenced much more by other people than we think we are. This is social influence, “the process whereby attitudes and behaviour are influenced by the real or implied presence of other people” (Vaughan & Hogg, 2014, p. 214).

In this chapter, we consider the main ways social influence manifests itself. These include conformity behaviour, obedience to authority, groupthink, group polarisation, social power, and crowd behaviour.

KEY TERMS

Social influence: efforts by individuals or groups to change the attitudes and/or behaviour of others.

Conformity: changes in behaviour and/or attitudes occurring in response to group pressure.

Conformity

Conformity involves yielding to group pressure, something we all do sometimes. For example, suppose you go with friends to see a film. You did not enjoy it much but all your friends thought it was brilliant. You might be tempted to conform by pretending to agree with their verdict on the film rather than being the odd one out.

Most textbook writers argue that conformity is undesirable. However, it often makes sense. Suppose your friends who are also studying psychology have the same view on a given topic in psychology but it differs from yours. If they know more about the topic, it is probably a good idea to conform to their views!

Image

Figure 15.1 Asch showed lines like the above to his participants. Which line do you think is the closest in height to line X? A, B, or C? Why do you think over 30% of participants answered A?

Solomon Asch: majority influence

Solomon Asch (1951, 1956) studied conformity using a situation where several people (typically seven) viewed a display (see Figure 15.1). Their task was to say aloud which one of three lines (A, B, and C) was the same length as a given stimulus line (X), with the experimenter working his way around the group members in turn. This task was very easy — 99% of judgements were correct under control conditions.

All group members (except for one genuine participant) were confederates working with the experimenter and had been told to give the same wrong answer on some trials (the critical trials). The only genuine participant was the last (or last but one) to provide their opinion.

Findings

What do you think the genuine participants did on critical trials when faced with the conflict between what the other group members said and what they knew was the right answer? They gave the wrong answer on 37% of these critical trials. Only 25% made no errors throughout the experiment. Conformity increased as the number of confederates rose from one to three but there was little change in conformity after that (Asch, 1956).

Asch’s research is famous within social psychology. However, there was nothing very social about it because he used groups of strangers! Abrams et al. (1990) predicted that participants would exhibit more conformity when the confederates appeared similar to themselves. In these circumstances, participants would identify more with the group and would be motivated to maintain group cohesion by conforming. As predicted, conformity occurred on 58% of critical trials when the confederates were described as psychology students but only 8% when they were described as students of ancient history.

Asch’s research was carried out in the United States in the late 1940s and early 1950s. Since then, there has been a significant reduction in people’s need for social approval (Twenge & Im, 2007) and so we might expect conformity to have reduced over time. That is the typical finding but a conformity effect is still found (Bond & Smith, 1996).

We would expect greater conformity in collectivistic cultures (which emphasise group belongingness) than individualistic ones (which emphasise personal responsibility). As predicted, conformity is greater in collectivistic cultures in Asia, Africa, and elsewhere (37% of trials) than individualistic cultures in North America and Europe (25%) (Bond & Smith, 1996).

When does conformity break down? The findings were very different when one confederate gave the correct answer (Asch, 1956). In those conditions, conformity occurred on only 5% of trials. The comforting feeling of not being entirely isolated made it much easier for the participants to avoid conforming.

Theoretical considerations

Deutsch and Gerard (1955) identified two reasons why individuals conform in Asch-type studies (see Figure 15.2). First, there is normative influence: individuals conform to be liked or respected by group members. There was strong normative influence in the study by Abrams et al. (1990; discussed earlier) when the other group members were perceived as similar to the participants.

Second, there is informational influence: individuals conform because of others’ superior knowledge. Lucas et al. (2006) found participants who doubted their mathematical ability showed more informational influence on conformity than those with greater confidence in their mathematical ability.

Image

Figure 15.2

KEY TERMS

Normative influence: conformity based on people’s desire to be liked and/or respected by others.

Informational influence: conformity based on the perceived greater knowledge or judgement of others.

We can understand more about the nature of conformity by considering participants’ responses in public (i.e., group members hear their responses) and in private (i.e., other group members are unaware of those responses). Conformity is due to compliance when it is found only publicly but is due to acceptance when it is found both publicly and privately (Kelman, 1958). Compliance is thought to involve primarily normative influence, whereas acceptance mostly involved informational influence.

The correct answer is so obvious in the Asch situation that it is often believed to involve only normative influence (compliance) (Sowden et al., 2018). However, reality is more complex. In one experiment, Asch (1956) found conformity on 43% of trials in the public condition and 12% in the private condition. These findings suggest that conformity in the public situation depends on normative and informational influences, with the former greater than the latter. This was confirmed by Bond (2005) in a review of 125 Asch-type studies.

Sowden et al. (2018) used a simple colour-judging task with public and private conditions. When there were only two other group members, normative influences (compliance) and informational influences (acceptance) on conformity did not differ significantly. However, compliance was more important than acceptance when there were more group members.

Conformity: good or bad?

Asch (1955) argued that independence (i.e., avoidance of conformity) is greatly to be preferred to conformity. From that perspective, it is regrettable that 75% of his participants succumbed to group pressure at least once and so exhibited mindless conformity. In my opinion, this argument is flawed in several ways.

First, nearly 70% of participants defied the group on over 50% of trials and 25% did not conform at all. Thus, we must avoid exaggerating the extent of conformity.

Second, many participants become aroused and somewhat distressed in the Asch situation, especially when exhibiting independent behaviour (Hatcher et al., 2018). Thus, participants experienced a conflict between producing the correct response and their desire not to ignore the group.

Third, most participants provided a mixture of incorrect and correct answers. This can be regarded as the optimal solution to the conflict they experienced (Hodges, 2014). Their incorrect answers allowed them to avoid being disrespectful to other group members and to exhibit group solidarity. The motive to promote group solidarity is greater when other group members are perceived as similar to participants and so they exhibit more conformity (Abrams et al., 1990). Their correct answers allowed them to show their perceptual accuracy and perhaps persuade other group members to reconsider their judgements.

EVALUATION

· There is much more conformity than most people would predict on a task as easy as the Asch task where the correct answer is obvious.

· Several factors influencing conformity (e.g., number of confederates; presence vs. absence of a supporter, type of culture) have been identified.

· The roles of normative and informational influence in conformity behaviour have been clearly established.

· There is insensitivity of measurement in the Asch situation, given that two of three response options are obviously wrong; in addition, the task itself is trivial and does not call into question participants’ deeply held beliefs.

· The behaviour of most participants on Asch-type tasks has been criticised. However, their behaviour can be seen as an effective way of showing respect to the other group members and also exhibiting accurate visual perception.

· The Asch situation raises important ethical issues. Participants are deliberately deceived (i.e., they do not know the other participants are confederates of the experimenter). As a result, they cannot give informed consent to take part in experiment even though informed consent is required in ethical research.

Serge Moscovici: minority influence

Asch studied the influence of the majority on a minority (typically of one) within a group. However, minorities can also influence majorities. Serge Moscovici was an influential pioneer in research on minority influence and so we start with his contribution.

Moscovici (1976, 1980) proposed a dual-process theory of influences within groups based on compliance and conversion. When a majority influences a minority, compliance is often involved — it is based on the majority’s power and usually involves public (but not necessarily private) agreement with the majority. Compliance often occurs rapidly and without much thought.

Conversion is how a minority influences a majority. It involves convincing the majority that the minority’s views are correct. This is most likely to occur under the following conditions:

1. Consistency: the minority must be consistent in their viewpoint.

2. Flexibility: the minority must not appear to be rigid and dogmatic in how they present their viewpoint.

3. Commitment: a committed minority will lead the majority to re-think their position.

Conversion often produces private as well as public agreement among majority members. It is generally more time-consuming than compliance and occurs after cognitive conflict and much thought.

KEY TERMS

Compliance: a majority’s influence on a minority determined by the majority’s power and authority.

Conversion: a minority’s influence on a majority by persuading the majority of the correctness of its views.

Moscovici and Pérez (2007) argued there are two different ways minorities can influence majorities. One way (emphasised earlier by Moscovici, 1976, 1980) involves an active minority creating conflict within a group leading to conversion. The other way involves a victimised minority creating a sense of guilt in majority members which causes the minority to exert overt influence. This often involves focusing on historical issues (e.g., slavery in the case of black people).

This form of minority influence has increased considerably within Western societies over the past 20—30 years: “Nowadays equal rights, non-discrimination, the elimination of sexual or racial harassment and emancipatory measures are demanded on the strength of victimhood” (Moscovici & Pérez, 2007, pp. 726—727).

Findings

Wood et al. (1994) identified three conformity effects predicted by Moscovici’s dual-process theory (see Figure 15.3):

1. Public influence: the individual’s behaviour within the group is influenced by others’ views. This should occur mostly when majorities influence minorities.

2. Direct private influence: there is a change in the individual’s private opinions about the issue discussed by the group. This should be found mainly when minorities influence majorities.

3. Indirect private influence: the individual’s private opinions about related issues change. This should also be found mostly when majorities are influenced by minorities.

Wood et al. (1994) reviewed research relevant to the above three effects. As predicted, majorities in most studies had more public influence than minorities. Also as predicted, minorities had more indirect private influence than majorities (especially when their opinions were consistent). However, majorities had more direct private influence than minorities, which is contrary to Moscovici’s theory.

Image

Figure 15.3

David and Turner (1999) argued minority influence will be found only when the minority is perceived as part of the ingroup. In their study, the participants were moderate feminists exposed to the minority views of extreme feminists. These minority views influenced the majority when the situation was set up as feminists vs. non-feminists — the extreme feminists were part of the ingroup. However, the views of the extreme feminists had little impact when there was an obvious contrast between a moderate feminist majority and an outgroup of extreme feminists.

Moscovici and Pérez (2007; discussed above) distinguished between active minorities who create conflict with majorities and victimised minorities who create guilt in majorities. Pérez and Molpeceres (2018) found victimised minorities had more impact on majority behaviour than did active minorities (e.g., willingness to pay compensation to an oppressed minority). However, active minorities had a greater impact on majority views of minorities than did victimised minorities. There was a tendency for majorities to believe that victimised minorities should exert more effort to avoid being victimised. Thus, active and victimised minorities both achieve an impact on majorities but the nature of that impact is different.

EVALUATION

· Minorities often influence majorities.

· The influence of active minorities on majorities is mainly in the form of private rather than public agreement. The opposite pattern is found when majorities influence minorities.

· There is some evidence that the effects of active minorities on majorities differ from those of victimised minorities.

· Minorities generally differ from majorities in several ways (e.g., power; status). Differences in the social influence exerted by majorities and minorities may depend on their power or status rather than on their majority or minority position within the group.

· Moscovici exaggerated the differences between majority and minority influence. As Smith and Mackie (2000, p. 371) pointed out, “Minorities are influential when their dissent offers a consensus, avoids contamination [i.e., obvious bias], and triggers private acceptance — the same processes by which all groups achieve influence.”

Obedience to authority

In nearly all societies, certain individuals are given power and authority over others. In our society, parents, teachers, politicians, and managers are all invested with varying degrees of authority. This generally makes sense — if a doctor tells us to take some tablets three times a day, they are the expert and so we do as we are told without thinking any more about it.

Obedience to authority resembles conformity in that both involve social influence. However, research on obedience to authority differs in three ways from that on conformity. First, participants are ordered to behave in certain ways rather than being fairly free to decide what to do. Second, participants are typically of lower status than the person issuing the orders, whereas they are usually of equal status to the other group members in conformity research. Third, participants’ behaviour in obedience studies is determined by social power, whereas in conformity studies it is influenced mostly by the need for acceptance.

It is of considerable theoretical (and real-life) importance to see how far individuals are willing to go in obeying an authority figure. What do you think would happen if you (or other individuals) were asked to do something you knew was wrong? The most famous (or notorious?) attempt to answer that question was by Stanley Milgram (see below).

Stanley Milgram

Milgram’s research (reported in book form in 1974) was conducted at Yale University in the early 1960s. Pairs of participants were used in a simple learning task (see Figure 15.4). The “teacher” gave electric shocks to the “learner” for wrong answers increasing the shock intensity each time. At 180 volts, the learner yelled, “I can’t stand the pain!” and by 270 volts his only response was an agonised scream. The maximum shock was 450 volts.

Image

Figure 15.4 The photographs show the electric shock machine used in Milgram’s classic experiment where 65% of the participants gave a potentially lethal shock to the “learner”, shown in the bottom left photograph. The learner was actually a confederate of the experimenter, a 47-yearold accountant called “Mr Wallace”. The photographs show the experimenter (in the overall) and the true participant, the “teacher”.

Would you give the maximum (and potentially fatal) 450-volt shock in this situation? Psychiatrists predicted that only one person in a thousand would do this. In fact, 65% of Milgram’s participants gave the maximum shock — 650 times as many as the expert psychiatrists had predicted!

Milgram found two main ways that obedience to authority could be reduced (see Figure 15.5):

1. Increasing the obviousness of the learner’s plight.

2. Reducing the authority or influence of the experimenter.

The first way was studied by comparing obedience in four situations (the percentage of totally obedient participants is in brackets):

· Remote feedback: the victim could not be seen or heard but his thumping on the wall could be heard (66%)

· Voice feedback: the victim could be heard but not seen (62%)

· Proximity: the victim was only 1 metre away from the participant (40%)

· Touch-proximity: this was like the proximity condition except the participant had to force the learner’s hand onto the shock-plate (30%)

In one experiment, Milgram reduced the authority of the experimenter by carrying out the experiment in a run-down office building rather than at Yale University. The percentage of obedient participants went down from 65% at Yale University to 48% in the run-down building. When the experimenter’s influence was reduced by having him give orders by phone, obedience fell to only 20%.

Image

Figure 15.5

All the factors discussed above reduced obedience because they increased the extent to which participants identified with the learner/victim (Reicher et al., 2012). In contrast, reducing the obviousness of the learner’s plight and increasing the experimenter’s authority led to greater obedience because participants identified with the experimenter (Reicher et al., 2012).

Very high levels of obedience were reported in most subsequent research using Milgram’s standard set-up. The percentage of totally obedient participants was 80% or more in Italy, Spain, Germany, Austria, and Holland (Bond & Smith, 1996). Such high obedience levels may reflect the precise way Milgram set up his experiments (Burger, 2011). For example, the learner protested for the first time only after the tenth shock. Probably more participants would have disobeyed the experimenter if the learner’s anguish had been obvious earlier. In addition, the shocks increased slowly in 15-volt increments. This made it hard for participants to notice when they were being asked to behave unreasonably.

Real-life situations

Milgram’s research was laboratory-based. However, similar findings have been reported in the real world. For example, in a study by Hofling et al. (1966), nurses were phoned up by someone claiming to be Dr Smith. He instructed them to give 20 mg of a drug called Astroten to a patient. The nurses should have refused because they did not know Dr Smith and the dose was much higher than the maximum safe dose. However, 95% of the nurses obeyed the instruction. In contrast, Rank and Jacobsen (1977) found only 11% of nurses obeyed a doctor’s instructions to give an overdose of Valium to patients. This low level of obedience occurred because the nurses discussed the issue with other nurses before deciding what to do.

Krackow and Blass (1995) asked 68 nurses to indicate occasions when they had obeyed or defied inappropriate doctors’ orders. When nurses obeyed, they mostly attributed responsibility for any harm the patient might suffer to the doctor. In contrast, those nurses who had defied the doctor had generally accepted personal responsibility for harm to the patient.

Theoretical considerations

There are several reasons why participants in Milgram’s research were obedient (or disobedient). Milgram (1974) argued that participants were obedient because they were in an agentic state: this involves becoming the instrument of an authority figure (i.e., the experimenter) and so ceasing to act according to their conscience. Someone in an agentic state thinks, “I am not responsible because I was ordered to do it!” As predicted by this account, there is much less obedience when participants are told they are responsible for any harm (Tilker, 1970). Note, however, that Milgram explicitly told many participants he would accept full responsibility if anything went wrong, thus loading the dice in favour of his theory.

Milgram claimed the tendency to adopt the agentic state, “is the fatal flaw nature has designed into us”. This led him to argue that his findings shed light on the horrors of Nazi Germany. This is misleading for various reasons. First, the values underlying Milgram’s research were positive (e.g., understanding more about human learning) whereas those of Nazi Germany were vile (e.g., racial superiority). Second, Milgram’s participants had to be watched closely to ensure their obedience, but this was not necessary in Nazi Germany. Third, many participants experienced strong conflict between the experimenter’s demands and their own conscience. In contrast, most Nazis were unconcerned about moral issues.

KEY TERM

Agentic state: feeling controlled by an authority figure and therefore lacking a sense of personal responsibility.

Haslam et al. (2015) favoured an alternative explanation to Milgram’s agentic state account. They argued that many participants identified with the experimenter and with the scientific goals underpinning the study. In essence, they displayed “engaged followership” rather than the passive obedience emphasised by Milgram. This interpretation was supported by questionnaire responses obtained from participants weeks or months after taking part, with 50% of them being positive about their involvement in the study.

Hollander and Turowetz (2018) cast doubt on the above explanation. They analysed interviews Milgram’s obedient participants gave immediately after their involvement in his research. When asked to explain their obedience, 24% referred to the importance of the experiment (consistent with the notion of engaged followership). However, 59% indicated they were following instructions (consistent with the notion of an agentic state), and 72% indicated they did not believe the learner was being harmed.

Perry et al. (2020) also re-analysed Milgram’s data. Participants’ level of belief in the learner being shocked was strongly associated with their obedience. Of those fully believing shocks were administered, only 38% were obedient. In contrast, 58% of those highly sceptical that shocks were administered were obedient. These findings seriously undermine the validity of Milgram’s research.

EVALUATION

· Milgram’s findings are often regarded as among the most surprising in the history of psychology (but are less surprising when you examine his set-up in detail).

· Milgram’s findings have direct relevance to many everyday situations (e.g., doctor—nurse interactions).

· Milgram systematically increased the chances that his participants would be in an agentic state by telling many of them that he would accept full responsibility.

· Milgram de-emphasised the finding that participants who were totally obedient were much more likely than resistant participants to be sceptical that shocks were actually being administered.

· Milgram exaggerated the similarities between his laboratory situation and the appalling horrors of Nazi Germany. For example, the goals of the former were positive whereas those of Nazi Germany were vile.

· There are serious ethical issues with Milgram’s research. Participants did not give their informed consent, deception was involved, and the participants were not free to leave the experiment if they wanted to. Many participants experienced very negative emotions during the course of the experiment and felt bad about themselves for being willing to inflict serious harm on another human being. Thus, Milgram failed to protect the participants from harm. Unsurprisingly, these ethical issues mean that practically no Milgram-type studies have been carried out for many years.

Group decision making

It is often believed that groups are generally more cautious than individuals in their decision making (see Figure 15.6). For example, group decisions may simply reflect the average views of all (or most) of its members. In fact, what often happens is group polarisation: “a change in individual preferences from pre-group discussion to post-group discussion in the direction that is favoured by individuals’ average pre-discussion preferences” (Sieber & Ziegler, 2019, p. 1482).

What factors influence group polarisation? First, there is social comparison. Individuals want to be positively evaluated by other group members. If they see other group members endorsing positions closer than their own to some socially desired goal, they will change their position toward that goal. Isenberg (1986) reviewed numerous studies. Social comparison strongly influenced group polarisation, especially when value- or emotion-laden issues were discussed rather than factual ones.

Second, group polarisation is influenced by persuasive arguments. Suppose most group members initially favour a given type of decision. During the discussion, individuals are likely to hear new arguments supporting their position (Larson et al., 1994). As a result, their views will often become more extreme. Isenberg’s (1986) review indicated persuasive arguments produced more group polarisation with factual than with emotional issues.

Image

Figure 15.6

Sieber and Ziegler (2019) argued, in their processing effort account, that the impact of persuasive arguments on group polarisation depends on ability and motivation. Individuals high in ability and cognitive resources are most likely to engage in effortful processing of persuasive arguments and thus be influenced by them. In contrast, those low in ability and motivation are least likely to process (and be influenced by) persuasive arguments.

Third, members of an ingroup often want to distinguish their group from other groups. Suppose an ingroup has a confrontation with a cautious outgroup (not part of the ingroup). They can distinguish themselves from that group by becoming riskier in their decision making (Hogg et al., 1990). In similar fashion, an ingroup encountering a risky outgroup can distinguish itself by becoming more cautious (Hogg et al., 1990).

Groupthink

The processes producing group polarisation can have serious consequences (see Figure 15.7). This is especially so when groups engage in groupthink, “a mode of thinking in highly cohesive groups in which the desire to reach unanimous agreement overrides the motivation to adopt proper rational decision-making procedures” (Vaughan & Hogg, 2014, p. 312). Features of groupthink include suppression of dissent, exaggerating group consensus, and a sense of group invulnerability.

KEY TERM

Groupthink: pressure to achieve agreement in groups in which dissent is suppressed; it can lead to disastrous decisions.

Groupthink has led to catastrophic decisions in the real world. Sorkin (2009) analysed factors behind the near-collapse of several very large American banks in 2008. There was a culture of risk taking that produced huge profits for a few years. Individuals in those banks working on risk assessment who expressed concerns about the excessive risks the banks were taking were ignored or sacked.

Image

Figure 15.7

Janis (1982) argued that five factors increase the chances of groupthink occurring:

1. The group is very cohesive.

2. The group considers only a few options.

3. The group is isolated from information coming from outside the group.

4. There is much stress (e.g., time pressure; threatening circumstances).

5. The group is dominated by a very directive leader.

Findings

Tetlock et al. (1992) considered the real-world cases Janis (1982) used to support his groupthink theory. Groups showing groupthink typically had a strong leader and high levels of conformity. Contrary to Janis’s theory, however, groups exhibiting groupthink were generally less cohesive than other groups and exposure to stressful circumstances was relatively unimportant. In the workplace, high group cohesion is associated with positive outcomes such as greater loyalty and increased productivity (Haslam et al., 2006).

Eder (2019) pointed out that it is entirely possible for groupthink to play an important role during some stages of a group decision-making but not others. He studied more than 200 interviews and speeches of those in the small group that made the decision to invade Iraq in 2003. Eder identified two phases in the decision-making process: (1) building a consensus about the nature of the threat posed by 9/11; and (2) agreeing on the appropriate response to that threat. Eder’s key finding was that groupthink played a much greater role during the first phase than the second.

We can also understand groupthink from the perspective of social identity theory (Tajfel & Turner, 1979; see Chapters 13 and 19). According to this theory, we possess several social identities based on the groups to which we belong. Thus, you may have a social identity as a student, a second social identity as a woman, and so on. Of key relevance here is that, “Humans are groupish animals with a propensity to engage in motivated cognition to support their group identities” (Van Bavel et al., 2020, p. 66). Thus, many people regard the preservation of ingroup identity as more important than having accurate beliefs.

Kahan (2015) found 75% of Americans with left-wing or liberal views agreed there is solid evidence of climate change compared to only 22% of those with right-wing or conservative views. The scepticism of right-wing individuals may seem strange given that over 97% of the world’s leading scientific experts accept the reality of climate change. However, their scepticism is explicable on the basis that such individuals identify themselves as rugged individualists who believe strongly in capitalism and who distrust government involvement in their lives.

A crucial issue is that of causation: discovering that a group exhibits symptoms of groupthink and also makes extremely poor decisions does not necessarily mean that groupthink caused the poor decision-making. Baron (2005) argued (with strong supporting evidence) that groupthink symptoms (e.g., suppression of dissent) are present in most groups (not only those making catastrophic decisions). Peterson et al. (1998) studied top management teams when making good and bad decisions. Contrary to Janis’s theory, the symptoms of groupthink were present almost as frequently during good as poor decision-making.

Baron (2005) reviewed research on the five factors Janis (1982) identified as leading to groupthink. He argued these factors sometimes increase the chances of groupthink occurring. However, they are not necessary for groupthink to occur because most groups naturally strive for consensus even in the absence of the five factors.

Baron (2005) concluded that groups making good and bad decisions often use the same processes. What, then, determines whether a group’s decision making is good or bad? Much depends on the quality of shared knowledge within a group. Bad decisions are mostly made by groups possessing biased or task-inappropriate shared knowledge, whereas good decisions are made by group with unbiased and appropriate knowledge (Tindale et al., 2012).

EVALUATION

· Groupthink has played a role in many real-world catastrophic decisions.

· As Janis predicted, factors such as a strong leader and pressures towards conformity increase the chances of groupthink.

· Nearly all the evidence consists essentially of associations between groupthink and poor decision-making. It is hard (or impossible) to establish that the former has caused the latter.

· The symptoms of groupthink are present in most groups (successful as well as unsuccessful) and are found far more often than assumed by Janis.

Power: social roles

Social roles are the parts we play every day as members of various social and work groups. For example, most days your social roles probably include those of friend and student. The best-known study to assess how social roles influence behaviour was the Stanford prison experiment (discussed below). It was based on the assumption that guards and prisoners have very clear roles with the guards exerting power over the prisoners.

KEY TERM

Social roles: the parts we play as members of social groups based on certain expectations about the behaviour that is appropriate.

Stanford prison experiment

In the 1960s, there were numerous brutal attacks by prison guards on prisoners in American prisons. Why did this brutality occur? Perhaps prison guards have aggressive or sadistic personalities. Alternatively, the social environment of prisons (e.g., a rigid power structure) might be responsible. Zimbardo carried out the Stanford prison experiment (Haney et al., 1973) to decide between these two possibilities. Emotionally stable students agreed to act as “guards” and “prisoners” in a mock prison. It was argued that if these stable guards behaved brutally, this would indicate that the social environment in prisons (and the role of guards) produces hostility and aggression.

What happened? Violence and rebellion broke out within two days. When the prisoners shouted and cursed at the guards, the guards responded by violently putting down this rebellion using fire extinguishers. Over time, the prisoners became more subdued and submissive, often slouching and keeping their eyes fixed on the ground. At the same time, force, harassment, and aggression by the guards increased. For example, the prisoners were sleep deprived, put in solitary confinement, and forced to clean the toilets with their bare hands.

Zimbardo argued that these findings indicate that, “the mere fact of putting on uniforms was sufficient to transform [participants] into passive prisoners and aggressive guards” (Zimbardo et al., 2012). However, this conclusion is totally unwarranted. It was only several decades later that Zimbardo admitted that he had omitted numerous crucial details about the study from his account. Zimbardo himself was far more centrally involved in what happened than he indicated at the time. For example, he briefed the guards as follows: “You can create in the prisoners … a sense of fear to some degree … We have total power in the situation. They have none” (Zimbardo, 2007, p. 55).

There are other problems with the Stanford prison experiment (Le Texier, 2019). For example, the guards’ behaviour was carefully monitored. Those deemed to be behaving with insufficient aggression were instructed to become more aggressive.

In December 2001, British researchers carried out the BBC prison study, which resembled the Stanford prison experiment. However, the findings were very different (Reicher & Haslam, 2006). The guards failed to identify with their role, whereas the prisoners increasingly identified with theirs. As a result, the guards were reluctant to impose their authority on the prisoners and so were overcome by them. The findings were different because the pressure Zimbardo applied to the guards to act aggressively was missing from the BBC study. It may also be relevant that the guards in the BBC prison study knew their actions would be seen by millions on television.

EVALUATION

· Zimbardo apparently showed that situational factors can have dramatic effects on behaviour even in stable individuals.

· Much of the brutality exhibited by the guards occurred because they were doing what Zimbardo had explicitly instructed them to do. According to one guard (John Mark), “Zimbardo went out of his way to create tension. Things like forced sleep deprivation … he knew what he wanted and then tried to shape the experiment” (Mark, 2011).

· Zimbardo claimed that personality plays no role in prison violence. However, he did not investigate individual differences in personality in his study and so cannot strictly draw any conclusions on this issue. Do you think the findings would have been the same if the prison guards had been timid nuns?

· Several guards did not act brutally which shows the situation was not all-important (Haslam & Reicher, 2012). This finding also suggests that individual differences in personality were important, even though Haney and Zimbardo (2009) claimed the experiment showed the opposite.

· The failure of the BBC study to repeat Zimbardo’s findings casts doubt on the importance of situational factors on behaviour.

· There are serious ethical problems in exposing people to degradation and hostility. The whole experience was very distressing to most of the participants and one “prisoner” had to be released early because he was so upset by the way he had been treated. Several prerequisites for ethical research (e.g., informed consent; avoidance of deception; protecting participants from harm; confidentiality of information about the participants) were not adhered to.

Crowd behaviour

Individuals often behave differently when in a crowd than when on their own or with a few friends. For example, lynch mobs in the southern United States murdered 2,000 people (mostly blacks) in the twentieth century. It is improbable that those involved would have behaved like that if they had not been part of a highly emotional crowd.

Le Bon (1895) argued that the anonymity of individuals in a crowd or mob often removes normal social constraints and so leads to violence and panic. In similar fashion, Baumeister et al. (2016, p. 10) emphasised the importance of deindividuation: “a temporary reduction in self-awareness, personal responsibility, and evaluation apprehension, usually brought about by immersing the self in a group”.

Negative views of crowd behaviour remain popular. Schweingruber and Wohlstein (2005) identified the following characteristics often ascribed to crowds: (1) spontaneous; (2) destructive; (3) irrational; (4) highly emotional; and (5) suggestible (the last two characteristics occur due to anonymity and unanimity within the crowd). If crowds do possess these characteristics, then we would predict that they should typically behave in antisocial and aggressive ways.

Reicher et al. (1995, 2016) disputed the above negative view of crowds in their social identity model. They disagreed that deindividuated individuals become uninhibited and freed from social constraints. Indeed, they argued that almost the opposite is the case — deindividuated individuals’ behaviour is strongly influenced by the crowd’s social norms (standards of behaviour) and so often does not lead to antisocial behaviour.

Drury (2018) developed the above ideas. He argued that,

Collective behaviour is possible in ad hoc crowds in an emergency or disaster through participants sharing an (emergent) social identity in the event … this shared social identity is the basis of the motivation to support others affected by the emergency … the source of shared identity in these situations is the common fate experienced by survivors (p. 39).

Drury and Reicher (2010, p. 61) used the term “collective resilience” to refer to the solidarity created by shared identity.

KEY TERMS

Deindividuation: loss of a sense of personal identity; it can occur in a large group or crowd.

Social norms: standards or rules of behaviour operating within a group.

IN THE REAL WORLD: mass panic?

As Drury and Reicher (2010, p. 60) pointed out, “Hardly any self-respecting Hollywood disaster movie would be complete without one scene of people running wildly in all directions and screaming hysterically.” That may help to explain why 80% of people believe crowds panic in threatening situations (Dezecache, 2015).

Reports of theatre fires provide apparent support for the above belief. However, careful analysis of such fires indicates that the inadequate structure of the theatres and insufficient emergency exits played a much larger role than panic (Chertkoff & Kushigian, 1999). In addition, the term “panic” has no precise scientific definition — many different behaviours (including rational ones) exhibited by members of a crowd could be interpreted as indicating panic (Haghani et al., 2019).

Most research is inconsistent with the notion that crowds panic when confronted by an emergency. For example, Drury and Reicher (2010) interviewed many of those present during the bombings in London on 7th July 2005. Most experienced a sense of togetherness on that day and used numerous positive terms (e.g., “unity”; “affinity”; “part of a group”; “warmness”) to describe their feelings. I was working in central London when the bombings occurred and witnessed first-hand the solidarity and calm exhibited by most Londoners.

Dezecache (2015) discussed many other examples of behaviour in mass emergency situations, most showing evidence of social identity. For example, Proulx and Fahy (2004) analysed the reports of 435 survivors of the attacks on the World Trade Center in New York on 11th September 2001 (see Figure 5.8). Mutual help was highlighted in 46% of the reports and 57% of the survivors perceived other people as behaving calmly in reaction to the situation.

Image

Figure 15.8

Drury et al. (2009) found in interviews with survivors of 11 mass emergency events that 57% identified strongly with other crowd members. Those identifying strongly were more likely than those not identifying strongly to help other survivors.

Findings

There is limited support for Zimbardo’s (1970) deindividuation theory. Mann (1981) analysed newspaper reports of individuals threatening to commit suicide by jumping from a bridge or building. Crowd members who were watching were most likely to encourage the potential suicide to jump when they were fairly anonymous and deindividuated (the crowd was large or the incident took place after dark). Silke (2003) analysed 500 violent attacks carried out in Northern Ireland. Disguised offenders (who can be regarded as deindividuated) inflicted more serious physical injuries on their victims.

Postmes and Spears (1998) reviewed studies on group and crowd behaviour. On average, deindividuation (produced by anonymity, large groups, and so on) was associated with a very small increase in antisocial behaviour. More importantly, deindividuation increased individuals’ adherence to group norms. This often produced positive effects (see earlier box).

EVALUATION

· Systematic analysis of crowd behaviour in numerous situations (including mass emergencies) indicates clearly that crowds do not typically panic and/or engage in violent behaviour.

· The social identity approach provides a coherent account of most of the findings, including the varying effects of deindividuation on crowd behaviour.

· Deindividuation generally increases adherence to group norms. However, the concept of ’deindividuation’ is often imprecisely defined and we have only a limited understanding of the processes associated with it (e.g., switch of attention from the self to the current situation).

· Most research on mass emergencies is necessarily limited. For example, it is assumed theoretically that social identify causes certain actions (e.g., helping behaviour). However, it is also possible that various actions cause increased social identity. More generally, it is not possible to use the experimental method (see Chapter 3) when studying crowd behaviour, which makes it hard to draw definite conclusions.

· Theoretically, it is claimed that whether crowds behave well or badly depends on the social norms to which they conform. However, predicting crowd behaviour depends on our ability to decide which norms will be used in any situation and when those norms will be used (Drury, 2018).

· A limitation with the social identity approach is that it is sometimes hard to measure ’social identity’ or ’group norms’.

Chapter summary

· There is a moderately strong conformity effect in the Asch situation, especially when participants perceive other group members as similar to themselves.

· Conformity in the Asch situation depends on normative influence (desire to be liked or respected) and informational influence (accepting that other group members have superior knowledge). Normative influence is typically greater than informational influence.

· It is arguable that most participants in the Asch situation behave appropriately in that they show respect for other group members and also exhibit accurate visual perception.

· Majorities often influence minorities through compliance, whereas minorities influence majorities through conversion. However, the specific impact that minorities have on majorities may vary between active and victimised minorities.

· In Milgram’s research on obedience to authority, the percentage of fully obedient participants went down when the obviousness of the learner’s plight increased and/or the experimenter’s authority was reduced.

· Milgram exaggerated the extent to which participants entered into a passive agentic state. In fact, may were positively engaged followers of the experimenter and believed their behaviour was appropriate.

· Milgram downplayed evidence indicating that a significant proportion of participants did not believe that the learner was actually receiving electric shocks.

· Group polarisation is influenced by social comparison, persuasive arguments, and group members’ desire to distinguish their group from other groups.

· Groupthink is often associated with very poor decision-making. However, it is hard to establish that groupthink actually causes bad decisions, and groupthink is almost as common in successful groups as in unsuccessful ones.

· The Stanford prison experiment suggested that brutality in American prisons was due to the power structure within prisons. However, the experiment was heavily biased in favour of the reported findings, and those findings were not replicated in the BBC study.

· It is often assumed that crowds are spontaneous, destructive, irrational, highly emotional, and suggestible and that some of these features are due to deindividuation within crowds. In fact, deindividuation increases adherence to group norms, and seldom leads to antisocial behaviour.

· It is often assumed that crowds confronted by a serious emergency will tend to panic. In fact, what is much more common is that crowds react calmly, develop a shared social identity, and engage in mutual help.

Further reading

· Drury, J. (2018). The role of social identity processes in mass emergency behaviour: An integrative review. European Review of Social Psychology, 29, 38—81. Theoretical accounts of crowd behaviour in mass emergencies are discussed thoroughly by John Drury.

· Hogg, M.A., & Vaughan, G.M. (2018). Social psychology, 8th Edn. Harlow: Pearson. Chapter 7 in this textbook by Michael Hogg and Graham Vaughan focuses on several topics related to social influence.

· Le Texier, T. (2019). Debunking the Stanford Prison Experiment. American Psychologist, 74, 823—839. Thibault Le Texier discusses at length the deficiencies with Zimbardo’s Stanford prison experiment and the very misleading way he described the experiment and its findings.

· Perry, G., Brannigan, A., Wanner, R.A., & Stam, H. (2020). Credibility and incredulity in Milgram’s obedience experiments: A re-analysis of an unpublished test. Social Psychology Quarterly, 83, 88—106. Gina Perry and colleagues raise several important issues about the proper interpretation of Milgram’s research on obedience to authority.

· Sowden, S., Koletsi, S., Lymberopoulos, E., Militaru, E., Catmur, C., & Bird, G. (2018). Quantifying compliance and acceptance through public and private social conformity. Consciousness and Cognition, 65, 359—367. Sophie Sowden and colleagues provide a very clear account of the major factors at work in Asch-type research on conformity.

Essay questions

1. What are some of the main reasons why people conform? To what extent should we criticise people for conforming?

2. Discuss the strengths and limitations of research on obedience to authority. How relevant is this research to real-life situations?

3. “The Stanford prison experiment proved that prison guards behave aggressively because of the power structure in prisons rather than because of their personality.” Discuss.

4. What factors are involved in group decision making?

5. How do crowds typically behave when confronted by a serious emergency? What factors are important in determining their behaviour in such circumstances?

You probably like most of the people you meet to a greater or lesser extent. However, there are probably a few whom you dislike. What kinds of information do you think influence your reactions to other people? Are you most affected by their personality, their intelligence, or their personal appearance? Why do you find some individuals more attractive than others? Do the factors determining romantic attraction to members of the opposite sex differ between men and women?