Social Intelligence curriculum

Situationist research found that minor details of a setting, such as being in a hurry or seeing a bystander do nothing, can change what people do more than their character would lead you to expect. Most researchers now hold an interactionist view: traits and situations both matter. So when a colleague ignores a request, ask what was pressing on him before you decide what kind of person he is.

Construal is your interpretation of a situation, and two people can read the same facts very differently. The trap is naive realism: assuming your reading is just how things are, so anyone who disagrees must lack facts or be biased. In one experiment, people given about half the facts felt as well informed as people given all of them. Before judging someone's reaction, ask what they think is happening.

Tacit knowledge is know-how you use but cannot fully put into words, like sensing that a joke would land badly in a new office. Researchers suggest much social intuition comes from implicit learning, where you absorb patterns without noticing you learned anything. Since even skilled people cannot say exactly what they know, spoken rules capture only part of it. Watch how insiders act, and test your reads in low-stakes moments.

Channel factors, a term credited to the psychologist Kurt Lewin, are small features of a situation that make an action easier or harder, such as whether the place, time and next step are spelled out. In a classic study, a leaflet showing a campus clinic's location and hours, with a prompt to plan a visit, raised students' turnout for tetanus shots. When people skip what they meant to do, look for friction first.

Social intelligence, the ability to understand people and social situations, is related to general reasoning but is not the same thing. A meta-analysis of 87 studies found people-centered abilities correlate more with each other than with abilities about things like numbers or space. Yet different social intelligence tests agreed only modestly. So a strong exam record is only partial evidence of how well someone reads a room.

Interpersonal accuracy is scored by checking a perceiver's guesses against a real criterion, often what the person reports they were thinking or feeling at the time. Ability researchers favor performance tasks over self-ratings, since a self-rating partly reflects how skilled you feel. A fluent theory about a coworker's mood is still a guess until something confirms it. Keep score of when your reads turn out right.

WEIRD stands for Western, educated, industrialized, rich and democratic, the kind of sample most psychology research has used. One review of leading journals found 96 percent of samples came from Western countries. Some classic cultural differences have held up while others are being re-examined, so neither sameness nor difference is safe to assume. Ask who was studied before you apply a finding to the people around you.

The primacy effect is the tendency for early information to set the tone of an impression and bend how later facts are read. In one experiment, advisers who started out dishonest were trusted less at the end, even though all advisers were equally honest overall. But a study tracking impressions over six days found new contrary news moved them more than it warranted. Revisit first reads on purpose.

The halo effect is assuming that someone with one good quality has other good qualities too; the horn or devil effect is the negative version. A well-documented case is attractiveness, which raises ratings of intelligence and trustworthiness, though one large study found the lift leveling off, and it can bring assumptions such as vanity. If a candidate's polish impresses you, rate each skill against its own evidence.

Expectations steer both questions and treatment. In one study, people who crave quick certainty leaned toward questions that could disprove their hunch when fresh, but toward confirming ones when drained. Treatment can draw out what you predicted: in four studies, people acted more selfishly toward men whose face shape is linked to aggression, which drew selfish replies. When tired, ask a question that could prove you wrong.

You can think of perceiving a person as three steps: selecting which cues to notice, organizing them into a pattern, and interpreting what they mean. Motives can tilt the process: one review of fairness research argues people weigh self-flattering cues more, so two perceivers reach different views from the same cues. People rarely ask what they missed. Before concluding, list what you noticed and what you might have missed.

Salience is how strongly something grabs attention, and people tend to treat whatever is most salient as most causally important. A person is often more noticeable than the forces acting on them, which is one proposed reason observers blame the actor rather than the situation. If one loud teammate dominates a failed project in your memory, ask what quieter factors, like deadlines or missing data, also shaped the outcome.

An implicit personality theory is your unspoken sense of which traits come together, used to build a full impression from a little information. If you assume a quiet person is unfriendly, or that a confident speaker is also competent, you are filling gaps with your theory rather than evidence. Gender stereotypes at work can run this way. Mark which parts of an impression you observed and which you inferred.

Perception checking is a repair move for an uncertain read: describe what you noticed, offer two plausible meanings, and ask which fits, as in 'You left early. Were you upset, or did something come up?' The reasoning: research scores a read of someone's thoughts against that person's own report, and accurate reads lean more on words than on looks. Ask before reacting to a glance or a short reply.

An internal attribution places a cause in the person, such as effort, ability or character; an external attribution places it in the situation, such as a hard task, bad luck or pressure from others. Research on everyday explanations shows people often go further and ask which specific goal, reason or mood drove an act. When a friend is late, list one internal and one external explanation before choosing.

The fundamental attribution error, or correspondence bias, is over-crediting personality when explaining others' behavior. In classic studies, readers inferred a writer's true opinion from an essay even when told the side was assigned. One account: the trait explanation comes automatically, while correcting for circumstances takes effort. Cross-cultural work suggests the bias is not universal but is strong in Western samples.

The actor-observer asymmetry is the classic claim that you explain your own actions by circumstances and other people's by personality. A 2006 meta-analysis found many studies failed to replicate it, with average effects near zero. Differences do show up in finer-grained explanations, such as which reasons and which features of a situation get mentioned. Hold the broad rule loosely, and listen to the reasons people give.

The self-serving bias is crediting yourself for success and outside factors for failure, like a cook who praises his skill when a dish works and blames the oven when it burns. It shows up across many groups, though its size varies with age and culture. In one lab task, people tended to blame the environment for their errors even though they could usually tell the real cause. After a win, ask what luck added.

The belief in a just world is the sense that people get what they deserve. It can steady you and support optimism, but research links stronger versions to derogating innocent victims, which protects the belief from evidence that bad things happen to blameless people. When a neighbor loses her job and your first thought is that she must have slacked off, ask what evidence you actually have.

Hostile attribution bias is the tendency to read ambiguous acts, such as a bump in a hallway or an unanswered message, as intentional attacks. Research links it to reactive aggression, meaning aggression in response to a perceived provocation. The cost is that answering an imagined attack with a real one hands the other person a real grievance. Before you respond, name an accidental explanation and check it.

Kelley's covariation model uses three checks: consensus (do other people respond the same way?), distinctiveness (does this person react this way only in this situation?) and consistency (does it happen every time?). Low consensus, low distinctiveness and high consistency point to the person. If everyone struggles with one manager, every time, look at the manager. People often skip these checks and blame the person anyway.

Implicit measures, such as the Implicit Association Test, infer attitudes from how fast people pair concepts, like faces with pleasant or unpleasant words. Average biases on these tests are among psychology's most replicated findings. But one person's score is unstable over weeks, and meta-analyses put the tests' average correlation with behavior at about .14 to .37, small to medium. Treat a score as a snapshot, not a verdict.

The attitude-behavior gap is the frequent mismatch between what people say they favor and what they do. Attitudes predict behavior better when they match it in specificity: liking exercise in general says less about Monday's gym visit than an attitude toward that gym habit itself. Time pressure, the setting and personality also shape whether an attitude shows up in action. To predict an act, ask about that act.

A schema is a stored template for a type of person or situation; a script is a schema for a sequence of events, like ordering at a counter. Scripts let you predict what comes next without effort. The cost is distortion: people tend to recall stories in their usual order and even add expected parts that were missing. If you are sure someone said something, check whether your template supplied it.

Automatic processing is fast, needs little attention and is hard to switch off; controlled processing is slow, effortful and limited in capacity. One influential model holds that deliberate thought steers a judgment only when a person has both the motivation and the opportunity to use it. Time pressure and distraction remove the opportunity. For decisions about people that matter, slow down enough to allow a correction.

Priming is when exposure to one idea makes related ideas easier to reach. In reaction-time tasks, people respond faster to concepts linked in memory to a prior cue, and some such effects have replicated well. Bolder claims fared worse: a famous finding that age-related words made people walk slower failed to replicate with automated timing, and appeared only when experimenters expected it. Be wary of dramatic priming claims.

A thin slice is a brief excerpt of behavior, usually under five minutes, used to judge someone. Accuracy depends on what is judged: extraversion shows quickly, agreeableness is harder to read from a glimpse, and posed emotions are easier than spontaneous ones. Longer exposure helped in some personality studies. Trust a quick read about how outgoing someone is more than one about how kind they are.

Stereotype accuracy research compares beliefs about groups with measured group averages. Results are mixed: many gender stereotypes track real differences, political and national-character stereotypes often miss, and one study found gender stereotypes about game preferences right in direction but overstated in size. An accurate average still describes a group, not a person, so let what this person does outweigh the label.

Theory of mind is the ability to recognize and reason about other people's thoughts, wants and beliefs. Its best-known test is false-belief understanding: knowing that someone who did not see a toy being moved will search where they last saw it, even though you know it is elsewhere. Passing requires setting your own knowledge aside. Adults make the same move when predicting how a colleague who missed an update will act.

On classic spoken false-belief tasks, most children pass somewhere between ages three and five. Language-free versions suggested success as early as 13 months, but those results are contested: a large multi-lab study reported that toddlers did not anticipate actions based on what someone knew. Timing across cultures is also uncertain, since most studies drew on Western samples. Treat any age claim as tied to the task used.

The curse of knowledge is a bias toward your own knowledge when judging what someone less informed knows. In a well-known demonstration, people tapping a song's rhythm overestimated how many listeners would name the tune. Adults show it in false-belief tasks too, though seven large studies found one form of it, a stronger bias when the wrong answer seems plausible, under half its reported size. Test explanations on a newcomer.

Projection is attributing to someone the state you would have in their situation, rather than the state they are actually in. Such egocentric errors are widespread, and simulation theory offers one explanation: imagining another mind means setting your own preferences aside, which takes effort. If you would love a surprise party, that says little about a shy friend. Adjust for their history, tastes and goals, or ask them.

Intent is what someone meant to do; impact is what actually happened. In one set of scenario studies, people blaming someone who harmed a third party leaned on intent, but when they imagined being the one harmed, the outcome counted for more. So 'I didn't mean it' may answer an onlooker but not the person who was hurt. When you cause harm, acknowledge the effect before explaining your intent.

First-order mentalizing is 'I think you think'; second-order adds a layer: 'I think you think she thinks.' Children master it later than simple false belief, and adults can hold only a few layers. Coordination relies on it. One set of studies found people treat publicly witnessed events as known to all and sometimes wrongly extend shared knowledge to private news, a pattern its authors tie to everyday faux pas.

Ways of reading minds vary between people, including between autistic and non-autistic people. The double empathy problem is the idea that understanding between those groups breaks down in both directions, not on one side only. In one study, general-population viewers tracked autistic narrators' emotions less accurately; in another, observers rated mixed autistic and non-autistic pairs least successful. A misread takes two.

Kinesics is communication through body movement. A classic scheme of gesture types includes emblems, gestures with a learned meaning that need no words, like a finger to the lips for quiet; illustrators, which accompany and picture speech, like tracing a shape in the air; and adaptors, such as twisting a ring, which are not meant to communicate. Emblems are culturally learned, so check what a gesture means where you are.

Proxemics, a term coined by Edward Hall, is the study of how people use space. Comfortable distance shifts with who the other person is, the setting and even mood, and it varies by culture: studies report that people in many East Asian samples prefer more room than people in Mediterranean ones. Coming closer than someone likes tends to bring discomfort or withdrawal. Judge closeness against the setting and the person's habits.

Oculesics is the study of eye behavior, such as gaze, mutual gaze and looking away. Norms differ: studies find Japanese participants make less eye contact than Canadian or Australian ones, and averted gaze can signal respect in some Asian cultures. A large review found no single standard for using or reading eye contact. Gaze aversion is also a popular but unreliable lie cue. Compare someone's gaze to their own habits.

Paralanguage is the vocal side of speech: pitch, loudness, rhythm and voice quality. These cues carry emotion apart from the words, and listeners can identify emotions in unfamiliar languages better than chance. They do best in their native language, an in-group advantage. So a flat 'fine' from a colleague is evidence worth noticing, while reading a speaker from another language background calls for extra caution.

Haptics is communication through touch. A study comparing the UK and Japan found that in both, the closer the emotional bond, the more of the body people let someone touch, though UK participants found social touch more pleasant. Context changes meaning: a hand on the arm can comfort coming from a friend and intimidate coming from a stranger. Match touch to the relationship and setting, and let the other person set the limit.

Chronemics is the use and perception of time in interaction: punctuality, waiting and pace. Artifacts are objects people keep around them, like desk items, and appearance includes clothing. All get read as messages: in one study, doctors pictured in white coats were rated warmer and more competent than doctors in casual clothes. Time habits vary by culture, so before reading a late reply as disrespect, consider their day.

Nonverbal behavior has no dictionary: a smile, a close approach or a touch can mean warmth at a reunion and pressure in a negotiation. One review lists the idea of a readable body language among four misconceptions. Read several cues together, against the setting and the person's usual style, but even clusters can mislead, so lean on what they say: accurate reads of thoughts and feelings rely more on words than on body cues.

Across a meta-analysis of more than 25,000 judgments of whether people were lying, average accuracy was about 54 percent, barely above the 50 percent of guessing. Groups you might expect to excel, such as police investigators and psychiatrists, did no better than lay people, and studies of real suspects look similar. Treat a gut sense that someone is lying as a weak hint, not a finding.

Many people, police officers included, believe liars avert their gaze, fidget and shift their posture. Reviews of the research find nonverbal cues to deceit faint and unreliable, and liars generally do not show more of the nervous behaviors observers watch for. Leaning on such cues can make judgments worse. If a friend avoids eye contact while explaining a missed call, that alone tells you almost nothing.

Lie-detection research separates the judge's ability from the sender's detectability, meaning how easy a given speaker is to read. In meta-analytic data, judges showed little reliable difference in skill, while some senders were consistently easy to read and others consistently hard. A 2018 study found the same when judges did not understand the speakers' language. So a catch says more about the sender than about the judge.

Some interviews add mental load, on the disputed idea that lying takes more effort than truth telling. Liars rehearse answers to expected questions, so unexpected ones can trip them. A meta-analysis of 25 studies of a method family that also prompts people to say more found a moderate gain. But results vary: reverse-order retelling helped in one study and made honest people describing plans look like liars in another.

Two well-studied interview methods use checkable facts. The verifiability approach asks people to include details someone could check, such as named witnesses or receipts, and truth tellers supply more of them. The strategic use of evidence holds back known facts and reveals them gradually, so a liar's story drifts from it. Both have stronger support than reading body language, though most of it comes from a few labs.

People tend to overestimate their skill at spotting lies, and body-language training does little to justify it: a meta-analysis found only a small effect, against a medium one for verbal-cue training. Observers taught the behavioral cues in a police interrogation manual did worse than untrained ones, and in one study micro-expression training left people at chance on real high-stakes lies. Weigh evidence over impressions.

Truth bias is the tendency to judge messages as honest regardless of whether they are. Truth-default theory argues this pays off: since the bulk of everyday talk is truthful, believing is usually right. The cost: lies often pass as truth. The bias changes the kind of mistake you make, missing lies more than wrongly accusing, without making you more accurate overall. Keep the default, but check claims when the stakes are high.

Active listening is a set of moves, not one skill. A paraphrase restates a point to check you got it; a summary ties several points together; an open question ('What happened next?') invites more; reflecting a feeling names the emotion you heard. When Dana vents about a missed promotion, 'So the review felt unfair' shows more than a nod. Done by rote, the moves can still leave someone unheard, so mean them.

Listening is a sequence: you take a message in, understand and remember it, weigh it, then respond. Research scales measure sensing, processing and responding as related but distinct parts. Omar hears every word of a project handoff and nods warmly, but by lunch he has forgotten the deadline: he took it in and replied, yet lost it in between. When talk goes wrong, ask which stage slipped before deciding nobody listened.

A listening style is a habitual focus. People-oriented listeners track feelings, action-oriented ones want a clear, organized point, content-oriented ones weigh details before judging, and time-oriented ones want it brief. Most people mix styles and shift with the setting. When Lee describes a hard week and Sam keeps asking for the bottom line, Lee may feel rushed. Notice what a speaker seems to need, and adjust.

Understanding speech takes attention, and so does planning what you will say, so the two compete; lab studies find that hearing speech while preparing your own makes the planning slower and more error-prone. Listening trainers also advise putting phones away and holding off judgment. When you catch yourself drafting a comeback while someone is mid-sentence, set it aside and finish taking in what they are saying.

Empathic listening tries to grasp the speaker's view and feelings while holding judgment back; evaluative listening, which one listening scale calls analytical, weighs facts and details before judging. Both have uses. When Priya vents about a rude client and her coworker starts auditing her account, she feels judged rather than heard. When a friend asks you to vet a lease, sympathy alone fails her. Ask which kind is wanted.

Backchannels are brief signals like 'mm-hm' or 'yeah' that show you are following without taking a turn; researchers also call them continuers. They are frequent and nudge the speaker to go on. But they can be faked, and generic ones can be given without taking in what was said. If Ana hears only 'mhm' while she explains a problem, she may doubt you followed. Mix in a specific response that proves you did.

High-quality listening means attending, trying to understand and not judging; it does not mean agreeing. In experiments by one research group, speakers who received it were less defensive and held less extreme, more two-sided views. If Marco objects to a budget cut, hearing 'So the cut hits your team hardest' before any rebuttal may make him less defensive. In a disagreement, reflect their point first, then state yours.

A transition-relevance place is a point where a turn could end and another speaker may start. Speakers mark these with grammar, intonation, breathing and gaze; in lab studies, listeners use grammar and intonation to anticipate turn ends. Gaps between turns average about 200 milliseconds across languages. When Ben pauses after 'eggs, milk, and,' nobody jumps in. If you keep being cut off, ask whether your pauses sound finished.

Phatic communion is a linguistics term for talk meant to make contact, not pass on facts, like a weather remark. One account holds that a neutral, self-focused or other-focused remark can signal status. Workplace researchers credit small talk with building belonging and warmth, though one employee survey also tied it to less focus on work. When a coworker says 'Busy morning?', answer the warmth, not just the facts.

An adjacency pair is two linked turns: a first part such as a question, offer or invitation, and the reply it makes relevant. Because the reply is expected, its timing carries meaning. Refusals of offers and invitations tend to arrive late, so a pause can signal a no before it is said. When Kim asks 'Free Saturday?' and hears a long 'Ummm,' she may soften the request on the spot. Listen to the gap, not only the words.

Repair is how talk handles problems of hearing or understanding. A study of 12 languages found someone asking for clarification about once every 1.4 minutes, with 'Huh?' when the trouble is vague and a pointed 'Which one?' when it is specific. Speakers most often fix their own slips, and listeners tend to wait before flagging one. When a friend misspeaks, a short pause or a puzzled look gives them room to self-correct.

Topic management is how people open, keep and change subjects. In one study, adults rating children's conversation transcripts judged topic shifts most natural when the new topic related to the old one, was marked with a phrase like 'by the way,' and came once the earlier topic closed. Unrelated jumps mostly rated lowest. If Rosa describes her trip and Theo jumps to his car trouble, she may feel dropped. Close, then bridge.

Open questions ('What happened?') let people answer in their own words; closed ones ('Was it Monday?') draw short, specific replies. Leading and option-posing questions suggest an answer, and a presupposing question takes a detail for granted, as in asking how big the dent was before anyone mentioned a dent. Interview research links such questions to wrong details in what people report. Start open, then narrow.

Overlap is not always rude. In one German conversation corpus, about a quarter of speech segments came from stretches of talking at once, and the early starter was often responding to what the other had just said. Silence varies too: research describes it as respect or careful thought in some cultural settings and as indifference or disagreement in others. If a colleague goes quiet after your proposal, ask what it means.

Descriptive norms describe what people commonly do; injunctive norms describe what others think you ought to do. In one economic game, when players learned that most earlier players acted selfishly but thought generosity was right, only the report of what people did changed behavior. A manager who grumbles that nobody files reports on time may advertise the wrong norm. Show both approval and compliance.

A role is what others expect from whoever holds a position, such as coach or parent. Inter-role conflict pits two roles against each other, like a school pickup against a late meeting. Intra-role conflict comes from clashing demands inside one role, as when a supervisor's boss wants speed and her staff want breaks. Role strain is the stress of juggling such demands. Name whose expectations collide before blaming yourself.

A script is a stored pattern for how a type of situation unfolds, learned from experience. Cues in the setting tell you which script applies and where you are in it. At a job interview, greetings, questions about your past, your own questions and a handshake mark the stages. A guest at a potluck who arrives with no dish has the wrong script. In an unfamiliar kind of event, watch which step comes next before acting.

Pluralistic ignorance is a shared misreading of opinion: most members privately reject a view but think most others hold it, so they go along publicly. In one study, even teachers underestimated how accepting colleagues were of tattooed teachers. Correcting the belief may not change action: across 11 countries, learning the real climate consensus barely moved willingness to act. Voicing your private view gives others evidence.

Many conventions begin arbitrarily. One philosophical account says a group lands on a pattern by chance, agreement or salience, and each repeat makes it more expected, until people follow it because others do. Another stresses that conventions spread by weight of precedent, not merit. If an office's first hires ate lunch at their desks, newcomers may copy them for years. Ask whether a rule serves a purpose or only a history.

Norms cover actions people can control, so first ask whether complying was possible. A newcomer may not know the rule; an injured guest may be unable to stand for a toast. Deliberate breaks send signals: in one team's studies, rule breakers seemed more dominant, and those breaking a wider rule to keep their group's norm also seemed prestigious. Explain to the unaware, accommodate the unable, engage the deliberate breaker.

Newcomers can learn a group's norms by watching members and updating their guesses as observations pile up. Everyday norms vary by situation and society, so what passes in a park can jar in an office, and missing them can send signals you never meant. When Wen joins a new team, she notes whether people interrupt, how meetings open and who speaks first before trying a joke. Watch a few rounds, then match.

Informational influence comes from wanting to be right: other people's answers count as data. It grows when a task is hard or ambiguous, and it can be sensible, since others may know more. Normative influence, by contrast, comes from wanting to be accepted. When Ravi, new in town, picks the busy noodle shop over the empty one, he is using the crowd as information. Ask whether the crowd actually knows more than you.

Normative influence comes from wanting to fit in and avoid disapproval. It often produces public compliance: outward agreement while private views stay put. Conformity studies repeatedly find more going along when answers are given publicly than when they are private. When Jo backs the group's restaurant pick aloud but votes for another on the anonymous poll, that gap is the sign. Private ballots expose it.

Asch had a person say which line matched a standard after planted confederates all gave the same wrong answer. On those trials about a third of answers went along; a 2023 replication found 33%, and a review of 44 replications found about 25%, so most answers stayed independent. Asch also found one confederate breaking ranks cut conformity from about a third to 5%. If you doubt a group, say so; you may free others.

Milgram had volunteers give supposed shocks of rising strength to a learner. In some versions about two-thirds went to the maximum, but across his many variants rates ran from near zero to near total. One reanalysis found completion far lower when stopping did not mean confronting the experimenter, and people often refused a blunt order. The experimenter also added arguments beyond his scripted prods. Read it as conditional.

Moscovici argued a minority holding its position consistently makes the majority rethink, and some may privately convert first; later work added it must also seem open, not rigid. Reviews find this works only in narrow conditions: persuasion studies see it mostly in private or on related issues, while discussion groups often show open shifts. As a lone dissenter, be consistent, give reasons, and show you heard others.

In a mock jail, students randomly cast as guards abused those cast as prisoners, and the study ended after six days. It was long read as roles taking people over. A reanalysis notes the lead researcher briefed guards on making prisoners feel powerless and that only some guards turned brutal; in the reanalysts' own later prison study, people did not slip automatically into assigned roles. Ask what those in charge encouraged.

Kelman split conformity three ways: compliance is agreeing in public without accepting in private, identification is agreeing provisionally because it fits your role in a group, and internalization is accepting a view as your own. Compliance tends to fade out of sight: conformity runs lower when answers are private. When a new hire repeats the boss's slogan only in meetings, expect it to vanish off the record.

Entitativity is how much a set of people is seen as a real unit. Studies tie it to interaction, common goals, shared outcomes, similarity and how much membership matters, but not to group size or how long the group has existed. Strangers on one flight are a collection; if a delay pushes them to split a rental van, they gain shared goals and outcomes. To tell whether you face a group, look for interaction and a common stake.

Cohesion is the pull that holds a group together: task cohesion is shared commitment to the goal, social cohesion is liking and closeness. It is generally linked to better team results. But students in one interview study described close-knit teams that avoided negative feedback and stalled, and an experiment finding no effect of cohesion used a task likely too easy to show one. If a close team never argues, invite a critic.

A classic scheme by Benne and Sheats sorts group roles into three kinds. Task roles move the goal along: proposing ideas, seeking information, summarizing. Building and maintenance roles keep people working together: encouraging, mediating, making room for quiet members. Individual roles serve personal needs, such as blocking or seeking recognition. If a meeting stalls or turns sour, check which kind of role nobody is filling.

In a centralized network, messages pass through one or a few members; in a decentralized one, links spread evenly. Early research found decentralized teams better on complex tasks; some later studies agree, others found the opposite. In one lab study, groups that picked their own hub, often the most active talker, did as well as or better than groups given a random one. Match structure and hub to the work.

Socialization is how a newcomer learns a group's values, role expectations and relationships. Recent work treats it as two-sided: groups offer guidance and support, and in one survey, newcomers who took more initiative reported better adjustment on most measures. Newcomers are a minority, yet they can shift how members act. A new hire who asks why a checklist gets skipped may change the habit. Ask, watch, then contribute.

Tuckman described groups moving through forming, storming, norming and performing, with a final adjourning stage when the work ends. The sequence is a useful map, not a timetable. An interview study of medical students' learning groups found it fit groups new to the task, while experienced groups could work well almost at once, then plateau or change abruptly. If your team skips the conflict, watch for complacency instead.

Bigger groups bring more resources, yet each added member raises the need for communication and coordination and cuts the airtime left for everyone else. In brainstorming, ideas per person fall as groups grow, and pairs do best per head. In a simulated cardiac-arrest response, by contrast, teams of up to five finished faster because tasks ran in parallel. Size up when work splits cleanly; stay small when everyone must talk.

Social facilitation is the change in performance when others watch or work alongside you. Zajonc proposed that an audience strengthens your most practiced response, helping on easy tasks and hurting on tricky ones; a rival account says attention narrows. A large replication of the first classic study found the effect, though a repeat of one of Zajonc's animal studies did not fully match. Rehearse before performing for a crowd.

Social loafing is lower individual effort on a shared task than on the same task done alone. Work by Karau and Williams tied it mainly to low identifiability and little chance of evaluation, and found it stronger when the task seems low in value, a teammate performs well, or inputs are redundant. On a group slide deck where nobody owns a slide, effort sags. Give each person a visible piece, and make the task feel worth doing.

Steiner called the gap between what a group could produce and what it actually produces process loss. Losses come from coordination, like duplicated work or waiting on each other, and from motivation, like reduced effort. Process gain is the reverse: interaction yields more than members could manage alone, sometimes topping the best individual. Before adding people, ask whether the handoffs will cost more than the help.

Brainstorming aloud runs into production blocking, where people wait their turn and lose ideas, plus fear of judgment and a drift toward the group's slower pace. Groups of four tend to produce about half the ideas of four people working separately. Writing ideas first or sharing them electronically recovers much of the loss. Before a team idea session, give everyone a few quiet minutes to list ideas alone.

Group work rewards everyone for the group's result, which tempts some to contribute little: free riding. The sucker effect is the backlash. In a classic experiment, people cut their effort when a partner who seemed able was slacking, to avoid being exploited, though a later study of four-person teams found no such pattern. When a teammate coasts, raise it directly and record contributions so hard workers do not quietly quit.

Steiner's task types describe how members' work adds up to a group result. Additive tasks sum everyone's effort, like a cleanup crew; compensatory tasks average inputs, like pooled estimates. Conjunctive tasks go only as well as the weakest member, like an assembly line, while disjunctive tasks need one person to succeed, like spotting the single typo in a contract. Before staffing a task, ask whose level will set the outcome.

Woolley and colleagues found small groups doing varied tasks showed a shared performance factor, weakly tied to members' average test scores and linked to equal speaking turns and skill at reading others. A meta-analysis of their own studies backed it, but independent work is mixed: one study found two factors and no turn-taking link, another that individual ability explained most differences. Equal airtime still costs little.

Groupthink, a term Irving Janis popularized through a case study of the Bay of Pigs invasion, describes a tight-knit group so intent on agreeing that it stops testing its choices honestly. Warning signs include weighing only a few options, never rechecking the favored plan's risks, and skipping outside experts. Because the idea grew from case studies, use the signs as a checklist, not a verdict.

Group polarization means a group's view after talking is more extreme than its members' average view before. If most of a hiring panel arrive mildly keen on a risky candidate, discussion tends to push the panel toward a bolder yes; a cautious majority grows more cautious. The pattern has been seen in juries and in panels of judges. Collect private views before the meeting so you can see how far the group moved.

Escalation of commitment is staying with an earlier choice, and putting more money or effort into it, as evidence against it grows. A team eight months into a custom app may fund another quarter to avoid writing off the first eight. A preregistered replication confirmed that people escalate more for choices they made themselves, but found no support for regret as a driver. Judge the next dollar by what it buys from here.

A shared mental model is a team's common understanding of the task, how the team is organized and who does what. It shows up as implicit coordination, like a nurse passing a scan result to the surgeon before being asked. In a study of 27 simulated trauma teams, behaviors signaling a shared model predicted effective care better than teamwork skills did. A quick test: could each member name the next step and its owner?

A decision rule is how a group turns many opinions into one choice. A simple majority feels fair when preferences differ, but it is only as good as the information behind each vote. Weighting votes by reliability can do better, though cues like rank and confident delivery are rough markers of it. The Delphi method gathers private views, shares an anonymous summary and repeats until the group agrees or hits a preset stop.

In a hidden-profile task, the best choice shows up only when members pool facts that just one or two of them hold. Talk drifts to common knowledge instead: shared facts come up more because more people hold them, and echoing what others know wins approval. A team picking a vendor may rehash prices everyone saw and never hear the engineer who knows the product fails in cold weather. Ask each member early what only they know.

Appointing a devil's advocate, someone told to argue against the emerging view, is one way to test a forming consensus. But positions argued on assignment tend to sway listeners less than sincere ones, possibly because they are voiced with less confidence. Procedures can help too: a no-interruption rule, equal speaking time and anonymous written rounds. If you name a critic, also invite anyone who privately doubts the plan.

A designated leader holds the title; an emergent leader is the one the group starts treating as leader. In one small study, 16 staff of one organization did a decision task in four rounds of new virtual groups. Formal rank shaped who was seen as leading, but communication style mattered more: those who spoke more, and earlier, tended to be named. In a new group, notice who speaks first and most, then ask if they know the most.

A directive leader gives clear instructions and expects them followed. A participative leader brings members into the decision but usually makes the final call. An empowering leader hands authority and responsibility to members. A review links participative leadership to commitment and to employees speaking up, but notes that leaning on it heavily can add stress. Ask who holds the knowledge and what load people can carry.

Psychological safety describes a team where people believe raising concerns, questions and mistakes is expected and possible. A systematic review of nine healthcare studies found its link to patient safety outcomes equivocal. One reason the authors note: a team that feels safe may report more incidents, so a higher count can reflect candor, not danger. If reports rise after a team opens up, ask if people are just saying more.

NASA's systems engineering handbook orders decisions this way: understand the decision needed, define criteria for judging options, then identify alternatives and evaluate them against those criteria. Jumping straight to options is risky, since free discussion can narrow fast onto a few early ideas. A team that picks software before agreeing what it must do may end up debating features. Write the problem and criteria first.

Meeting design covers choices made before and during a meeting: whether to hold it, who attends, what the agenda covers and how decisions are captured. In a survey of 190 hospital nurses and midwives, staff rating meetings effective also rated team performance in them high, though cause was not shown; workplace culture was the factor most tied to both. The authors advise a clear purpose, a set agenda and an owned action log.

Media naturalness theory holds that the further a medium drifts from face-to-face talk, losing shared space, real-time exchange, facial expression and body language, the more mental effort it takes. Video calls lose some cues; a cameras-off call loses more. In a study of 549 online comments on remote meetings, most frustration centered on cameras and microphones. The authors urge agendas and ground rules, like hand-raising.

Speaking time in groups tends to track things unrelated to the task, such as power and gender, and research cited in meeting studies links domination by high-status members to lower participation by others. Procedures can spread the floor: give everyone a fixed turn, enforce a no-interruption rule, or collect written ideas before open talk, so quieter members' views are already on the table.

French and Raven named five bases of power. Legitimate power is a role's formal authority; reward and coercive power are the perceived ability to grant benefits or impose penalties. These three come with a position. Expert power rests on knowledge and skill, referent power on being admired and identified with; these belong to the person. When a manager retires, the first three end with the job; the last two can outlast it.

Social psychology usually defines power as asymmetric control over valued resources, and status as rank in others' eyes, the standing that leads people to defer. A 70-culture study of over 18,000 people, summarized in a review of power research, found the power-to-status link varies widely: in some societies the two nearly coincide, in others they barely relate. Power seen as serving the common good earned more status.

The dual model of status describes two routes to rank. Dominance relies on fear, intimidation and control of resources; prestige comes from skill, knowledge and respect, and is granted by those who defer. In one cited group study, both routes predicted who gained influence. Dominance-motivated leaders are described as putting their own needs first. Some researchers argue the two are defined too loosely to separate cleanly.

Differences in who holds the floor show up early. When researchers put three strangers together for a 10-minute casual chat, people differed in how much of the talk they took, how often they started topics and how often they interrupted. Other studies find that speaking early and often in a new group predicts being seen as its leader. In a new group's first minutes, notice who claims the floor; others are reading those cues.

Experiments that make people feel powerful, for example by having them recall a time they held power, find that power can reduce attention to others' perspectives and dampen empathy, under some conditions. One found lower momentary empathy after a power recall. But cross-cultural work finds that people who see their power as entrusted by others take others' views more. When you gain authority, ask whose trust it rests on.

Power-dependence theory holds that your influence over someone rests on how much they rely on resources you control, and on whether they have alternatives. A lead who alone approves purchases has leverage; a self-service tool for small ones shrinks it. In one study, employees who relied more on AI tools felt more powerful with their leaders. To gauge leverage, ask how badly something is needed and where else it can be had.

A broker is someone who bridges people or groups that are otherwise unconnected and helps them interact. A study of school staff measured brokerage three ways: formal title, position in the real network of contacts, and actual bridge-building behavior. The three were related but not interchangeable, so a title alone did not reveal who linked departments. To find real connectors, ask whom people go to for help across teams.

Impression management is any attempt to steer how others see you. A 2017 meta-analysis split it into self-focused tactics, like touting wins, and other-focused ones, like flattering the rater or stressing what you share. Self-focused tactics related to interview ratings, but not reliably to job performance ratings in the few studies; other-focused ones did. The authors suggest why: supervisors already know your record.

Organizational citizenship behavior is discretionary work, like covering a colleague's shift or training a newcomer, that the formal reward system does not directly recognize. A survey of 299 employee-supervisor pairs in South Korea linked more of it to higher overall evaluations. Demanded rather than chosen, it becomes compulsory citizenship, which a meta-analysis found strongly tied to burnout and work-family conflict.

Employee silence is withholding concerns that could improve things. Defensive silence comes from fear of what speaking up will cost; acquiescent silence, from having disengaged. A healthcare review notes that speaking-up programs often disappoint when the culture around them is unsupportive. A suggestion box nobody answers invites disengagement; a usable channel reports back what happened to each concern.

Edgar Schein's model sees an organization's culture at three levels: visible artifacts like rituals and slogans; espoused values, its stated principles; and basic assumptions, beliefs so taken for granted they rarely get voiced. Schein argued the assumptions drive behavior. A clinic can print patients-first on badges while booking around staff convenience. When stated values and daily practice disagree, read the practice.

Organizational socialization is how a newcomer adapts to a new role, marked by role clarity, task mastery and social acceptance. A review found only five usable studies of formal onboarding, with low-certainty evidence; structured on-the-job training had the best support, for role clarity and competence. A survey of 419 new tech employees tied self-starting behavior to clearer roles. Early on, ask what good work looks like.

Political skill combines four abilities: social astuteness, reading others accurately; interpersonal influence, adapting your approach; networking ability, building useful ties; and apparent sincerity, seeming genuine. In a one-time survey of 200 employee-supervisor pairs, politically skilled staff had better ties with supervisors, linked to promotability, mainly when their work was strong. Pair people skills with good work.

Common rating errors include halo, letting one strong area color every score; leniency, rating everyone high; central tendency, parking everyone mid-scale; and similar-to-me, favoring people like yourself. In one study, raters taught these errors learned to name them, but raters trained on a shared definition of good performance rated more accurately. Tie each score to specific observed behavior, not an overall feeling.

Team conflict comes in three kinds: task conflict over the content and goals of the work, relationship conflict over personal friction, and process conflict over who does what and how. Relationship conflict is generally found harmful. Task conflict was once seen as useful, but two meta-analyses found no overall positive link to team outcomes; intensity seems to matter. Keep disputes on the work and cool them early.

The dual concern model crosses concern for your own needs with concern for the other party's. High on both gives integrating, a solution that fully serves both; high self, low other gives dominating; low self, high other gives obliging; low on both gives avoiding; middling on both gives compromising. Integrating and compromising tend to go with better outcomes. Before a hard talk, ask how much each concern matters here.

Models describe conflict climbing in stages unless someone steps in. One has tactics harden from requests to demands, anger and threats; another adds recruiting supporters. Glasl's model groups nine stages into three phases: disputes over facts, clashes that turn personal, and efforts to harm the other. Little tested, it got support in one survey-based validation. Name the stage early; the remedy changes as it climbs.

Motive attribution asymmetry is the tendency of groups in conflict to explain their own side's acts by love of their group and the other side's by hatred. Both sides do it, so each sees itself as defending and the other as attacking. A study of US Republicans and Democrats replicated the pattern; feeling threatened went with harsher views of the other party. Before reacting, ask what the other side thinks it is protecting.

A position is the specific thing a party says it wants; an interest is the need behind it. Interest-based problem-solving, often used in labor-management bargaining instead of trading positions, keeps personal friction apart from the issue, maps each side's interests, generates options and settles on a standard both accept. Two coworkers fighting over a quiet desk may both need focus time, which a booking schedule could give.

Third parties differ in what they take from the disputants. A facilitator guides the process and stays mostly out of the substance. A mediator, an impartial helper with no power to decide, works with the parties, sometimes separately, toward a deal they accept. An arbitrator hears both sides and issues a decision that may be binding. Parties may hold back in mediation if the same person will later rule, so settle roles first.

Trust rests on judgments of another person's ability, benevolence and integrity, so repair starts by naming which one was harmed. In one preregistered experiment, a person who apologized for a past breach was trusted more than one who denied it, and the gap held after good news arrived. Other experiments found the best form of apology depended on whether competence or integrity was at issue. Settle the issue, then mend trust.

Distributive bargaining splits a fixed amount, so one side's gain is the other's loss. Integrative negotiation creates value, for example by trading concessions on issues the parties rank differently, or by adding issues. Many people approach negotiations as zero-sum. A quick test: is there more than one issue, and might the sides care about them unequally? If so, there is likely value to create before anyone claims it.

Anchoring is the pull a first number exerts on the judgments that follow. In negotiation the first offer acts as an anchor: counteroffers and final prices tend to settle near it, a finding researchers treat as robust. Moving first has costs too when you know little about the other side. If a contractor opens with a high quote, do not bargain only from that number; restate your own research on comparable jobs.

The best alternative to a negotiated agreement, or BATNA, names what you will actually do if talks end with no deal. It sets your reservation point, the least favorable terms worth accepting, since anything worse loses to walking away. Game theory makes a similar point: bargaining outcomes depend on each side's fallback, and a stronger fallback is a basic source of power. Before asking for a raise, line up an outside offer or a plan B.

The zone of possible agreement, or ZOPA, is the set of deals that each side prefers to having no deal. In a single-price talk it lies between the seller's lowest acceptable price and the buyer's highest. If a buyer will pay at most 200 dollars and the seller needs at least 240, there is no ZOPA on price. Then walk away, or add terms like delivery dates or extras and see whether some package beats both fallbacks.

Logrolling means conceding on issues you value less in exchange for gains on issues you value more, which works when the parties rank issues differently. A caterer may care most about a large deposit and a couple about changing the menu late; trading late changes for a bigger deposit can leave both better off than splitting each issue. It requires knowing the other side's priorities, so ask early which terms matter most.

The fixed-pie bias is the assumption that the parties' interests are directly opposed, so the only question is how to split a set amount. Researchers describe it as common, and it hides trades that cost one side little and help the other a lot. Two roommates arguing over the bigger bedroom may find one cares about space and the other about paying less rent. Ask what each side wants, and why, before assuming opposition.

Game theory separates one-shot encounters from repeated ones. In a repeated game, a player who defects can be punished in later rounds and a cooperator rewarded, so cooperation's value depends on the odds of meeting again, called the shadow of the future. A tenant renegotiating with a landlord they will keep for years is in a repeated game. Before pressing for the last concession, ask how often you will face this person again.

Following Allport, researchers treat prejudice as having two parts: an overgeneralized belief about a group, the stereotype, and a negative feeling toward it. Discrimination is unfair treatment in action. The links are loose: a prejudiced person may never act on it, and someone without prejudice may still discriminate by following a biased rule or norm. Ask separately what people believe, what they feel and what they do.

Stereotype threat is the concern of being judged by, or confirming, a negative stereotype about your group. Lab studies have shown it lowering test scores, but recent meta-analyses report widely varying effect sizes, and critics doubt it appears in high-stakes settings like hiring. A workplace meta-analysis tied it to lower job satisfaction, mostly in cross-sectional data. Treat it as a plausible stressor of disputed size.

A self-fulfilling prophecy is an expectation that helps make itself true: what you expect of someone shapes how you treat them and how they perform. Rosenthal proposed that teachers who expect more of a student offer a warmer climate and teach them more. In a study of the negative version, the Golem effect, with 101 leaders and 202 followers, a group's dim view of the typical follower went with lower effort and performance.

Researchers contrast blatant racism, open hostility now less socially acceptable, with subtle racism in everyday routines where the motive is unclear. Aversive racism theory holds that racial bias now tends to appear in ambiguous situations where a reason unrelated to race is available. A manager who always finds a fit problem with one group's candidates is the kind of case it describes. Look at patterns across many decisions.

Social categorization, sorting people into groups, is a low-effort default that lets you apply what you think you know about a category. A side effect is the outgroup homogeneity effect: people see less variety among other groups' members than among their own. A rival department may seem full of one type of person, while your own team looks like a mix of individuals. When someone surprises you, think about the individual.

The stereotype content model holds that groups are judged on warmth, tied to whether they seem cooperative or competitive, and competence, tied to perceived status. The BIAS map adds emotion and behavior: warm, competent groups draw admiration and help; cold, competent ones envy; warm but incompetent ones pity; cold, incompetent ones contempt. Newer work suggests spontaneous stereotypes span more than two dimensions.

Antibias training often tries to change automatic associations or raise awareness. Meta-analyses find lab manipulations shift those associations only for hours or days, and one review argues that awareness alone does not change a culture. Changing the decision may hold more promise: blind auditions and reviewer instructions stressing idea quality narrowed gender gaps. Before hiring, fix the criteria and hide the irrelevant.

Social identity theory, from Tajfel and Turner, holds that people draw part of their self-concept from the groups they belong to and want those groups to compare well with rivals. A group win then feels like a personal win. Picture an employee who says 'we' after a record quarter but 'they' after a scandal. When you grade your own side generously, ask whether evidence or identity is doing the judging.

In the minimal group paradigm, researchers sort strangers by something trivial, such as which of two shapes they prefer, then let them rate or reward each other. With no shared history between the groups, people still tend to favor their own label, both in how they judge and in how much they cooperate. The bias is smaller than between real groups, but it shows an us-them line needs no real difference to tilt judgments.

Realistic conflict theory says that when two groups want the same limited thing, like jobs, their interests clash and hostility follows. In Sherif's Robbers Cave study, two groups of boys at camp fell into conflict, and cooperation came once staff set goals neither group could reach alone. Evidence is mixed, though: prejudice toward minorities also appears in boom times. When teams fight over one budget, seek a shared goal.

The contact hypothesis predicts that frequent, positive contact between members of two groups reduces prejudice. Allport named helpful conditions: equal status in the encounter, shared goals, working together, and institutional support. A later meta-analysis suggested those are not necessary for contact to help, while negative encounters worsen attitudes. If you bring rival teams together, protect the quality of each meeting.

Scapegoating is unfairly blaming an out-group for a group's misfortunes. The frustration-aggression account says that when people cannot strike at the true source of their stress, they may redirect anger at an innocent target. Critics noted this cannot explain who gets picked; later theory points to groups that are weak, already excluded and easy to spot. When a layoff gets blamed on the newest hires, ask who decided it.

To dehumanize people is to see them as less than human: animal-like, lacking civility and reason, or object-like, lacking warmth and emotion. Many researchers argue it weakens inhibitions against harm, and lab studies link it to support for harsher treatment. Critics reply that much cruelty targets people seen as fully human, such as supposed traitors. Treat 'they are animals' talk as a warning sign, not the whole story.

Everyone belongs to many groups at once: a nationality, a job, a faith, a team. Social identity complexity is how distinct those memberships seem, and seeing them as distinct is linked to more tolerance of out-groups. In one experiment, showing several identities of the people in an unfair exchange reduced harsh treatment of the out-group member. When one label is all you see in a rival, look for their other groups.

The bystander effect is the finding that people are less likely, and slower, to help when others are present. It is among the more reliable results in social psychology, though a meta-analysis found it weaker in clearly dangerous emergencies. Its origin story, 38 neighbors ignoring the 1964 murder of Kitty Genovese, was later discredited; witnesses did report the attack. In a crowd, assume nobody has acted until you see it.

Diffusion of responsibility is the thinning of personal duty when it is shared among many possible helpers. Each assumes someone else will step in, or asks why it should be them. It can show up far from emergencies: in one email experiment, a plea for help visibly sent to many others drew fewer replies, though not when the message was framed with cues about helping. The logic suggests a remedy to try: ask one named person.

Psychological egoism says all helping is self-serving: easing your own distress, dodging blame or earning praise. Batson tested each selfish motive in turn. Highly empathic people still helped when they could easily leave, when no one would know, and when skipping would cost little guilt. Philosophers reviewing it call this bad news for egoism, though debate continues. To test a motive, remove its payoff and watch.

Latane and Darley described emergency helping as a chain: notice the event, read it as a real emergency, take responsibility, decide how to help, then do it. Each link can break. A commuter staring at a phone may never notice; a calm crowd can make a collapse look harmless; a trained first-aider may assume the nurse nearby will act. When you hesitate, find the link you are stuck on and deal with that one.

Audience inhibition is the brake onlookers put on helping: you fear being judged if you step in and turn out to be wrong. You might misread the scene, since a man slumped on a bench may just be asleep, or be blamed, even sued, if your help goes badly. With diffusion of responsibility and pluralistic ignorance, it helps explain why crowds help less. Weigh it: a false alarm mostly costs embarrassment; delay can cost a life.

Two biological accounts explain how helping can evolve. Kin selection, formalized by Hamilton, predicts more help toward closer relatives, who likely share the helper's genes. Reciprocal altruism, from Trivers, predicts help among non-relatives who meet repeatedly and can spot cheaters. Neither speaks to conscious motive; a trait favored by selection can still feel like real care. Use them to predict who gets helped, not why.

Some bystander programs teach a menu of options, such as the 5D model: distract, delegate, document, delay and direct action. Teens in one workshop study saw risks in some, such as filming an incident or distracting a harasser. A review of youth sexual-violence prevention found readiness to step in did not reliably become action, and effects sometimes faded. Pick the lowest-risk tactic that still reaches the person.

In a social dilemma, acting selfishly tempts each person, yet if everyone gives in, all end up worse off than if all had cooperated. The prisoner's dilemma is the two-person case, where defecting pays whatever the other does; funding shared services and using shared resources are group cases. Picture roommates who each skip cleaning to save time and all live in a mess. Ask: is my best move bad for us if everyone copies it?

In the prisoner's dilemma, each player gains by defecting whatever the other does, yet mutual defection leaves both worse off than mutual cooperation. In repeated play with a known final round, backward reasoning says always defect, though real players rarely do. If the end is uncertain and more rounds are likely, future payoffs can make cooperating pay. The more future dealings you expect, the more cooperation pays.

A public goods dilemma asks each person to pay into something all enjoy, such as a team's shared documentation; since you benefit either way, you are tempted to free ride. A commons dilemma runs the other way: a shared resource, like a town's groundwater, gets used up because each user keeps the gain while the damage is spread over all. Ask which way the temptation runs: taking too much, or giving too little.

Tit for tat cooperates on the first move, then copies the partner's last move. It won both of Axelrod's repeated prisoner's dilemma tournaments. One weakness is noise: after one mistake, two tit-for-tat players can trade revenge for many rounds. Variants that forgive some defections do better when occasional errors happen, and no strategy wins everywhere. If a colleague slights you once, answer once, then cooperate again.

Direct reciprocity is 'you helped me, so I help you.' Indirect reciprocity is 'you helped someone, so I help you': cooperation steered by visible reputation. Models show it can work even where the same two people rarely meet again. Much reputation travels by gossip, which can be false; one experiment found that cooperators also passed on more honest gossip. Before acting on a reputation, ask where the story came from.

In public goods games, paying to fine low contributors often raises cooperation, at a price. In many studies punishing cost more than it gained, some players punished generous contributors, and when targets could hit back, punishment and cooperation fell. In one study with stable identities, adding messages to sanctions raised cooperation despite counter-punishment. When you sanction a slacker, say what you expect and why.

To trust is to rely on someone while accepting that they could let you down. In repeated trust games, where one player invests and the other decides how much to return, a meta-analysis found the partner's actual return rate was by far the strongest driver of trust over time. One experiment found a bad reputation kept trust low even after good behavior. To be trusted, make your reliability visible early and often.

Individualism-collectivism, one of Hofstede's dimensions, describes the bond between person and group: how interdependent a society keeps its members. Direct, explicit speech is more common in individualistic cultures, and in one study of judicial scenarios, Chinese participants, higher in collectivism, assigned more punishment and compensation than Americans. Read it as a society's tendency, not a forecast for one person.

Power distance, in Hofstede's terms, is how far the less powerful members of a society expect and accept unequal power. In interviews, a Chinese participant described keeping silent rather than questioning a superior, even when disagreeing, as respect learned young. Researchers also tie Chinese comfort with silence in Sino-British meetings partly to power distance. Before correcting a boss, learn how dissent is voiced.

Goffman used face for the positive social value a person claims in an encounter, a claim others can honor or undercut. Facework is the effort to guard face, your own and other people's. Politeness theory names two face needs: positive face, the wish to be approved of, and negative face, the wish to act unimpeded. Rewriting a coworker's slide unasked strains both. Before correcting someone, ask which face your words threaten.

Intercultural competence is the ability to interact effectively and appropriately across cultures. Many models combine attitudes like openness, knowledge of norms, and capabilities like flexibility. A recent review reports that capabilities predict effectiveness most robustly, with attitudes acting mainly through them. Self-reports, the most common measure, are open to flattering bias. Judge yourself by what you can do.

Edward Hall called communication low-context when meaning sits in the explicit words, and high-context when much of it sits in the setting, the relationship and what goes unsaid. Mismatches breed misreads: researchers report Western teachers and colleagues reading Chinese silence as weak language skill or low commitment, when it can signal respect or a wish to avoid conflict. When a pause or a vague yes puzzles you, ask.

Politeness theory counts acts that impose on someone, such as unasked advice, as threats to face, and its most cautious option is to skip the act entirely. Indirectness spares feelings but leaves more to infer. Low-context cultures tend to voice criticism directly; in some high-context settings, silence itself is the polite move. A manager's 'that could be interesting' may mean no. When stakes are high, confirm the meaning.

Hofstede's country scores average many people's survey answers, so they describe a society, not a member. Applying them to one person is the ecological fallacy. Hofstede warned that personalities vary widely within every national culture and that scores should not be used to stereotype anyone. A high US individualism score says little about the engineer you just met. Use a profile to know a society, then learn the person.

Festinger's social comparison theory says people have a drive to assess their abilities and opinions, and that without clear norms or objective information they measure themselves against others. Such comparisons can happen even when objective facts exist. A new runner who runs a 5K in 31 minutes cannot tell if that is good until she sees how others her age did. When unsure how you are doing, check whom you measure against.

Relative deprivation is the discontent people feel when they see themselves as unfairly behind others they compare with. It depends on the comparison, not on objective hardship. Runciman's conditions: you lack something, see others with it, want it, and think it attainable. A nurse pleased with a 5 percent raise can feel cheated on hearing a newer colleague got 8. When resentment flares, check whom you are comparing with.

Tesser's self-evaluation maintenance model starts from the wish to think well of yourself. When someone close to you excels, two things can happen: if the area matters little to you, you bask in their glory; if it is central to who you are, the comparison stings. People restore balance by caring less about the area, pulling away, or changing the gap. If a friend's promotion irritates you, notice which move you are making.

Charles Cooley wrote in 1902 that much of our self-feeling comes from imagining how we appear to another person, imagining their judgment, then feeling pride or shame. The imagined judgment does the work. Researchers call these guesses reflected appraisals, and they can miss: one study found Chinese late adolescents underestimated how peers saw their personalities. Treat your sense of how others see you as a guess to check.

An upward comparison measures you against someone who seems better off; a downward one, against someone who seems worse off. Downward comparison usually lifts mood and self-evaluation, while upward comparison, the more common kind, can bring negative feelings, though it can also spur self-improvement. A junior designer eyeing a star's portfolio can feel small or driven. Notice which way you just compared, and what it did.

Festinger's similarity hypothesis says people prefer comparison targets near their own level, since those comparisons tell them more. Similarity shapes the result too: a target who feels like you can pull your self-view toward theirs, one who feels different pushes it away, and pushing away is more common. Yet among German ninth graders, the class average mattered and friends added little. Notice the group you are ranked in.

Image-led apps such as Instagram often show people at their best. In two survey studies, heavier use came with more upward comparisons, and those comparisons went with lower self-esteem and, in the second study, more depressive symptoms. Yet use itself was not tied to lower self-esteem overall, effects were modest, and the design cannot show cause. When a scroll deflates you, notice whom you just measured yourself against.

In the Clark and Wells model of social anxiety, a person in a feared situation turns attention inward to monitor how they come across. That crowds out the room: they miss the nods showing others are responding well, and judge how they look from how they feel, as in 'my voice feels shaky, so everyone can hear it.' Sam, mid-talk, tracks his pulse instead of his audience. Cognitive therapy includes training attention outward.

Safety behaviors are moves to prevent a feared outcome: scripting topics, avoiding eye contact, saying little. Cognitive models say they backfire: if nothing bad happens, the person credits the move instead of learning the danger was small, and the avoidant kind can make someone seem cold. Priya stays quiet at lunch so no one finds her dull, and never tests that fear. The model asks of each habit: what would happen without it?

Cognitive models of social anxiety rest on two beliefs: a good impression matters greatly, and one tends to come across badly. Rules like 'never allow a silence' make every chat a test. Before, the person rehearses what could go wrong; afterward, a post-mortem replays the worst moments from feelings, not from what others did. NIMH describes social anxiety disorder as persistent fear of being judged that disrupts daily life.

The spotlight effect, studied by Gilovich's team, is the tendency to overestimate how much others watch and evaluate you; people also expect harsher judgments than others make. A likely cause: you judge yourself by your feelings, while others see only your behavior. Jordan spills coffee at work and replays it all day; most colleagues barely noticed. When you feel exposed, ask how much you noticed the last person who slipped.

The liking gap, first reported by Boothby and colleagues, is the tendency to guess that someone you just met enjoyed your company less than they really did. It runs against usual self-flattery: people rate their own traits generously, yet assume others found them less likable. Work with children suggests it appears by about age five. After a chat that felt awkward, lean toward assuming it went better than it felt.

Therapies build practice in different ways. NIMH describes exposure therapy, one CBT method, as facing feared situations step by step, and says CBT can help people practice social skills. A different model, cognitive therapy for social anxiety, skips exposure ladders and skills training: a client states a prediction others could observe, drops a safety behavior, checks the result and reviews video. Dana predicts listeners will look away.

Research on socially anxious children separates three explanations for low social skill. A skill deficit: the behavior was never learned. A performance deficit: the skill is there, but anxiety blocks it. A perceived deficit: the child rates their skills low while observers see nothing wrong, as one study found. Researchers draw a different implication for each: teach the skill, ease the anxiety, or challenge the self-view.