Types of conformity, and why we conform

Kelman (1958) distinguished levels of conformity. The 7182 specification requires two:

• Compliance — going along with the group publicly while privately disagreeing. It is temporary and stops when the group is gone.
• Internalisation — you genuinely accept the group's view. The change is deep and permanent, and persists in the group's absence.

Spec check: the current specification names only internalisation and compliance. Identification (adopting a behaviour because you value membership of the group) appears in older textbooks but is no longer named in 7182 — know it as context, not as an exam requirement.

Deutsch and Gerard (1955) proposed a two-process explanation:

• Normative social influence (NSI) — an emotional need to be liked and accepted. Leads to compliance. Strongest with strangers, in unfamiliar situations, and with people you want to impress.
• Informational social influence (ISI) — a cognitive need to be right. You assume the majority knows better. Leads to internalisation. Strongest when the task is ambiguous, difficult, or in a crisis, or when others are experts.

Evaluation: the two processes are hard to separate in practice — a dissenter in Asch's study may reduce conformity by giving social support (NSI) or by providing an alternative source of information (ISI). Also, individual differences: people high in the need for affiliation (nAffiliators) are more likely to conform, which NSI alone does not predict.

Which process?

Asch's line-judgement study

Procedure: 123 American male students were tested individually in groups of 6–8 confederates. On each trial they saw a standard line and three comparison lines and said aloud which comparison matched. The answer was obvious. On 12 of 18 trials — the critical trials — the confederates unanimously gave the same wrong answer.

Findings:

• Participants conformed to the wrong answer on 36.8% of the critical trials.
• 75% conformed at least once; 25% never conformed.
• In a control condition (no confederates) the error rate was under 1% — so the task really was unambiguous.

In interviews afterwards, most conformers said they knew the answer was wrong but went along to avoid disapproval — evidence for NSI and compliance.

Evaluation: the task was artificial and trivial (low mundane realism) and the groups were not real groups, so the findings may not generalise. The sample was American men in the 1950s — an era of unusually high conformity — so the study may be a child of its time (Perrin and Spencer, 1980, found much lower conformity with UK engineering students). It is also androcentric and culture-bound: conformity rates tend to be higher in collectivist cultures.

The three variables Asch manipulated

• Group size — conformity rose as the majority grew, but levelled off: with three confederates it was already about 32%, and adding more made little difference. So a small majority is enough.
• Unanimity — introducing a dissenter (a confederate who gave the correct answer, or even a different wrong answer) reduced conformity substantially. Breaking the unanimity of the majority frees the naive participant to act independently.
• Task difficulty — when the comparison lines were made more similar in length, conformity increased. The situation is now ambiguous, so ISI takes over: participants genuinely don't know, so they look to others.

Exam craft: each variable maps onto an explanation. Unanimity and group size point to NSI (fear of standing out); task difficulty points to ISI (need to be right). Saying that in an essay is AO3 gold.

Quote the figure

Milgram's baseline obedience study

Procedure: 40 American men, recruited by newspaper advert for a study on 'memory'. A rigged draw always made the participant the Teacher and a confederate ('Mr Wallace') the Learner. The Teacher gave apparently real electric shocks, rising in 15V steps from 15V to 450V, whenever the Learner made a mistake. An experimenter in a grey lab coat issued four scripted prods ("The experiment requires that you continue", "You have no other choice, you must go on").

Findings:

• 100% of participants continued to 300V.
• 65% continued to the maximum 450V.
• Participants showed extreme tension — sweating, trembling, nervous laughter; three had full-blown seizures.

Beforehand, 14 psychology students had estimated that only about 3% would go to 450V. Obedience was massively under-predicted.

Evaluation: Orne and Holland argued participants guessed the shocks were fake, so the study lacks internal validity. But Sheridan and King found high obedience even when real shocks were given to a puppy, and Hofling et al. (1966) found 21 of 22 nurses obeyed an unknown doctor's phone instruction to give a dangerous drug dose — supporting external validity. Ethically, Milgram deceived participants and arguably failed to make withdrawal feel possible, though he debriefed thoroughly and 84% later said they were glad to have taken part.

Milgram's situational variables

Milgram ran variations to isolate what the situation was doing:

• Proximity — Teacher and Learner in the same room: obedience fell from 65% to 40%. In the touch-proximity condition (Teacher forces the Learner's hand onto a shock plate) it fell to 30%. When the experimenter gave orders by telephone, obedience dropped to 20.5% — and some participants secretly gave weaker shocks.
• Location — moved from prestigious Yale to a run-down office block in Bridgeport: obedience fell to 47.5%. Institutional prestige adds legitimacy, but not much.
• Uniform — when the experimenter was replaced by an 'ordinary member of the public' in everyday clothes, obedience fell to 20% — the lowest of these three variations. Uniform is a powerful symbol of legitimate authority.

Evaluation: the variations were run with high control, and Bickman (1974) supports the uniform effect in the field — people were far more likely to obey a confederate dressed as a security guard than one in a jacket and tie. But the added manipulations make the setup even more artificial, and the obedience alibi criticism (Mandel) warns that explaining atrocities purely by situation lets perpetrators off the hook.

Which variation?

Agentic state & legitimacy of authority

• Agentic state — we act as an agent carrying out another person's wishes, so we feel no personal responsibility. The opposite is the autonomous state, where we direct our own behaviour and feel responsible. The shift is the agentic shift, and it usually occurs in front of someone we perceive as an authority figure.
• Binding factors keep us in the agentic state by reducing the 'moral strain' we feel — shifting blame onto the victim, or denying the damage done.
• Legitimacy of authority — societies are hierarchies, and we are socialised from childhood to accept that some people have the right to exercise power. Legitimacy is signalled by uniform and setting. Its dark side is destructive authority, as when Milgram's experimenter ordered cruelty.

Evaluation: agentic state explains the reduced responsibility participants reported ("I was just doing what I was told"), and Blass and Schmitt found observers blamed the experimenter, not the teacher. But it can't explain gradual disobedience or Hofling's nurses (who showed no anxiety, as agentic-shift theory predicts they should). Legitimacy of authority has strong cross-cultural support — obedience rates vary by country (e.g. higher in Australia than the USA in some replications, lower elsewhere), reflecting how societies teach deference — and it explains real-world atrocity where orders come through a legitimate hierarchy.

The Authoritarian Personality (Adorno et al., 1950)

Adorno et al. studied over 2,000 middle-class white Americans and their unconscious attitudes towards other ethnic groups, using scales including the F-scale (potential for fascism).

People scoring high on the F-scale — the authoritarian personality — showed:

• Extreme respect for authority and submissiveness to it; contempt for those of 'inferior' status.
• Rigid, black-and-white thinking; fixed stereotypes; conventional attitudes to sex, race and gender.
• An origin in a harsh childhood: strict discipline, conditional love and impossibly high expectations create hostility towards parents which cannot safely be expressed, so it is displaced (scapegoated) onto those seen as weaker.

Elms and Milgram (1966) interviewed 20 of Milgram's fully obedient participants and found they scored higher on the F-scale than a comparison group of disobedient participants — support for a dispositional element.

Evaluation: the link is only a correlation, so a third variable (e.g. low education) may explain both. It cannot explain whole populations obeying — millions of Germans could not all share one personality type, which makes social identity a better explanation. The F-scale also suffers acquiescence bias (all items are worded in the same direction) and is politically biased, equating authoritarianism with right-wing ideology only.

Agentic state

Consistency or flexibility?

Whose evidence is it?

Tap a card, then the study or explanation it belongs to. Get the names attached to the right findings — examiners reward accurate referencing.

Social support and locus of control

• Social support — the pressure to conform is reduced if someone else also resists. In Asch's variations, a dissenter freed the participant to act independently even when the dissenter gave a different wrong answer, showing that what matters is breaking the unanimity, not being right. In Milgram's variation with two disobedient confederates, obedience dropped to 10%. The dissenter acts as a model of independence.
• Locus of control (Rotter, 1966) — a continuum from high internal (I control what happens to me) to high external (my life is shaped by luck, fate and other people). Internals are more likely to resist: they take personal responsibility, are more self-confident and more achievement-oriented, and are less concerned with being liked.

Evaluation: Holland (1967) re-ran Milgram and found 37% of internals resisted (did not go to 450V) compared with 23% of externals — supporting the role of LOC. But Twenge et al. (2004) found that over 40 years Americans became more resistant yet more external — the opposite of the prediction — and LOC only really matters in new situations, not familiar ones where past experience dominates. So LOC is a limited, though genuine, explanation.

Who resists?

Moscovici — how a minority converts a majority

Procedure: Moscovici et al. (1969) put groups of six women — four naive participants and two confederates — in front of 36 slides that were all clearly blue, varying in intensity. Participants stated the colour aloud.

• Consistent condition: the confederates called every slide green. Naive participants agreed on 8.42% of trials, and 32% gave the green answer at least once.
• Inconsistent condition: confederates said green on two-thirds of trials. Agreement fell to just 1.25%.

The three key behaviours of an effective minority:

• Consistency — the same message over time (diachronic) and agreement between members (synchronic). It makes people rethink.
• Commitment — demonstrating personal sacrifice or risk (the augmentation principle) shows you are not acting out of self-interest.
• Flexibility — Nemeth argued that rigid, dogmatic consistency is off-putting; a minority must be prepared to compromise and adapt.

Minority influence works through deeper processing: those exposed to a consistent minority think harder about the message, so change is more likely to be internalisation — private, lasting conversion — rather than mere compliance.

Evaluation: Moscovici's task (naming colours) is artificial, far from real-world minority influence such as jury deliberation or political campaigning, where the stakes are far higher. But Wood et al.'s meta-analysis of nearly 100 similar studies found consistent minorities were the most influential — a solid evidential base.

The key ingredient

Term and definition

Tap an item on the left, then its partner on the right.

The big ideas to know

You have covered the key ideas in AQA 4.1.1 Social influence. Press Finish to see your score.