Social
Psychology.

A first wordWhat it can show.

Social psychology is the experimental study of how people think, feel, and act in the actual or imagined presence of other people — conformity, obedience, attribution, prejudice, persuasion, the small alterations in behaviour caused by being watched.

The field is famous, even outside the academy, because its mid-century findings were genuinely shocking. Solomon Asch, Stanley Milgram, Leon Festinger, Philip Zimbardo — these are names known to undergraduates who could not name a single cognitive psychologist. They are also, in part, names whose findings have aged unevenly. The discipline that grew up on dramatic single studies has spent the past decade learning to live with replication.

This deck surveys the field in thirty leaves: its mid-century landmarks; the methodological critiques that demolished some of them and rehabilitated others; the calmer, sturdier programmes that have outlasted the controversies; and the books a serious reader should reach for.

Chapter IWhat it studies.

Gordon Allport's working definition, in the 1954 Handbook of Social Psychology, is still the standard: social psychology is the attempt to understand and explain how the thought, feeling, and behaviour of individuals are influenced by the actual, imagined, or implied presence of others.

Five core questions

1. Conformity and obedience. When do people go along with the group, or with authority, against their own judgement?
2. Attitudes and persuasion. What does it take to change a mind, and how stable is the change?
3. Attribution. How do people explain their own and others' behaviour, and where does the explanation systematically err?
4. Group processes. What does being in a group do to a person — identity, polarisation, deindividuation, performance?
5. Intergroup relations. Prejudice, stereotype, contact, conflict, cooperation between groups.

What it cannot show: causation in the wild. The lab gives clean inference at the cost of ecological scope. The field is constantly negotiating the trade.

Chapter IILewin and the postwar shift.

Kurt Lewin (1890–1947), trained in Berlin, fled Nazi Germany in 1933 and reinvented himself in the American Midwest. His field theory — the claim that behaviour is a function of the person and the environment, B = f(P, E), each understood as a dynamic field of forces — became the founding metaphor of modern social psychology.

Lewin's MIT Research Center for Group Dynamics, founded in 1944, trained the cohort that built the field: Festinger, Cartwright, Deutsch. His students went on to invent the T-group, the laboratory study of leadership styles, and the techniques of action research that would later become organisational development.

The postwar shift Lewin presided over was a turn from descriptive to experimental. Where European social psychology had been speculative and synoptic, the new American field was experimental and small-scale: bring strangers into a room, manipulate one variable, measure the response. Almost everything that follows in this deck depends on that move.

Chapter IIIThe birth of group norms.

Muzafer Sherif's The Psychology of Social Norms (1936) is the first classic of the discipline. Using the autokinetic effect — an optical illusion in which a stationary light in a dark room appears to drift — Sherif showed that subjects in groups quickly converged on a shared estimate of how far the light moved, and that the convergence persisted even when subjects later returned to estimating alone.

The finding has been replicated for generations. It establishes, in cleanest form, the social construction of judgement under uncertainty: when you do not know the right answer, you take other people's answers as data. The norm becomes invisible because it is internalised.

Robbers Cave 1954

Sherif's other landmark: a summer camp in Oklahoma where two groups of 11-year-old boys, randomly assigned, generated genuine intergroup hostility within days — and whose hostility was reduced not by mere contact but only by superordinate goals requiring the groups to cooperate. The study has its critics — Gina Perry's The Lost Boys (2018) has questioned how scripted the camp was — but the cooperative-goal finding has held up in subsequent work.

Chapter IVAsch's line studies.

Solomon Asch's conformity experiments at Swarthmore in 1951–55 are the cleanest demonstration of normative pressure ever published. The task was a perceptual one: which of three lines matches the target? The correct answer was always obvious to a control subject working alone — over 99% accuracy.

In the social condition, with confederates unanimously wrong, the picture shifted sharply. Roughly 75% of subjects conformed at least once across the trials, with about 32% conforming on average across critical trials. Subjects in post-experiment interviews split between those who knew the group was wrong but went along, and those who came to doubt their own perception.

What Asch's work showed, against the older "majority rules" cliche, was the importance of unanimity: a single confederate breaking ranks — even by giving a different wrong answer — cut conformity by about three quarters. The lever is not majority size; it is unbroken consensus.

The finding has replicated cross-culturally, with collectivist cultures showing somewhat higher conformity (Bond & Smith, 1996, meta-analysis). The basic phenomenon is solid.

Chapter VMilgram and obedience.

Stanley Milgram, a 27-year-old Yale assistant professor and a former student of Asch's, published his first obedience paper in 1963. The headline figure — 65% of subjects continued administering shocks to the maximum 450 volts despite the learner's escalating distress and apparent incapacitation — reset the discipline.

The original paradigm was actually 18 variants. The 65% headline came from one. In other variants, with the experimenter on the phone rather than in the room, with two experimenters disagreeing, with a peer refusing to participate, full obedience dropped sharply — sometimes below 10%. The lesson was less "people will obey" than "the situation does most of the work."

"Obedience is as basic an element in the structure of social life as one can point to."— Milgram, 1974

Milgram extended the work into Obedience to Authority (1974). The book is more nuanced than the famous result; it is also — as the field has lately had to admit — methodologically less clean than it claimed.

Chapter VIMilgram’s troubles.

Gina Perry, working through the Milgram archives at Yale in the late 2000s, produced two findings that have changed how the field reads the original studies. The first: many subjects, in post-experiment interviews and follow-up questionnaires, said they had not believed the shocks were real. Belief in the cover story varied across conditions and was not consistently reported.

The second, more troubling: the experimenter departed from the four-prod script. In many sessions he improvised, used coercive language, and refused subjects who tried to leave. The experiment was less an experimental manipulation than a varied, sometimes ad-hoc social interaction.

This does not mean Milgram's work was fraudulent or that obedience is a myth. The basic phenomenon — that ordinary people, under specific situational pressures from a perceived authority, will perform acts they find distressing — is robust and has been partly replicated by Jerry Burger (2009) at Santa Clara, with stricter ethics and stopped at 150 volts. About 70% continued past 150, comparable to Milgram's 79% at the same point.

The truer lesson: situational pressure is real, single-study certainty is not, and the gap between Milgram's findings and the parable Milgram told about them was always larger than he let on.

Chapter VIIStanford prison.

Philip Zimbardo's Stanford Prison Experiment, conducted in August 1971, was for forty years the field's most-cited cautionary tale about situational pressure. The story Zimbardo told — that ordinary college students, given guard uniforms and authority, became sadistic; that prisoners broke down; that the situation had taken over — was retold in textbooks, the 2015 film, and Zimbardo's own The Lucifer Effect (2007).

The 2018 demolition

The story has, since Thibault Le Texier's archival work and the 2018 reporting by Ben Blum, largely fallen apart. Zimbardo and his graduate-student warden coached guards on how to behave ("be tough"); prisoners reported they were performing what they took to be expected of them; one famous breakdown was, the prisoner later said, partly faked to get out of an exam he had to study for.

The SPE was not an experiment so much as a piece of demonstration theatre with a predetermined conclusion. Stephen Reicher and Alex Haslam's BBC Prison Study (2002), an actual experiment with random assignment and proper controls, found something close to the opposite of Zimbardo's claim: people in groups did not slide passively into their roles; they actively negotiated identities, and tyranny required collusion, not mere assignment.

The SPE is now textbook canon as a cautionary tale about how a story can outrun its evidence.

Chapter VIIIThe crisis the field went through.

By 2015 it was clear the discipline had a problem. Daniel Kahneman wrote an open letter in 2012 calling for replication of priming research; Bem's 2011 ESP paper, methodologically standard for the time, had passed peer review; the Stapel fraud case had broken in the Netherlands; the Open Science Collaboration's Estimating the reproducibility of psychological science (2015) provided the numbers.

What did not replicate

Much of social priming (professor priming, money priming, elderly walking-speed). Power posing's hormonal effects. Ego depletion at the original effect size. Many implicit-bias-as-behaviour findings. The weapons effect on aggression. Some facial-feedback claims. Stereotype threat at the original size in some populations.

What did

Asch conformity. The basic Milgram phenomenon. The fundamental attribution error in Western samples. Cognitive dissonance under post-decisional regret. The basic contact-hypothesis effect (Pettigrew & Tropp 2006 meta-analysis, 515 studies). Social facilitation. Most attitude-behaviour findings under proper measurement. Group polarisation.

The field's response — pre-registration, registered reports, Many Labs collaborations, larger samples, transparent analyses — has been more thorough than economics's or biomedicine's. The discipline is in better shape, intellectually, than it was a decade ago.

Chapter IXCognitive dissonance.

Leon Festinger's A Theory of Cognitive Dissonance (1957) is, with Asch and Milgram, the foundation of the modern field. The theory: holding two inconsistent cognitions produces an aversive state, and the person reduces it by changing one cognition.

The Festinger & Carlsmith $1 / $20 study (1959) is the canonical demonstration. Subjects performed a tedious peg-turning task, then were paid either $1 or $20 to tell the next subject it was interesting. The $1 subjects later reported the task as actually interesting; the $20 subjects did not. The reasoning: $20 was sufficient justification for the lie ("I lied for the money"), so no dissonance; $1 was not, so the only way to reduce the inconsistency was to come to believe the lie.

The theory generalises widely: the more freely a choice is made, the harder it is to reverse; people who suffer to join groups value them more (Aronson & Mills 1959, severe initiation); cult members whose prophecies fail double down rather than disconfirm (When Prophecy Fails, 1956). Dissonance theory is among the very few mid-century social-psych frameworks to have survived replication essentially intact.

Chapter XBem and the alternative.

Daryl Bem's self-perception theory (1972) proposed an alternative reading of dissonance results: maybe subjects are not reducing aversive arousal, maybe they are simply inferring their attitudes from their behaviour, the way an outside observer would. ("I told someone the task was interesting; people who say that find tasks interesting; therefore I find it interesting.") No internal aversive state needed.

For a decade the two theories looked empirically indistinguishable. The resolution came from Mark Zanna and Joel Cooper's 1974 misattribution studies: when subjects were given a placebo described as causing arousal, dissonance effects disappeared. They had attributed their arousal to the pill, not to the inconsistency, so the cognitive change did not occur. That predicts the dissonance pattern but not the self-perception pattern.

The synthesis, which most textbooks now teach: dissonance applies when the attitude-discrepant behaviour produces a real arousal response (the subject cared, the choice was theirs); self-perception applies when arousal is low and the inference is essentially observational. Both processes are real; they operate in different conditions.

Chapter XIHow we explain behaviour.

Fritz Heider's The Psychology of Interpersonal Relations (1958) launched the attribution programme: people are naïve scientists, perpetually trying to explain why others (and they themselves) act as they do, and the explanations cluster around two poles — dispositional (it's their character) and situational (it's the circumstances).

Edward Jones and Keith Davis (1965) and Harold Kelley (1967) formalised the inferential rules. The empirical surprise — documented by Lee Ross in 1977, who named it — was the fundamental attribution error: observers reliably under-weight situational factors when explaining others' behaviour. Even when subjects know an essay was assigned, they still infer the writer's true views from it. Even when Milgram's experimenter is in the room, observers blame the teacher.

The actor-observer asymmetry

Jones & Nisbett (1971): I trip because the sidewalk was uneven; you trip because you are clumsy. Subsequent work (Malle 2006 meta-analysis) has tempered this — the asymmetry is smaller and more conditional than the original claim. The basic FAE pattern, however, is robust in Western samples and partial in collectivist ones (Choi, Nisbett, Norenzayan 1999).

Chapter XIITajfel and identity.

Henri Tajfel, a Polish-Jewish refugee who had survived the war as a French prisoner-of-war (the Nazis never confirmed he was a Jew), spent his Bristol career trying to understand how prejudice and intergroup violence become possible. His minimal group paradigm (1971) was a stripped-down attempt to find the lower bound of in-group bias.

Subjects were sorted into groups on the most trivial possible basis — a tossed coin, a preference for Klee over Kandinsky — with no contact, no history, no shared interest. Asked to allocate small monetary rewards between an in-group and out-group member, they reliably favoured the in-group, even at the cost of maximum joint gain. Group identity was sufficient. No content needed.

Social Identity Theory

Tajfel and John Turner's 1979 SIT proposed that part of the self-concept derives from group memberships, and that people maintain self-esteem partly by maintaining favourable comparisons between in-groups and out-groups. The framework underlies most modern work on prejudice, nationalism, organisational behaviour, and political polarisation. Its later refinement, self-categorisation theory (Turner et al., 1987), is the technical extension.

Chapter XIIIBandura's social learning.

Albert Bandura's Bobo doll experiments (1961, 1963, 1965) were the most direct refutation of strict behaviourism's claim that learning requires reinforcement. Children who simply watched an adult attack the doll later attacked it themselves — with no reinforcement, no instruction, just observation.

Bandura's Social Learning Theory (1977) generalised the finding into a framework: people learn by observation, by model, by imitation, especially from high-status or similar models. The implications for media violence, for parenting, for behavioural therapy, were immediate and large. The framework was renamed social cognitive theory in 1986 to mark its incorporation of cognitive variables.

Self-efficacy

Bandura's later concept — the belief that one is capable of executing the actions needed to produce given attainments — is the most cited construct in personality and motivation research. Self-efficacy predicts academic performance, athletic performance, smoking cessation, response to therapy. It is increased by mastery experiences (the strongest source), vicarious experience, social persuasion, and physiological feedback.

Bandura was, by the time of his death in 2021, the most-cited living psychologist.

Chapter XIVThe bystander effect.

The Genovese case prompted Bibb Latané and John Darley's series of experiments at Columbia and NYU in 1968–70. The basic finding: the more people present at an emergency, the less likely any one of them is to act. The pattern arises from diffusion of responsibility (anyone could; therefore no one must) and pluralistic ignorance (everyone is reading everyone else's calm as evidence the situation is not an emergency).

The smoke-filled room study (1968): a single subject reported smoke 75% of the time; subjects in groups of three with two confederates who ignored it reported it 10% of the time. The seizure study (1968): subjects who believed they were the only listener intervened 85% of the time; those who believed five others were also listening, 31%.

The bystander effect is robust, though the Genovese parable that launched it was, as later journalism (notably Rachel Manning, Mark Levine, and Alan Collins's 2007 reanalysis) showed, factually distorted. The phenomenon survives the correction.

Chapter XVWhat groups do to thinking.

The classic claim, going back to Le Bon's 1895 The Crowd, was that groups make people stupider and more violent. The empirical record is more interesting and more specific.

Group polarisation

Stoner's 1961 risky-shift finding — groups make riskier decisions than their members alone — turned out (Moscovici & Zavalloni 1969) to be a special case of a more general phenomenon: groups amplify the dominant tendency in their members' initial views. Conservative groups become more conservative; cautious groups more cautious; juries pre-disposed to convict become more so. The effect is mediated by social comparison and persuasive-arguments processes both.

Groupthink

Janis's case-history method has been criticised for selection on the dependent variable (he picked failures and looked for the syndrome). Controlled tests have found weaker effects than the case studies suggested. But the core claim — that high cohesion plus directive leadership plus stress can shut down internal dissent — is broadly supported.

What works against it: explicit dissent norms, devil's advocates, anonymous initial proposals. The applied lesson has reshaped corporate-board and military-staff procedure.

Chapter XVIHow attitudes change.

The Yale persuasion programme of the 1950s, led by Carl Hovland, mapped the variables that mattered: who says it (source credibility), what they say (message factors), to whom (audience characteristics), in what setting (channel and context). The programme was cumulative and empirical and produced a generation of textbooks.

Richard Petty and John Cacioppo's Elaboration Likelihood Model (1986) gave it a process theory. When motivation and ability to think are high, persuasion proceeds through the central route: arguments matter, and the resulting attitudes are stable and predictive of behaviour. When motivation or ability are low, peripheral cues do the work, and the resulting attitudes are weaker.

What changes minds in practice

Two-sided arguments outperform one-sided ones for educated audiences and where opposing views will be encountered. Inoculation (McGuire 1964) — pre-exposing people to weak versions of counterarguments — protects against subsequent persuasion. Self-generated arguments are more persuasive than externally provided ones. Most political-campaign persuasion attempts have effects close to zero (Kalla & Broockman 2018, meta-analysis). Persuasion happens, but it is harder and rarer than the field's old optimism implied.

Chapter XVIIThe compliance professionals.

Robert Cialdini's six (later seven) principles of influence are, in their original form: reciprocity (small unsolicited gifts produce disproportionate compliance — the Hare Krishna flower); commitment and consistency (small initial agreements lead to large later ones — the foot-in-the-door technique, Freedman & Fraser 1966); social proof (people do what others are doing — canned laughter, "best-seller" lists); authority (Milgram on a smaller scale — the lab coat, the title); liking (we comply with people we like — physical attraction, similarity, compliments); scarcity (limited availability increases perceived value — the deadline, "while supplies last"). The seventh, added in the 2016 Pre-Suasion, is unity (shared identity beats mere liking).

Cialdini's work has been more replicable than most social psychology — possibly because the effects were observed in field settings before they were formalised in the lab. The principles are a working toolkit for advertising, fundraising, and behavioural-policy interventions, and an excellent guide to recognising when one is being worked on.

Chapter XVIIIWhat people say vs. what they do.

The gap between professed attitude and observed behaviour was social psychology's worry from the 1930s through the 1960s. Allan Wicker's 1969 review concluded that attitudes typically predicted behaviour at correlations of about 0.30 or less. Many in the field considered the attitude concept on its way out.

Martin Fishbein and Icek Ajzen's Theory of Reasoned Action (1975) and its successor, the Theory of Planned Behaviour (Ajzen 1991), rescued it. The corrections were both theoretical and methodological. Theoretically: behaviour is predicted by intention, which is predicted by attitude toward the specific behaviour, subjective norm, and perceived behavioural control — not by global attitudes toward the object. Methodologically: measure attitudes at the same level of specificity as the behaviour (attitude toward jogging tomorrow morning, not toward exercise in general).

With those corrections, attitude-behaviour correlations rise to 0.6 and above, and the theory of planned behaviour is one of the more useful applied tools in health and environmental psychology. The original "attitudes don't predict behaviour" claim was wrong, but for an instructive reason: the attitude was being measured at the wrong grain.

Chapter XIXThe contact hypothesis.

Gordon Allport's The Nature of Prejudice (1954), commissioned in the wake of Brown v. Board, is one of the few books in social psychology that genuinely changed national policy. Its central claim — that prejudice is reduced by contact between groups under specific conditions — underwrote desegregation arguments before the Supreme Court and a generation of school and workplace integration efforts.

The empirical case has been examined more thoroughly than almost any in the field. Thomas Pettigrew and Linda Tropp's 2006 meta-analysis of 515 studies and over 250,000 subjects found a robust contact-prejudice reduction effect (mean r ≈ -0.21). Allport's four optimal conditions strengthen the effect but are not strictly necessary; even casual contact reduces prejudice, on average.

Where the effect is weaker: highly threatened in-group members; situations in which the contact reinforces stereotypes; intergroup violence whose roots are structural rather than perceptual. The contact hypothesis is one of the field's success stories — a consistent, replicable, policy-relevant finding — though not a universal solvent for prejudice.

Chapter XXImplicit Association Test.

Anthony Greenwald, Mahzarin Banaji, and Brian Nosek's Implicit Association Test (1998) was the first widely-used reaction-time measure of implicit attitudes. The Project Implicit website has run more than thirty million IATs since 2002 and remains one of the largest psychology data collections in existence.

The IAT showed two robust things. First: the average American shows a faster pairing of "white" with "good" than "Black" with "good" — including, on average, Black respondents themselves, though less strongly. Second: explicit attitudes and IAT scores correlate weakly, suggesting they measure something at least partially distinct.

The contested interpretation

What the IAT is measuring — individual prejudice, cultural knowledge of stereotypes, mere familiarity, or measurement noise — remains debated. Hart Blanton, James Jaccard, and Frederick Oswald have argued the test-retest reliability is too low (around 0.4) to support individual diagnosis. Bertram Gawronski's 2019 meta-analysis found IAT-behaviour correlations of about 0.15 — nontrivial but small. The diagnostic use of IAT scores in employment or training settings is not warranted by the data; the use as a population-level measure of attitudinal residue is more defensible.

Chapter XXIPerformance under threat.

Claude Steele's account, developed with Joshua Aronson and elaborated in Whistling Vivaldi (2010), proposed that stigmatised group members carry a cognitive load — the worry of confirming the negative stereotype — that interferes with performance under high-stakes evaluation. The original 1995 paper was hugely cited and influenced testing policy.

The replication record has been mixed. Several large pre-registered replications have found weaker effects or null effects, particularly among children and in the math-and-women variant. Paulette Flore and Jelte Wicherts's 2015 meta-analysis found a small overall effect with strong evidence of publication bias; subsequent registered replications have moderated rather than abolished the claim.

The current consensus, more carefully stated: stereotype threat exists as a phenomenon, particularly in older students under high-stakes conditions, but the effect sizes are smaller and the moderators more numerous than the original literature suggested. It is not a complete explanation of group-level test gaps. It is a real source of variance under specifiable conditions.

Chapter XXIIWhat makes people aggressive.

The early frustration-aggression hypothesis (Dollard, Doob, Miller, Mowrer, Sears, 1939) was simple: blocked goals lead to aggression. Leonard Berkowitz refined it: frustration produces a readiness to aggress, which aggressive cues then channel.

Craig Anderson and Brad Bushman's General Aggression Model (2002) is the field's current best synthesis — person variables (trait aggression, beliefs), situation variables (provocation, frustration, cues), and internal states (arousal, cognition, affect) interact to produce aggressive thought, feeling, and behaviour.

Violent video games

The most public application has been the video-game debate. Anderson and Bushman's meta-analyses (1998, 2010) found small-to-moderate effects of violent-game exposure on short-term aggressive thought and affect. Christopher Ferguson and Markey have argued the effects are largely artifacts of measurement and publication bias. The 2020 APA task-force review and several large pre-registered studies have failed to replicate the original effects at meaningful sizes. The current honest answer: any effect is small enough that it is hard to detect cleanly. The strong claims of the 2000s overran the data.

Chapter XXIIIThe science of close relationships.

Robert Sternberg's triangular theory (1986) decomposes love into three components — intimacy, passion, commitment — whose various combinations produce romantic love (intimacy + passion), companionate love (intimacy + commitment), consummate love (all three), and the rest. The framework is taxonomic, not predictive, but it has held up as a useful descriptive map.

Elaine Hatfield and Ellen Berscheid's earlier distinction between passionate love (intense, transient) and companionate love (deep, durable) is the empirical scaffolding underneath Sternberg.

The Gottman programme

John and Julie Gottman's Seattle "love lab" has spent decades videotaping couples discussing conflicts and coding their interactions in microscopic detail. The signature finding: certain interaction patterns — particularly the ratio of positive to negative exchanges and the presence or absence of contempt — predict divorce with reported accuracy in the 80–90% range over multi-year follow-ups.

The 90% accuracy claim has been challenged on methodological grounds (the predictions were made post-hoc on the same sample, in some studies). But the general finding — that the way couples handle disagreement, more than the disagreements themselves, predicts marital outcome — has held.

Chapter XXIVThe cultural turn.

For most of its history social psychology assumed its findings applied to people in general. Hazel Markus and Shinobu Kitayama's Culture and the Self (1991) was the major intervention against the assumption. They argued that the "independent" self — bounded, stable, internally consistent — that animated American social-psychology theory was a cultural construction, contrasted with the "interdependent" self more characteristic of East Asian and many other societies, in which the self is constituted by relationships and shifts contextually.

The downstream evidence has been substantial. The fundamental attribution error is weaker in East Asian samples (Choi, Nisbett, Norenzayan 1999). Cognitive dissonance shows different patterns under self vs. other framings (Hoshino-Browne et al. 2005). Self-enhancement is reliable in American samples and largely absent in Japanese ones (Heine et al. 1999).

Joseph Henrich, Steve Heine, and Ara Norenzayan's The weirdest people in the world? (2010, BBS) sharpened the methodological lesson: psychology had been generalising from a narrow sliver of humanity. The field's response — broader sampling, registered cross-cultural replications, more careful theoretical framing — is ongoing and incomplete.

Chapter XXVSocial psychology in the platform era.

Two distinct questions sit under the same word. First: how do social-media platforms work, as social-psychological systems — what do likes, retweets, algorithmic feeds do to attention, conformity, identity, polarisation? Second: are the field's classic findings still robust in a population whose social experience is increasingly mediated?

On the first: there is now solid evidence that platform-mediated outrage spreads faster than other emotional content (William Brady & Jay Van Bavel, 2017); that political identity has become more affectively polarised in the US over the past two decades (Iyengar et al.); that brief exposure to opposing political views on Twitter increases rather than decreases polarisation (Bail et al. 2018). The social-comparison and pluralistic-ignorance literatures are being substantially reworked under platform conditions.

On the second: the cross-cohort question is unresolved. Jean Twenge's iGen (2017) argues for a sharp generational shift around 2012; Andrew Przybylski and others have argued the effects are smaller than headline claims; the U.S. Surgeon General's 2023 advisory on adolescent mental health found enough signal to warrant action; the Jonathan Haidt Anxious Generation (2024) version is the highest-profile public statement of the strong claim. The honest current position: there is a real signal in adolescent-girl mental-health data after 2012; the precise causal contribution of social media is contested; the size of the effect is debated; some effect is now widely accepted.

Chapter XXVIBooks and landmark papers.

• 1936The Psychology of Social NormsMuzafer Sherif
• 1951Effects of group pressure on judgement (paper)Solomon Asch
• 1954The Nature of PrejudiceGordon Allport
• 1956When Prophecy FailsFestinger, Riecken, Schachter
• 1957A Theory of Cognitive DissonanceLeon Festinger
• 1958The Psychology of Interpersonal RelationsFritz Heider
• 1972The Social Animal (many eds.)Elliot Aronson
• 1974Obedience to AuthorityStanley Milgram
• 1977Social Learning TheoryAlbert Bandura
• 1979Social Identity and Intergroup RelationsHenri Tajfel (ed.)
• 1984InfluenceRobert Cialdini
• 1991Culture and the Self (paper)Markus & Kitayama
• 1995Stereotype threat (paper)Steele & Aronson
• 2002The General Aggression ModelAnderson & Bushman
• 2006Pettigrew & Tropp meta-analysis (JPSP)Pettigrew & Tropp
• 2007The Lucifer EffectPhilip Zimbardo
• 2008Mistakes Were Made (But Not by Me)Tavris & Aronson
• 2010The weirdest people in the world? (paper)Henrich, Heine, Norenzayan
• 2010Whistling VivaldiClaude Steele
• 2013Behind the Shock MachineGina Perry
• 2015Estimating Reproducibility (Science)Open Science Collab.
• 2017iGenJean Twenge
• 2018The Lost BoysGina Perry
• 2024The Anxious GenerationJonathan Haidt

Chapter XXVIIWatch and read.

Fig. · Zimbardo · The psychology of evil · TED 2008

Two more

Asch's original conformity setup is short and worth seeing for the body language: Asch · Conformity Experiment.

The 1959 Festinger and Carlsmith dissonance study, demonstrated and explained in five minutes — one of the cleanest experiments in the field's history: Festinger & Carlsmith · Cognitive dissonance.

And read

Aronson's The Social Animal for the field as a whole. Cialdini's Influence for the applied side. Allport's Nature of Prejudice — still the best book on the topic, seventy years on. Then Henrich, Heine, and Norenzayan to recalibrate.