What people say (reported behaviour) and what people do (revealed behaviour) routinely diverge. Across 422 studies, intentions capture only about 28% of the variance in actual behaviour (Sheeran, 2002). And when researchers deliberately push intentions up by a lot, behaviour still moves only a little (Webb & Sheeran, 2006). This isn't a measurement quirk. It's one of the most replicated findings in the field. The practical upshot: treat every "would you / will you" answer as a weak signal, and calibrate it against real behaviour before you bet on it.

The things that actually drive behaviour are usually **latent**: not directly observable, and (the uncomfortable part) not reliably accessible to the person themselves. When you ask people why they did something, they don't retrieve the real cause. They **build a plausible story** out of their own theories about themselves (Nisbett & Wilson, 1977). In lab and supermarket demonstrations of "choice blindness," people will even defend a choice they never actually made (Johansson et al., 2005; Hall et al., 2010). The practical rule: treat a stated reason as a *hypothesis*, and recover real drivers from behaviour, comparison, and experiment. Not from the "why."

When independent teams redid 100 published psychology studies, only about a third got the same result, and the effects that did survive were roughly half as strong (Open Science Collaboration, 2015). Beloved findings like "ego depletion" and the physical claims behind "power posing" largely vanished under careful testing. The causes are mundane: small samples, flexible analysis, and journals that reward surprising results. The takeaway is not that behavioural science is worthless. It is that any single study is provisional, and the findings worth betting on are the replicated, well-powered, preregistered ones.

There is no clean way to measure what people think and do. Surveys are great but people misremember and manage their image. Behavioural logs are honest about what happened but totally silent on why. Clever tools like trade-off tasks and reaction-time tests each capture something real while missing something else, and even the most famous reaction-time test, the IAT, predicts individual behaviour poorly. The fix is not a perfect instrument. It's triangulation: combine methods whose weaknesses don't overlap, and trust the finding that shows up across several of them.

In 1968 Walter Mischel showed that a personality trait predicts any one action only weakly, the famous correlation of about 0.30. The field heard "traits are fake." That was a misreading. Average someone's behaviour across many occasions and the stability jumps right back (Epstein, 1979). The "powerful" situations psychologists love turn out to be no stronger than traits (Funder & Ozer, 1983). And people carry stable 'if-then' signatures: calm at home, combative at work, the same person every time. Traits are real, and they predict major life outcomes (Roberts et al., 2007). The usable rule: who you are predicts tendencies and averages, where you are predicts the specific moment.

Marketers and campaigners borrow each other's tools constantly, and often they should. Voting and buying both run on identity, habit, gut feeling, and mental shortcuts rather than careful comparison (Converse, 1964; Taber & Lodge, 2006). But the two choices are built differently. A purchase is frequent, private, switchable, and hands you the thing you chose. A vote is rare, secret, almost never decisive on its own, and mostly expressive, a statement of who you are (Riker & Ordeshook, 1968; Achen & Bartels, 2016). The rule that keeps you out of trouble: borrow the methods, not the mental model.

We feel like we take in a rich, complete picture of the world. We do not. Attention is a hard bottleneck, and we consciously register only the slice it lands on. Ask people to count basketball passes and about half miss a person in a gorilla suit strolling through the scene (Simons & Chabris, 1999). Swap the stranger someone is talking to mid-conversation and half do not notice (Simons & Levin, 1998). For anyone trying to be noticed, the lesson is blunt: presence is not perception. Being in the room, on the page, or in the feed earns you nothing until you win attention.

Memory feels like playback but it's actually closer to reconstruction. Each time you recall an event you rebuild it from fragments, general knowledge, and anything you have heard since, and the rebuild can be edited without you noticing. Ask witnesses how fast cars were going when they "smashed" rather than "hit" and they report higher speeds, then later remember broken glass that was never there (Loftus & Palmer, 1974). Show people a list of related words and about half will confidently remember a word that was never shown (Roediger & McDermott, 1995). The practical warning is simple: a recalled account is not a record, and the way you ask for it changes it.

Behaviour is shaped by experience along three routes: association (a cue starts to predict an outcome), consequences (rewarded actions repeat, punished ones fade), and observation (we copy what we see others do, with no reward of our own). Two modern corrections matter. Conditioning is really about prediction, not mere repetition: a cue teaches you only if it reliably predicts something (Rescorla, 1988), and the brain learns from the gap between the reward it expected and the reward it got (Schultz et al., 1997). And the schedule of reward is decisive: unpredictable, occasional rewards, the variable-ratio schedule behind slot machines and pull-to-refresh, produce the most persistent, hardest-to-quit behaviour.

A habit is a behaviour your brain has wired to a cue (a time, place, or preceding action) so it fires automatically, with almost no thought. That's why "trying harder" rarely works and why the famous "21 days to a new habit" is a myth (the real figure averages ~66 days, but varies enormously). The practical lever isn't motivation; it's context. Change someone's surroundings, or catch them in a moment when their surroundings have already changed (a house move, a new job, a new baby), and behaviour becomes editable.

The mind runs in two modes. One is fast, automatic, and effortless; the other is slow, deliberate, and demanding. The fast mode handles almost everything and is usually right, but when a problem is tricky it still hands up a quick, confident answer, and the slow mode tends to rubber-stamp it rather than check. Ask people a simple-looking sum (a bat and ball cost $1.10, the bat is $1 more than the ball) and most blurt the wrong intuitive answer, even at top universities (Frederick, 2005). Two cautions: the careful version of the theory is two kinds of processing, not two tidy "systems" in the head (Evans & Stanovich, 2013), and "fast equals irrational" is but a myth. The practical point: most decisions are made on the fast pass, so design for the gut, and recruit the slow mode on purpose only when it matters.

Persuasion runs along two routes. On the central route, people weigh your actual arguments; this happens when they are both motivated and able to think it through, and the attitudes it builds are durable and tend to drive behaviour. On the peripheral route, people lean on shortcuts (who is speaking, how expert or attractive they seem, how many reasons there are, what everyone else thinks); this happens when they do not much care, and the attitudes it builds are weak and fade. The deciding factor is involvement, how much the topic matters to the person. In a classic study, strong-versus-weak arguments swayed people who cared about the product, while a celebrity endorser swayed people who did not (Petty, Cacioppo & Schumann, 1983). The practical move is to match your message to how much your audience actually cares.

When people judge something, they often consult how they feel about it and treat that feeling as information, a process called affect-as-information. The catch is that feelings travel. A mood picked up from the weather, a previous argument, or a good lunch can bleed into a judgment that has nothing to do with its source. In the classic demonstration, people rated their whole lives as more satisfying on sunny days than rainy ones, but the effect vanished the moment someone asked about the weather first, because naming the real source breaks the spell (Schwarz & Clore, 1983). And feelings are not just "good" or "bad": specific emotions of the same flavour can pull in opposite directions, with fear making people cautious about risk and anger making them bold (Lerner & Keltner, 2001). The practical upshot: feelings are real data, but check where they came from before you trust the reading.

We tend to treat motivation as one quantity: add a reward, get more of the behaviour. The science says the type and direction matter more than the amount. The brain's "want it" system can come apart from its "like it" system, so people chase things they no longer enjoy. Rewarding an activity someone already finds interesting can backfire, turning play into work and lowering their spontaneous interest once the reward stops (Deci, 1971; Lepper, Greene & Nisbett, 1973). What sustains motivation over time is not pressure but three needs being met: autonomy, competence, and relatedness (Ryan & Deci, 2000). And two famous ideas have not held up: Maslow's strict pyramid of needs and the notion that willpower is a fuel tank that runs dry.

The textbook says a rational chooser ranks options by their odds times their payoffs. People reliably do something else. We feel outcomes as gains or losses measured from wherever we happen to be standing, not as final totals, so the same result can please or sting depending on the reference point (Kahneman & Tversky, 1979). We overweight rare events, and we flip between caution and risk-taking in a predictable fourfold pattern. These patterns are robust: they reproduced across 19 countries (Ruggeri et al., 2020). We also lean on fast shortcuts, and there is a real fight about what they mean: one camp sees them as biases measured against the ideal maths (Tversky & Kahneman, 1974), another sees them as smart tools that often beat the maths when the world is genuinely uncertain (Gigerenzer & Brighton, 2009). The practical takeaway: find the reference point, watch the framing, and match the tool to how knowable the odds actually are.

People do not judge risk by the numbers. They judge it by feel, and the feeling runs on a few specific levers. We dread risks that seem uncontrollable, catastrophic, and involuntary (a plane crash, a meltdown) and we shrug at risks that are familiar and chosen (driving, junk food), almost regardless of the actual death tolls (Slovic, 1987). We also read a single good-or-bad feeling about a thing and use it to judge both its risk and its benefit, so anything we like feels safe and anything we distrust feels dangerous (Finucane et al., 2000). Emotion, not arithmetic, tends to win when the two disagree (Loewenstein et al., 2001). And on charged topics like climate, more knowledge does not calm the fear; it sorts people deeper into their existing camps (Kahan et al., 2012). The practical upshot: to change how risky something feels, work on the feeling and the qualities behind it, not just the statistics.

People place far more weight on now than on later, a tendency called discounting. The trouble is not just that we are impatient, but that we are impatient in a lopsided way: we discount the near future steeply and the far future gently, so our preferences flip as a reward gets close. Most people will wait a week for an extra ten pounds if the whole thing is a year away, then grab the smaller sum the moment it is available today (Ainslie, 1975; Laibson, 1997). That present bias drives procrastination, overspending, and broken diets, and the things that actually fix it are commitment devices and good defaults, because they work around our future self instead of trusting it (O'Donoghue & Rabin, 1999; Thaler & Benartzi, 2004). One famous claim, that a four-year-old's willpower with a marshmallow predicts their whole future, has been substantially deflated (Watts et al., 2018).

A bias is a shortcut that misfires in a predictable direction. The famous catalogue, anchoring, availability, framing, confirmation bias, and the rest, came from showing people behave in ways a textbook of pure logic would not predict (Tversky & Kahneman, 1974). Two things have happened since. A replication reckoning sorted the catalogue into the solid (anchoring, framing) and the shaky (several crowd-pleasers that failed to reproduce), so a single eye-catching study is no longer enough (Open Science Collaboration, 2015). And a long-running argument questions whether "bias" is even the right word: present the same problem differently and many errors shrink or vanish, which suggests some of them live in the question, not the person (Gigerenzer, 1996). The practical skill is to lean on the biases that replicate, distrust the ones that do not, and remember that the fix is often to change the format, not to scold the thinker.

Personality science tries to find the smallest set of dials on which people reliably differ, and the dominant answer is the Big Five: Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. It came from a simple idea, that the differences that matter get baked into language, plus a lot of factor analysis (Goldberg, 1990). These traits are not trivia: they predict health, divorce, and career success about as strongly as wealth or intelligence do (Roberts et al., 2007). They are stable enough to count on but not frozen, drifting in predictable directions as people mature (Roberts, Walton and Viechtbauer, 2006). The big caveat is that the tidy five-factor map holds best in literate, Western samples and frays in others (Gurven et al., 2013). Treat it as an excellent description of people, not an explanation of them, and never as a set of fixed types.

Temperament is the earliest layer of who we are: biologically rooted differences in reactivity (how strongly and fast you respond to things) and self-regulation (how well you rein that response in), visible in the first months of life. The field has largely settled on three broad dimensions, and one well-studied type, behavioural inhibition (the wary, slow-to-warm child), is one of the biggest early risk factors for later anxiety (Clauss and Blackford, 2012). But temperament is a tilt, not a verdict: roughly half of even strongly inhibited children do not become anxious adults, because what happens next depends heavily on how well the environment fits the child. The honest cautions: most of the data come from parents rating their own kids, and "difficult" says as much about the setting as about the child.

Beyond raw ability, people have stable preferences about thinking itself. Two matter most. Need for cognition is how much you enjoy effortful thought: high-need-for-cognition people seek out hard thinking and are moved by the strength of an argument, while low-need people lean on quick cues (Cacioppo and Petty, 1982). Need for closure is how badly you want a definite answer and how much ambiguity grates: high-closure people decide fast, grab the first reasonable answer, and resist changing their minds (Kruglanski and Webster, 1996). These styles are real and useful for tailoring how you communicate, but two cautions run through everything below: they overlap heavily with the personality trait Openness, so they may be less new than they look, and the famous link between closure-seeking and political conservatism is genuine but modest and hotly contested.

Emotion regulation is how people steer which feelings they have and how they show them. The most studied contrast is reappraisal (reframing a situation so it hits differently) versus expressive suppression (keeping a straight face while the feeling runs on underneath). On average, habitual reappraisal goes with better mood, relationships, and health, while habitual suppression goes with worse, and in the lab suppression cuts the outward display but leaves the feeling and even spikes the body's stress response (Gross, 1998a; Gross and John, 2003). But the tidy "reframe good, bottle bad" story oversimplifies. No strategy is best in the abstract: the strongest predictor of coping well is flexibility, matching the strategy to the moment and switching when it is not working (Bonanno and Burton, 2013). Suppression has its place, the effects are modest, and the costs of bottling up are smaller in cultures that prize it.

Attachment theory says our earliest bonds leave us with a working model of closeness, an unspoken sense of whether people show up and whether we are worth showing up for, that colours later relationships (Bowlby, 1969; Hazan and Shaver, 1987). Researchers now capture it as two dials rather than four boxes: anxiety (how much you fear being abandoned) and avoidance (how much closeness makes you uneasy). Low on both is what "secure" means, and it tracks with better relationships, steadier emotions, and better mental health (Mikulincer and Shaver, 2007). The real finding is solid. The popular version oversells it: the styles are points on continua, not types a quiz can stamp on you (Fraley and Spieker, 2003); the link from infancy is moderate, not destiny (Fraley, 2002); and the neat causal story from parent to child still has a hole in it (Verhage et al., 2016).

People who do well on one kind of mental test tend to do well on others, even very different ones, and that web of positive correlations is the most replicated finding in the field. We summarise it as general intelligence, or g (Spearman, 1904), and tests of it are reliable, stable, and predict real outcomes like school results, work, and health, modestly to moderately (Deary et al., 2007; Plomin and Deary, 2015). That much is solid. The overreach starts straight after. The famous "best predictor of job performance" figure was recently cut once the statistics were re-checked (Sackett et al., 2022); measured intelligence rose for most of the 20th century and has since fallen in places, both for environmental reasons, so it is clearly not a fixed innate number (Flynn, 1987; Bratsberg and Rogeberg, 2018); and intelligence being substantially heritable does not make it unchangeable or explain gaps between groups (Nisbett et al., 2012). A real, narrow, partly movable signal, routinely treated as a verdict on a person's worth.

Psychologists study a cluster of socially aversive traits: narcissism (grandiosity and entitlement), Machiavellianism (cold, strategic manipulation), and psychopathy (callousness and impulsivity), together the Dark Triad, with everyday sadism (enjoying cruelty) sometimes added as a fourth (Paulhus and Williams, 2002; Buckels, Jones and Paulhus, 2013). They overlap so much that many researchers treat them as flavours of one "dark core", the tendency to put your own gain above others and to believe you are justified in doing so (Moshagen, Hilbig and Zettler, 2018). The popular story is that a measured dose of these traits makes people effective, especially as leaders. The data do not support it: dark traits help people climb and get noticed a little, but predict worse behaviour at work and, if anything, slightly worse leadership, not better (O'Boyle et al., 2012; Landay et al., 2019). They are also dimensions almost everyone has a little of, not boxes to file enemies into.

Personality and disposition are measured mostly by self-report, like "tick how much each statement sounds like you". Done well it is reliable and genuinely predictive, so this is not a reason to throw the method out. But it does bend in known ways. People shade answers to look good, especially when it counts (Crowne and Marlowe, 1960; Paulhus, 1984); they fall into content-free habits like always agreeing (Baumgartner and Steenkamp, 2001); in high-stakes settings they fake, and the tests may have been weak predictors to begin with (Morgeson et al., 2007); and because everyone rates themselves against their own local crowd, self-report averages cannot be compared across cultures and can even come out backwards (Heine et al., 2002). The lesson is not to distrust all of it, it is to read a self-report number knowing what could have moved it. This is the article that puts an asterisk on every "people high in X report" sentence in this whole layer.

A belief is how strongly you hold that something is true, and the ideal is to shift that confidence in step with the evidence. Real people fall short: they update sluggishly, lean toward conclusions they already want (Kunda, 1990), and hold beliefs in connected webs, so moving one tugs on others (Thagard, 1989). For a decade the headline was the backfire effect, the claim that corrections make false beliefs stronger (Nyhan and Reifler, 2010). It turned out to be rare: large replications found people mostly do move toward the facts, just modestly (Wood and Porter, 2019; Guess and Coppock, 2020). The beliefs that truly resist are the ones fused to identity, where more knowledge can mean more polarisation, not less (Kahan et al., 2012). The practical upshot is hopeful: evidence usually helps, so the job is to make updating feel safe rather than threatening.

A belief is how strongly you hold that something is true, and the ideal is to shift that confidence in step with the evidence. Real people fall short: they update sluggishly, lean toward conclusions they already want (Kunda, 1990), and hold beliefs in connected webs, so moving one tugs on others (Thagard, 1989). For a decade the headline was the backfire effect, the claim that corrections make false beliefs stronger (Nyhan and Reifler, 2010). It turned out to be rare: large replications found people mostly do move toward the facts, just modestly (Wood and Porter, 2019; Guess and Coppock, 2020). The beliefs that truly resist are the ones fused to identity, where more knowledge can mean more polarisation, not less (Kahan et al., 2012). The practical upshot is hopeful: evidence usually helps, so the job is to make updating feel safe rather than threatening.

A value is a broad, lasting guiding principle, freedom, security, achievement, kindness, and what matters is not whether you hold it (almost everyone endorses almost all of them) but how you rank it against the others. Shalom Schwartz showed that these priorities fall into a remarkably consistent circle: ten basic values (refined to nineteen) arranged on two axes, openness to change versus conservation, and self-enhancement versus self-transcendence (Schwartz, 1992; Schwartz et al., 2012). Neighbours on the circle support each other; opposites conflict, so you cannot top-rank both power and universal compassion. This structure replicates across dozens of countries, one of the most robust findings in the field (Bilsky, Janik and Schwartz, 2011). The honest catch: values predict broad leanings far better than specific actions, and the link to behaviour is modest (Bardi and Schwartz, 2003).

A goal is a value (L3-03: Values) made concrete: a state you are steering toward or away from. The deepest split is direction, approach (move toward a reward) versus avoidance (prevent a bad outcome), which shows up in personality as a promotion focus on gains and a prevention focus on safety (Higgins, 1997). Avoidance goals tend to cost more in motivation and performance (Elliot and Church, 1997), but the real driver is fit: a goal pursued in a way that matches your focus feels better and pulls harder (Higgins, 2000). Three tools reliably turn goals into action: making them specific and challenging (Locke and Latham, 2002), attaching a simple if-then plan (Gollwitzer, 1999), and tying them to identity (Oyserman and Destin, 2010). The catch: aggressive goals have a documented dark side, tunnel vision, gaming, and cut corners (Ordóñez et al., 2009).

People do not weigh gains and losses evenly: the pain of a loss tends to outweigh the pleasure of an equal gain, often put at roughly two to one (Kahneman and Tversky, 1979; Tversky and Kahneman, 1992). It follows that what counts as a gain or a loss depends entirely on your reference point, and that owning something makes giving it up feel like a loss, the endowment effect (Kahneman, Knetsch and Thaler, 1990). The pattern is part of a broader rule that bad outweighs good across most of mental life (Baumeister et al., 2001). But the clean "2:1 law" has had a reckoning: it is real and common, yet smaller, absent, or reversed in many contexts (Gal and Rucker, 2018), with moderators rather than a fixed value (Mrkva et al., 2020). Treat the asymmetry as a strong, useful tendency, not a constant.

"Confidence" hides three separate ideas. Self-concept is your picture of who you are (Markus, 1977). Self-esteem is how good you feel about yourself overall (Rosenberg, 1965). Self-efficacy is your belief that you can do a specific task (Bandura, 1977). They are not interchangeable. Self-efficacy is the one that reliably predicts effort, persistence, and performance, and the one you can deliberately grow, mainly by stacking up small real successes (Stajkovic and Luthans, 1998). Self-esteem, by contrast, turns out to be mostly a result of how life is going rather than a cause of success, which is why the campaign to raise everyone's self-esteem largely failed to deliver (Baumeister et al., 2003). And "low self-esteem causes bad behaviour" is itself a myth: the real risk is fragile, inflated self-regard (Bushman and Baumeister, 1998).

Identity comes in layers. There is the personal self that makes you an individual, the social selves you draw from the groups you belong to, and the role selves attached to the positions you hold (Stryker and Burke, 2000). The social ones are deceptively powerful: put people into even meaningless, coin-flip groups and they immediately favour their own (Tajfel et al., 1971). Which identity is running depends on the situation (Turner et al., 1987), which is why the same person can be calm at work and tribal at the match. Two correctives matter most. In-group bias is not mainly about self-esteem (Rubin and Hewstone, 1998), and, crucially, loving your own group is not the same as hating others (Brewer, 1999): belonging does not require an enemy. At the extreme, when a personal and a group self fuse, people will do drastic things for the group (Swann et al., 2012).

The single most useful thing here is a clean set of definitions. Sex is the biological side (chromosomes, hormones, anatomy), largely male or female but with some natural variation. Gender identity is your inner sense of being a man, a woman, both, or neither. Gender expression is how you present, and which norms you follow. Sexual orientation is who you are drawn to. These four vary partly independently, so knowing one tells you little about the others. A few findings are well established: men and women are more psychologically alike than different (Hyde, 2005); orientation is not a choice, has biological roots, but is not a single "gene" and is not perfectly fixed (Bailey et al., 2016; Ganna et al., 2019). Some genuine debates remain, and the heated questions of clinical care and policy are a separate matter from these definitions and findings.

From late adolescence on, people build an internalised, evolving life story that knits the remembered past to the imagined future and gives a life some unity and meaning. Psychologists call it narrative identity, and treat it as a third layer of personality, sitting above traits and goals (McAdams and Pals, 2006). The story is built, not recorded: memory is reconstructive, so we assemble it for meaning and revise it as we change (Conway and Pleydell-Pearce, 2000). The kind of story matters. People whose stories find growth in hardship, so-called redemption sequences, and who cast themselves as agents, tend to have higher wellbeing than those whose stories run the other way (McAdams et al., 2001; McAdams and McLean, 2013). Two honest caveats: the cause-and-effect is not settled, and the tidy redemptive arc is partly a cultural, very American, ideal.

Under specific attitudes run broad worldviews, integrative beliefs about how the social and moral world works, and four have been studied most. Belief in a just world is the assumption that people get what they deserve; it calms personal anxiety but can tip into blaming victims (Lerner and Simmons, 1966; Dalbert, 1999). System justification is the motive to defend the existing order as fair and legitimate (Jost and Banaji, 1994), though the strong claim that the disadvantaged justify the system most failed a large-scale test (Brandt, 2013). Right-wing authoritarianism and social dominance orientation are two dispositions, one toward submission and conformity, one toward group hierarchy, that together predict prejudice well (Pratto et al., 1994; Duckitt, 2001). The honest cautions matter: these frames correlate with politics but are not the same as it, the labels oversell, and the science can be misused both to manipulate people and to dismiss them.

Moral Foundations Theory (Haidt and Joseph, 2004) says moral judgment is intuition-first and built on several distinct concerns, usually listed as care, fairness, loyalty, authority, and sanctity, with liberty added later (Iyer et al., 2012). Its best-supported finding is descriptive: across many studies, political liberals lean heavily on care and fairness, while conservatives draw on all of them more evenly (Graham, Haidt and Nosek, 2009). That is a real and useful insight into moral disagreement. The theory's deeper claims are contested. The idea that these are innate mental modules is disputed (Suhler and Churchland, 2011), a strong rival argues that all moral judgment really tracks perceived harm (Schein and Gray, 2018), and the exact five-part structure holds up unevenly across cultures (Iurino and Saucier, 2020). Treat the foundations as a useful map of moral emphasis, not as settled architecture.

An ideology is supposed to be an organised, internally consistent system of political beliefs. The uncomfortable finding is that most ordinary citizens do not have one: their issue positions hang together loosely and wobble over time (Converse, 1964; Kinder and Kalmoe, 2017). Left-right is still a real psychological dimension with modest dispositional correlates (Jost, 2006), and people's self-labels often do not match their actual policy views (Ellis and Stimson, 2012). What does most of the organising is identity, not doctrine: people back a policy largely because their side proposed it, while denying any such influence (Cohen, 2003), and the dominant feature of modern politics is affective polarisation, disliking the other side more than disagreeing with it (Iyengar, Sood and Lelkes, 2012). Treat "ideology" as a loose identity, not a tight belief system.

Possible selves are your images of who you might become, the hoped-for, the expected, and the feared (Markus and Nurius, 1986). They give goals their emotional charge, because the gap between who you are and who you want to be is felt, not just noted (Higgins, 1987). But a hopeful image on its own is mostly a daydream. What turns it into a force is balance and strategy: pairing a hoped-for self with the feared self it guards against, and linking both to concrete next steps (Oyserman and Markus, 1990; Oyserman, Bybee and Terry, 2006). And which futures you chase matters for how the life feels: aspirations centred on growth, relationships, and contribution support wellbeing, while those centred on wealth, fame, and image tend to erode it, even when reached (Kasser and Ryan, 1996; Deci and Ryan, 2000).

Taste looks like private aesthetics but works partly as social signalling: it marks and reproduces class position (Bourdieu, 1984), though the modern high-status move is breadth, the cultural omnivore, rather than snobbery (Peterson and Kern, 1996). Underneath, preferences are formed more than revealed. Mere repeated exposure makes us like things, often with no memory of the exposure (Zajonc, 1968), a robust effect (Bornstein, 1989) that reverses if you overdo it (Berlyne, 1970). And the same person will rank the same options differently depending on how they are asked, which means a stable "true" preference may not be sitting there to be read off (Lichtenstein and Slovic, 1971). Much of what feels like taste is exposure, position, and framing.

Wellbeing splits along an ancient line. The hedonic view equates it with feeling good, measured as subjective wellbeing: plenty of positive feeling, little negative, and satisfaction with your life (Diener, 1984). The eudaimonic view equates it with functioning well, measured as psychological wellbeing: autonomy, growth, purpose, mastery, good relationships, self-acceptance (Ryff, 1989), and closely tied to self-determination theory (Ryan and Deci, 2000). The two correlate so highly that whether they are truly separate is contested (Kashdan et al., 2008). And the famous question of whether money buys happiness has an unusually honest answer, because two rivals who disagreed ran a joint study and resolved it (Killingsworth, Kahneman and Mellers, 2023).

Religion and spirituality function as comprehensive worldviews: meaning-systems that bind together beliefs about ultimate reality, moral norms, ritual, identity and community into a single frame held with such "an aura of factuality" that it feels simply like how things are (Geertz, 1973). Behavioural science studies the mechanisms of that frame, how it arises, spreads and shapes behaviour, and it does so methodologically, without adjudicating whether any religious claim is true, which is a distinction worth stating plainly up front. The cognitive science of religion finds that supernatural belief arises partly as a byproduct of ordinary mental equipment: a quick-triggered tendency to detect agents and intentions behind events, the same theory-of-mind we use on each other, and a memory bias toward "minimally counterintuitive" ideas, mostly ordinary with one striking violation, that are easy to remember and pass on (Barrett, 2000; Boyer, 2001). Belief spreads not mainly through argument but through credibility-enhancing displays: people adopt a belief more readily when they watch others pay real, hard-to-fake costs for it, fasting, tithing, risky ritual (Henrich, 2009). And worldviews do real coordinating work, binding morality, identity and community. Two honest cautions travel with all this. Grand causal stories about religion and civilisation are empirically fraught: the high-profile claim that moralising "big gods" drove large-scale cooperation is genuinely unresolved, and a 2019 Nature paper that seemed to settle the timing was retracted in 2021 after its result was shown to flip depending on how missing data were coded (Beheim et al., 2021). And the well-known links between religion and wellbeing are real but small and heavily confounded with plain social connection (Koenig et al., 2012). The practical upshot is that if you want to reach people across a worldview, you have to understand the frame, what it holds sacred, what it treats as a threat, what it will accept as credible, rather than simply presenting facts, because facts that collide with a sacred frame tend to bounce off.

The desire for stable, caring relationships is a fundamental human motivation, not an optional extra, and it shapes emotion, thought, and health across the board (Baumeister and Leary, 1995). Self-esteem seems to work partly as its gauge, a sociometer that reads how accepted you are (Leary et al., 1995). Thwart the need and it hurts fast: even being ignored by strangers in a trivial computer game reliably stings (Williams, 2007), exclusion can turn people aggressive or self-defeating (Twenge et al., 2001), and chronic disconnection carries a mortality risk on the scale of smoking (Holt-Lunstad, Smith and Layton, 2010). Which groups you belong to, and how identity works, is the story of L3-07; this is about the need itself, and how easily it is both met and exploited.

A group is more than the sum of its members; it exerts real force on each of them (Lewin, 1947). The best-established example is conformity: faced with a unanimous majority giving an obviously wrong answer, many people go along with it, against the plain evidence of their eyes (Asch, 1956), an effect found across cultures though it varies in size (Bond and Smith, 1996). The most useful and least-known part is the cure: a single dissenting ally shatters the pressure. The famous cautionary tales, groupthink and the Robbers Cave study, are shakier than their fame implies: groupthink barely survives testing (Esser, 1998), and Robbers Cave was effectively stage-managed (Perry, 2018). In group dynamics, the loudest findings are often the weakest.

Social norms come in two kinds that are easy to muddle. A descriptive norm is what most people do; an injunctive norm is what most people approve of (Cialdini, Reno and Kallgren, 1990). They move us more than we realise, and largely without our noticing (Nolan et al., 2008). The catch is that a descriptive norm pulls everyone toward the average, which helps if you are above it and backfires if you are below it, the "boomerang" (Schultz et al., 2007). So a campaign that announces "most people do the bad thing" advertises the bad thing as normal and can increase it. The honest caveats: the effects are real but modest and context-dependent (Bergquist, Nilsson and Schultz, 2019), and a norm only grips you if it belongs to a group you count as yours.

The desire for status appears to be a fundamental human motive, universal and consequential for wellbeing (Anderson, Hildreth and Howland, 2015). Crucially, rank comes by two distinct routes: dominance, where deference is extracted by intimidation and fear, and prestige, where deference is freely given to people seen as skilled or valuable (Henrich and Gil-White, 2001). Both win influence, but through opposite signatures, the dominant are watched warily, the prestigious are sought out and copied (Cheng et al., 2013). Hierarchies, once formed, tend to reinforce themselves (Magee and Galinsky, 2008). And a hopeful twist: the status that actually raises wellbeing is local respect from the groups you see face to face, far more than money or rank in the abstract (Anderson et al., 2012).

Loneliness is the felt gap between the connection you want and the connection you have, which is why you can be lonely in a crowd and content on your own; it is distinct from objective isolation (Weiss, 1973). The leading model treats it as an evolved aversive signal, like hunger, that motivates re-connection and, when chronic, dysregulates stress, sleep, and immune function (Cacioppo and Hawkley, 2009). Weak social connection is associated with markedly higher mortality (Holt-Lunstad, Smith and Layton, 2010; Holt-Lunstad et al., 2015), though the popular "as bad as smoking" line overstates a real but smaller effect. Loneliness is real, consequential, and, importantly, a signal to act rather than a verdict on your worth.

Trust is the willingness to be vulnerable to someone on the expectation they will not exploit you; reciprocity is the near-universal norm that we return good for good, and often harm for harm (Gouldner, 1960). Economists made trust measurable with the trust game, where you hand a stranger money that grows, and see how much they send back (Berg, Dickhaut and McCabe, 1995). Reciprocity keys on intentions, not just outcomes, so the same gift can be repaid or resented depending on what the giver could have done instead (Falk and Fischbacher, 2006). Repeated interaction lets cooperation sustain itself through simple reciprocity, the logic of tit-for-tat (Axelrod and Hamilton, 1981), and at scale, trust is economic infrastructure (Knack and Keefer, 1997). The honest caveat: what people say about trust predicts what they do surprisingly poorly (Glaeser et al., 2000).

Treat people as nodes and relationships as links, and the structure starts to explain behaviour. Novel information reaches you mostly through weak ties, acquaintances who bridge to social worlds your close friends do not touch (Granovetter, 1973); people who span the gaps between clusters, filling structural holes, get an information and control advantage (Burt, 2004). A few long-range links make the whole world small (Watts and Strogatz, 1998), and a handful of hubs dominate many networks (Barabási and Albert, 1999). The famous claim that obesity and happiness spread through your friends to three degrees (Christakis and Fowler, 2007) is where honesty is needed: observational network data cannot cleanly separate contagion from the fact that similar people cluster together (Shalizi and Thomas, 2011). Experiments can, and they show behaviour spreads best not through lone influencers but through clustered reinforcement (Centola, 2010; Centola and Macy, 2007).

Family is where we are first socialised, and kinship is the wider web of descent and alliance we are born into. Biology built in a bias toward relatives: we help kin roughly in proportion to how related we are (Hamilton, 1964), and reciprocity handles the rest (Trivers, 1971). But humans raise children cooperatively, through a wide net of helpers, not in isolated pairs (Hrdy, 2009), and family relationships run on a distinct moral grammar of sharing rather than exchange (Fiske, 1992). The catch is how much we mistake the local for the eternal: the "traditional" nuclear family was a mid-century anomaly, not the human default (Coontz, 1992), and the familiar claim that family structure determines a child's fate is deeply confounded with class and resources (McLanahan, Tach and Schneider, 2013).

Class is your structured position in the economic order, and it predicts an astonishing amount: how long you live, whether you climb or stay put, how you vote, and the chances your children get. There are two founding lenses. Marx defined class by who owns the means of production (Marx, 1867); Weber split it into three independent dimensions, economic class, social status, and organised power, that need not line up (Weber, 1922). Bourdieu reframed it as three kinds of capital, economic, cultural, and social, with cultural capital, taste and credentials and know-how, as the engine that quietly passes advantage down the generations (Bourdieu, 1979), a mechanism you can watch directly in how different classes raise their children (Lareau, 2002). The honest complications: most people badly misjudge their own class, the "death of class" is oversold because class still predicts hard outcomes, and how far you can climb depends heavily on where you happen to grow up (Chetty et al., 2014).

Cross-cultural researchers try to capture the differences between societies with a small set of dimensions. Hofstede's are the most famous, individualism, power distance, uncertainty avoidance and more, each country given a score (Hofstede, 1980). Later work sharpened the picture: the independent versus interdependent sense of self (Markus and Kitayama, 1991), a values map that shifts predictably as societies get richer (Inglehart and Baker, 2000), and Gelfand's tight versus loose, how strictly a culture enforces its norms, rooted in a history of threat (Gelfand et al., 2011). The catch, and it is a big one, is that these are averages of national populations, not descriptions of individuals, and much of the underlying evidence comes from a single unrepresentative slice of humanity (Henrich, Heine and Norenzayan, 2010). Use the dimensions to form hypotheses, never to predict a person.

Rituals are patterned symbolic acts whose real payload is emotional and social, not practical. Durkheim's founding claim is that performing them together produces collective effervescence, a shared emotional charge that bonds people and marks off the sacred from the ordinary (Durkheim, 1912). They also carry us across life's thresholds (van Gennep, 1909) and, in Collins's account, run through daily life as small charged encounters that leave us wanting more (Collins, 2004). This is not just old theory: controlled experiments show rituals genuinely ease grief (Norton and Gino, 2014), make food more enjoyable (Vohs et al., 2013), and, in intense collective form, boost generosity to the group (Xygalatas et al., 2013). The honest caveat is that "ritual" gets stretched to cover any habit, and the classic functionalist story can be circular (Bell, 1992).

Researchers study religion along two axes: as an institution (organised bodies with doctrine, hierarchy, and membership) and as practice (belief, ritual, identity, and a way of living). Durkheim saw its social function, binding a community through shared symbols (Durkheim, 1912); Weber saw a causal force, arguing a particular Protestant discipline helped set off modern capitalism, a thesis still argued over (Weber, 1905). A later school modelled religion as a kind of market, where competition between faiths can raise participation (Stark and Bainbridge, 1987). The long-standing prediction that modernity would simply dissolve religion has not held up cleanly, and its own authors walked it back, though a revised version tied to how secure people feel survives (Norris and Inglehart, 2004). None of this settles what is true; it describes what religion does.

Researchers study religion along two axes: as an institution (organised bodies with doctrine, hierarchy, and membership) and as practice (belief, ritual, identity, and a way of living). Durkheim saw its social function, binding a community through shared symbols (Durkheim, 1912); Weber saw a causal force, arguing a particular Protestant discipline helped set off modern capitalism, a thesis still argued over (Weber, 1905). A later school modelled religion as a kind of market, where competition between faiths can raise participation (Stark and Bainbridge, 1987). The long-standing prediction that modernity would simply dissolve religion has not held up cleanly, and its own authors walked it back, though a revised version tied to how secure people feel survives (Norris and Inglehart, 2004). None of this settles what is true; it describes what religion does.

How something is described changes what people choose, even when the underlying facts are identical. The classic demonstration reverses people's appetite for risk simply by describing outcomes as lives saved rather than lives lost (Tversky and Kahneman, 1981). In communication research, to frame is to select some aspects of reality and make them salient, steering how a problem, its cause, and its fix are understood (Entman, 1993). Deeper still, the metaphors built into everyday language may shape how we reason about a topic, whether we treat crime as a beast to be hunted or a virus to be cured (Lakoff and Johnson, 1980; Thibodeau and Boroditsky, 2011). The honest caveats matter: framing effects are real but usually moderate, not overwhelming (Kühberger, 1998), the strongest metaphor claims are shakier than the popular version suggests, and even negating a frame can reinforce it (Lakoff, 2004).

Power is the capacity to produce intended effects; authority is power seen as legitimate, and legitimacy is what makes it stick. Weber's classic map sorts legitimate authority into three kinds, traditional, charismatic, and legal-rational (Weber, 1922), while French and Raven catalogued the everyday bases of power, from reward and coercion to expertise and personal appeal (French and Raven, 1959). The famous evidence is Milgram's, where most people delivered what they believed were dangerous shocks on an experimenter's say-so (Milgram, 1963). But the modern reappraisal, built on the original archives, reframes this: people were not mindlessly obeying so much as actively identifying with what looked like a worthy cause, "engaged followership" rather than blind obedience (Reicher, Haslam and Smith, 2012). That is arguably more troubling, not less.

A cohort is a group who were born around the same time and so meet history at the same age; a generation is a cohort claimed to share a distinctive character shaped by its formative years (Mannheim, 1928). The trouble is that any snapshot difference between age groups blends three things that look identical: an age effect (people change as they get older), a period effect (something is true of everyone right now), and a cohort effect (a lasting imprint from when a group grew up). Telling them apart needs longitudinal data and careful modelling, and when researchers do it, most claimed "generational" differences shrink toward nothing (Costanza et al., 2012; Rudolph and Zacher, 2017). Some cohort effects are real, but the familiar labels, Gen X, Millennial, Gen Z, are marketing categories, not scientific findings, which is why even Pew stopped leaning on them (Parker, 2023).

A life stage is an age-graded position with its own expected tasks and norms, such as adolescence, parenthood, or retirement; the lifecourse is the whole trajectory linking them, sensitive to timing, sequence, and the historical moment (Elder, 1994). The influential stage theories, Erikson's eight psychosocial stages and Levinson's "seasons", are evocative but hold up poorly outside their original samples (Erikson, 1950; Levinson, 1978). The modern, better-grounded view is Elder's life-course paradigm, whose most useful ideas are that the timing of a transition changes its meaning (a divorce at 25 is not a divorce at 55) and that lives are linked, so one person's transition ripples through their network. The honest complication: the tidy "standard biography" of school, work, marriage, children, retirement was a brief historical anomaly and has fragmented, so life stages are far less fixed than they feel (Beck, 1986; Bauman, 2000).

Institutions are the rules, both formal ones like laws and contracts and informal ones like conventions and norms, that structure repeated human interaction and reduce uncertainty (North, 1990). Their quality, whether property is secure, contracts are enforced, power is constrained, is one of the deepest explanations for why some societies grow and others stagnate (Acemoglu, Johnson and Robinson, 2001). Ostrom showed that ordinary people also build their own effective institutions to manage shared resources, often outperforming both the state and the market (Ostrom, 2010). And organisations within a field tend to copy each other until they all look alike, chasing legitimacy more than efficiency (DiMaggio and Powell, 1983). The honest caveat is that "institutions matter" is slippery: good institutions and good outcomes cause each other, and the concept can expand until it explains nothing.

Media ecology studies media as environments, not channels: the form of a medium changes us, not only the content it delivers, which is McLuhan's famous "the medium is the message" (McLuhan, 1964). Each shift, oral to print to electronic to the feed, rewires public life; Postman argued the move from print to television turned serious discourse into entertainment (Postman, 1985), and Meyrowitz saw electronic media collapsing the boundary between backstage and front-stage decades before social media (Meyrowitz, 1985). The honest counterweight matters here: the grand theorists are evocative but hard to test, measured media effects are usually smaller than the rhetoric implies, and the popular "filter bubble" story is substantially overstated, since people encounter more cross-cutting views than assumed (Bakshy, Messing and Adamic, 2015), and deliberately exposing them to the other side can even deepen division (Bail et al., 2018).

Geographers distinguish space (abstract, geometric) from place (location made meaningful by the life lived in it), and the bond people form with places, topophilia, is real and powerful (Tuan, 1977). Sociologists show that neighbourhoods exert their own effects on health, safety, and opportunity beyond the individuals in them: a neighbourhood's collective efficacy, its residents' shared willingness to act for the common good, predicts lower violence even after accounting for poverty (Sampson, Raudenbush and Earls, 1997), and the geography of opportunity is enormous and persistent (see L4-09). The honest caveat is the selection problem: people are not randomly assigned to neighbourhoods, so some apparent "neighbourhood effects" are really sorting, and when researchers ran an actual experiment moving families, the effects were real but more modest and more age-dependent than observation implied (Chetty, Hendren and Katz, 2016).

Lifestyle is the patterned bundle of practices, consumption, leisure, taste, the way you carry yourself, that both signals and quietly reproduces your social position. Bourdieu's central idea is habitus: a system of durable, largely unconscious dispositions absorbed through your upbringing in a particular social position, which then generates your "spontaneous" tastes and choices, so taste is a class-marked pattern, not a free individual act (Bourdieu, 1979). Giddens countered that in late modernity lifestyle becomes more of a reflexive, chosen project of the self (Giddens, 1991), though reflexivity itself is unevenly available by class. The honest complications are real: Bourdieu's neat map of taste to class does not hold as cleanly as claimed, with high-status people now tending to be cultural omnivores who enjoy both highbrow and popular culture (Chan and Goldthorpe, 2007), and the commercial "lifestyle segmentations" built on all this are often empirically weak (Plummer, 1974, began the tradition; the validity problems came later).

Voting is two separate questions that get constantly conflated: whether someone votes (turnout) and for whom (choice). Choice is remarkably hard to move: it is anchored in early, stable party identification and group identity far more than in issue positions (Campbell et al., 1960; Achen and Bartels, 2016), and a meta-analysis of 49 field experiments found the average persuasive effect of campaign contact in general elections is essentially zero (Kalla and Broockman, 2018). Turnout, by contrast, genuinely moves: the strongest reliable lever is social pressure, telling people their neighbours will know whether they voted (Gerber, Green and Larimer, 2008). The honest headline for anyone in this business: most of what is sold as persuasion is really mobilising the base in disguise, and the "persuadable middle" is far thinner than the briefs claim.

Decades of buying data reveal patterns so regular they are called laws. Brand loyalty is mostly a by-product of size, not devotion: small brands suffer double jeopardy, they have fewer buyers and those buyers are slightly less loyal (Ehrenberg, Goodhardt and Barwise, 1990). A brand's customers also buy its rivals, roughly in proportion to those rivals' market shares, so the "loyal tribe" is largely a myth. The practical doctrine that follows, from Sharp's How Brands Grow, is that growth comes overwhelmingly from penetration (reaching more buyers) rather than deepening loyalty, and that the job is to build mental availability, being easily called to mind at the moment of purchase, cued by the category entry points that trigger a need (Romaniuk and Sharp, 2016). The honest limits: these laws are astonishingly robust in stable markets but bend in disrupted ones, and the rival "brand equity" tradition argues that meaning and differentiation are exactly what the behavioural laws leave out (Keller, 1993).

Most consumers are not loyal to one brand; they are poly-loyal, buying from a repertoire of 3 to 5 brands in a category, roughly in proportion to those brands' sizes (see L5-02). So switching is not a crisis to be prevented but the normal texture of a category. The useful distinction, from Dick and Basu, separates attitudinal loyalty (you genuinely prefer and are committed to the brand) from behavioural loyalty (you just keep buying it), which can come apart completely: plenty of repeat purchase is habit, convenience, or lock-in, not devotion (Dick and Basu, 1994). The popular "loyalty as the key to profit" doctrine, and its famous Net Promoter Score, is weaker than advertised, NPS predicts growth no better than ordinary satisfaction measures (Keiningham et al., 2007), and for most brands growth still comes from reaching more buyers, not from deepening loyalty.

Rogers gave us the enduring vocabulary of diffusion: the S-curve of cumulative adoption, the five adopter categories (innovators, early adopters, early majority, late majority, laggards), and the traits of an innovation that speed its spread (Rogers, 2003). Bass turned it into a forecasting model driven by two forces, external influence (advertising and media) and internal social influence (imitation), and it remains the workhorse for predicting a new product's take-up (Bass, 1969). What has changed since is the why. Information now spreads almost instantly, so the bottleneck is access, price, and trust, not awareness. The romantic idea that a few special "influentials" tip the market has not survived scrutiny: large cascades usually come from a critical mass of ordinary, easily-influenced people, not rare tastemakers (Watts and Dodds, 2007), and influence turns out to be spread thinly across a network, not concentrated (Aral and Walker, 2012).

A genuine consumption ritual does two things a brand prizes. It locks in use, because it is repeated behaviour with symbolic resistance to substitution, which makes it stickier than mere preference (the Oreo twist, the Corona lime). And it is the mechanism of meaning transfer: meaning does not sit inside a product, it is pumped from the culture into goods through advertising and then into people's lives through rituals of use, gift-giving, and grooming (McCracken, 1986). That is why a ritual-laden brand carries so much more than its function, and why the meaning is so hard for a rival to copy. The operational catch: you can support and protect a ritual customers already have, but you cannot decree one, and calling any repeated purchase a "ritual" fools only the marketer saying it.

People trust recommendations from other people far more than they trust advertising, which makes word of mouth economically huge and endlessly chased. Two findings do most of the useful work. First, on what gets shared: content that stirs a high-arousal emotion (awe, excitement, anger, anxiety) travels further than content that is merely positive or sad, and things get shared when they give people social currency, are triggered by everyday cues, and carry practical value (Berger and Milkman, 2012). Second, on how sharing spreads: most things that look "viral" are not deep person-to-person cascades at all but broadcasts, one large source reaching many people at once, so sustained self-propagating growth is rare (Goel et al., 2016). The honest conclusion is that virality is an outcome, not a strategy. You cannot reliably manufacture it, but you can build something distinctive and triggerable, measure the spread of talk rather than just its volume (Godes and Mayzlin, 2004), and use referral mechanics where the category rewards them (Schmitt et al., 2011).

The serious study of collective action treats a protest not as spontaneous emotion boiling over but as organised, resource-dependent, framed activity. The pivotal reframe is that grievance is not the scarce ingredient, it is always present, so the real question is what supplies the resources, money, time, networks, and organisational infrastructure, that convert a standing complaint into action (McCarthy and Zald, 1977). Mobilisation also needs an opening in the political situation, an organisational base, and a shared belief that change is actually possible (McAdam, 1982), plus a frame that names an injustice, an "us" who suffers it, and an agent who can fix it (Benford and Snow, 2000). The most-cited modern finding is that nonviolent campaigns have historically succeeded far more often than violent ones (roughly 53% versus 26% across a century of cases), with breadth of participation mattering more than intensity (Stephan and Chenoweth, 2008). The famous "3.5% rule" drawn from that work, that no campaign mobilising 3.5% of a population in active participation has failed, is a striking descriptive pattern, not a formula, and its own authors caution against reading it as a forecast, especially as authoritarian counter-tactics have sharpened.

Health behaviour change is the largest applied laboratory in behavioural science, and its central lesson is humbling: knowing is not doing. Decades of cognitive models mapped the beliefs behind health choices (perceived risk, benefits, barriers, and above all self-efficacy, the belief you can actually do it), and the modern synthesis, COM-B, sums them up as behaviour needing Capability, Opportunity, and Motivation together (Michie, van Stralen and West, 2011). But the hard empirical finding is the intention-behaviour gap: even when you successfully change what people intend to do, behaviour follows only modestly (Webb and Sheeran, 2006; Sheeran and Webb, 2016). The field's clearest success, the long decline in smoking, came not from persuading individuals but from structural, multi-channel action (taxes, smoke-free laws, plain packaging, services), while diet and exercise interventions tend to produce small effects that fade. And the deepest finding of all is that social conditions, income, education, control over one's life, explain more of the variation in health than any psychological programme, so interventions that ignore that gradient can quietly widen it (Marmot, 2015).

The founding insight of behavioural finance is mental accounting: money is not treated as fungible, people partition it into separate mental pots (the windfall you will splurge, the salary you are careful with, the "current income" you spend versus the "future wealth" you protect), and where a sum is filed changes how it is spent (Thaler, 1985). Underneath sits loss aversion and reference dependence: losses loom larger than equal gains, and evaluating too frequently makes people feel those losses too sharply (Kahneman and Tversky, 1979; Benartzi and Thaler, 1995). The applied punchline is blunt and useful: the interventions that reliably change financial behaviour are structural, pre-commitment and defaults, while teaching people about money barely works. Pre-committing future pay rises to saving (Save More Tomorrow) and automatic enrolment lifted saving and participation dramatically (Thaler and Benartzi, 2004; Madrian and Shea, 2001), whereas financial-literacy education explains almost none of the variation in what people actually do (Fernandes, Lynch and Netemeyer, 2014).

Compliance research studies how a direct request turns into a yes, and it produced a toolkit of sequential-request techniques that genuinely work. Foot-in-the-door: get a small yes first, and a larger request is more likely to follow (Freedman and Fraser, 1966). Door-in-the-face: open with a big request that gets refused, then the smaller one you wanted feels like a concession worth matching (Cialdini et al., 1975). Low-ball: secure a commitment, then raise the cost, and people tend to stick (Cialdini et al., 1978). These sit inside Cialdini's broader six principles (see L5-13) and shade into outright obedience to authority (see L4-15). The honest catch, sharpened by the replication crisis, is that the reliable core is narrower than the popular retellings suggest: some effects are robust but smaller than first claimed (Burger, 1999), and many of the colourful one-off field demonstrations replicate weakly. Treat the techniques as hypotheses to test in your own context, not proven levers, and expect a large share of plausible applications to do nothing.

Psychographic or lifestyle segmentation sorts people by values, attitudes, and lifestyle rather than by age and income, and the commercial systems (VALS, PRIZM and their descendants) are familiar to every marketer. The uncomfortable methodological fact is cluster instability: data-driven cluster solutions are highly sensitive to the algorithm, the distance measure, the starting seeds, the sample, and the number of clusters you ask for, so the segments you are shown are usually just one of many statistically equivalent ways to slice the same data (Dolnicar, 2002; Dolnicar, Grün and Leisch, 2018). Two related problems follow: such segments often add little predictive power over plain demographics (Yankelovich and Meer, 2006), and individuals frequently get reassigned to different clusters when you measure again months later, which undercuts the whole idea of a stable "type". The resolution keeps segmentation but separates two very different things: a segmentation used as a communication device to align a team around a few archetypes, and a segmentation used to steer strategic investment, which must be model-based, grounded in real behaviour, and validated for stability and out-of-sample prediction before anyone bets a budget on it.

Robert Cialdini distilled persuasion into six principles, reciprocity, commitment and consistency, social proof, liking, authority, and scarcity, later adding a seventh, unity (Cialdini, 1984; 2016). Each is real: reciprocity rests on a documented social norm (Gouldner, 1960), social proof on the classic conformity studies (Asch, 1956), consistency on dissonance and self-perception (Festinger, 1957; Bem, 1972), scarcity on psychological reactance (Brehm, 1966), authority on the obedience tradition (see L4-15). The catch, and the point of this piece, is that each principle has a predictable failure condition. The sharpest example: social proof backfires when you broadcast an unwanted behaviour as common, a sign saying "many visitors take wood" increased theft in a petrified forest (Cialdini et al., 2006). The honest stance is to treat the six as a hypothesis-generation kit, not a checklist of guaranteed levers: name the failure condition before you deploy a principle, and test it in your own context, because the families of mechanism are real while many of the colourful specific findings replicate weakly.

Thaler and Sunstein's Nudge (2008) launched a movement built on choice architecture: defaults, framing, salience, and norm messages that steer behaviour without banning anything, and governments institutionalised it through behavioural-insights units worldwide. Some early wins were real and large, especially defaults (pension auto-enrolment, organ-donation registration). But the reckoning that began around 2020 has been brutal. A large meta-analysis reported sizeable average nudge effects (Mertens et al., 2022), and a reanalysis found that once you correct for publication bias, the average effect is statistically indistinguishable from zero (Maier et al., 2022). Meanwhile the most credible real-world evidence, 126 trials run by actual government nudge units on 23 million people, found an average effect of about 1.4 percentage points, several times smaller than the effect the published academic literature advertised (DellaVigna and Linos, 2022). The honest synthesis: defaults, form simplification, and timely reminders genuinely work, the low-cost-tweak-as-substitute-for-policy claim does not, and when you plan, you should discount the headline effect sizes several-fold.

COM-B says any behaviour is the product of three things a person needs at once: Capability (the physical and mental capacity to do it), Opportunity (an environment and social world that allow it), and Motivation (wanting to, both reflectively and automatically) (Michie, van Stralen and West, 2011). The applied power is not the model itself but the discipline it forces: diagnose which of the three is the binding constraint before you design anything, because the right fix for a Capability gap is useless for an Opportunity gap. The Behaviour Change Wheel then maps that diagnosis onto nine intervention functions (from education and training to environmental restructuring and restriction), and the Behaviour Change Technique Taxonomy gives a shared vocabulary of 93 techniques so interventions can be described and compared (Michie et al., 2013). The honest caveat: COM-B is integrative, not predictive, it organises what we already know and forces a rigorous diagnosis, rather than supplying new causal laws, and the techniques it points to have moderate, context-dependent effects. Its value is the discipline, and that discipline is rare enough to be worth a great deal.

That framing works, the same facts worded two ways pulling different choices, is settled (Tversky and Kahneman, 1981; the phenomenon is L4-14's territory). The applied questions are the ones that decide whether it helps you: which frame to use, and how much it will actually do. Both have uncomfortable answers. Framing is not one tool but three, and they are not equally reliable (Levin, Schneider and Gaeta, 1998): attribute frames (90% lean vs 10% fat) dependably shift how one thing is judged, risky-choice frames swing with the audience's reference point, and goal frames (act versus fail to act) are the weakest and easiest to counter, so picking the wrong type for the task is a common and costly error. And the effect you can expect in the field is far smaller than the one in the study: framing shrinks dramatically once a competing frame, a distrusted source, prior attitudes, or real stakes enter the room (Druckman, 2001; Chong and Druckman, 2007). So the applied job is to match the frame type to the task and test it against the frames it will actually meet, treating any single-frame lab number as a ceiling, not a forecast.

A default is what happens when a person does nothing, and it moves behaviour with a force that dwarfs most persuasion. The starkest case: countries where organ donation is opt-out register roughly 85 to 99% of citizens as donors, while otherwise-similar opt-in countries manage around 4 to 28%, a gap that is essentially the entire policy, produced by a single choice about the default (Johnson and Goldstein, 2003). Automatic enrolment did the same for retirement saving, lifting participation from about 37% to 86% and sticking for years (Madrian and Shea, 2001). Defaults work because several forces pull the same way at once: the effort of switching, the way the default becomes the reference point you would feel a loss to leave, the sense that someone set it for a reason, and plain inattention. The applied catch is that a default is never neutral: it is an editorial choice by whoever designs the form, so the only honest question is whether the default is the option that genuinely serves the typical person, or one that serves the person who set it.

Persuasion has always been "message x source x audience", and the source is often the part that decides whether a message lands. The classic finding is that a source's power comes from credibility, which is expertise and trustworthiness together (Hovland and Weiss, 1951): you need both, and a source high on one and low on the other underperforms. The applied twist that matters most today is that credibility is conferred by the audience, not by credentials. An expert speaking to a community that distrusts their institution can carry negative credibility, and for contested or polarised topics a messenger who shares the audience's identity routinely beats a better-qualified outsider (Berinsky, 2017). Source cues also work hardest on disengaged audiences, who lean on "who said it" as a shortcut, and least on motivated ones who scrutinise the argument itself (the elaboration-likelihood logic of L1-06). And the reflex move, hiring a celebrity, is an unreliable bet: endorsement effects are heterogeneous and often null unless there is genuine fit (Knoll and Matthes, 2017). So choose the messenger for the specific audience, and test the pairing.

That we copy others is settled (the phenomenon and the norm concept live in L4-03 and L5-12). Two applied facts are the ones that change how you use it. First, social proof is path-dependent: in the MusicLab experiment, simply showing people how many others had downloaded a song made hits bigger and made which song became a hit far less predictable, so success fed on itself rather than tracking quality (Salganik, Dodds and Watts, 2006). Popularity is momentum, not merit, which means your own bestseller or viral hit may be a cascade, not proof that it is good. Second, descriptive-norm messages (what others do) cut both ways: telling people they use less energy than average made the already-frugal drift up toward the mean, a boomerang that only disappeared when an injunctive signal (approval, a smiley face) was added (Schultz et al., 2007). The applied rules follow: show a favourable, specific norm, never advertise the unwanted behaviour as common (the petrified-forest error of L5-12), pair descriptive with injunctive to stop the boomerang, and read your own popularity metrics as momentum rather than merit.

Here is the uncomfortable fact this whole layer has been circling: the apparatus of a helpful nudge and the apparatus of manipulation is the same apparatus. Defaults, framing, salience, friction, social proof, the messenger you pick. So "I'm just presenting the options" is never an exemption, because there is no neutral way to present them (Thaler and Sunstein, 2003). What separates assistance from manipulation is therefore never the technique. It is three things you can actually test. Transparency: would it still work if your audience saw exactly what you were doing and why (Bovens, 2009)? Welfare-alignment: does it serve the chooser's goals, by their lights, or only yours? Legitimacy: do you have standing to steer here at all? Run those three on any intervention before it goes out. If it survives being seen, serves the person you are steering, and you have the right to steer, it sits on the assistance side. If it only works because the audience cannot see it, it is manipulation wearing a nudge's clothes.

The Theory of Planned Behaviour (TPB) says a behaviour is driven by the intention to do it, and that intention is built from three things: your attitude to the behaviour, the social pressure you feel around it, and how much control you believe you have (Ajzen, 1991). Decades of studies confirm it predicts intention well and behaviour moderately (Armitage and Conner, 2001). Here is the applied catch, and the reason this matters for real work: intention is a strong predictor but a weak lever. Experiments that successfully shifted people's intentions produced only small changes in their actual behaviour (Webb and Sheeran, 2006), the famous intention-behaviour gap. So the useful move is to stop treating TPB as a way to manufacture action and start using it as a diagnostic: measure the three components in your specific audience, find which one is actually holding the behaviour back, aim your effort there, and then cross the gap with concrete plans (Gollwitzer, 1999) and by removing the real-world barriers the model only gestures at.

BJ Fogg's model says a behaviour happens only when three things line up in the same moment: Motivation to do it, Ability to do it easily, and a Prompt that triggers it right then (Fogg, 2009). The formula B = MAP is clean and genuinely useful, but its practical value is not the equation, it is two counter-intuitive moves it points you toward. First, the prompt is the factor everyone underrates: no cue at the right moment and nothing happens, no matter how motivated someone is, so most "they just aren't doing it" problems are missing or mistimed prompts, not weak desire. Second, when a behaviour stalls, make it tiny rather than trying to pump up motivation, because ability is something you can design and motivation is volatile and fades. The honest caveat matters just as much: B = MAP is a design heuristic with little independent testing behind it, not a validated causal theory, so use it to diagnose and design, and reach for the more rigorously developed COM-B (covered in L6-03) when a decision needs to stand on evidence.

COM-B says any behaviour needs three things present at once: Capability, Opportunity, and Motivation (Michie, van Stralen and West, 2011). The applied lens on how to use that to diagnose a problem lives in L5-14. What matters here, at the framework level, is something more useful than one more model to memorise: COM-B was not invented so much as distilled, built by systematically reviewing nineteen existing behaviour-change frameworks and boiling them into one. That origin is the whole point. It makes COM-B a master map on which every other tool in this series has a place. The Theory of Planned Behaviour sits in reflective motivation, Fogg's model is a coarser three-way cut of the same territory, nudges and defaults act on physical opportunity, habit lives in automatic motivation, social proof in social opportunity. When you do not know which framework to reach for, COM-B tells you where on the map your problem sits. The honest catch, and the reason not to over-trust it: it is a framework, not a theory. It tells you where to look, not why a given intervention will work or how large the effect will be (Ogden, 2016), and its wide institutional adoption is not the same thing as being proven right.

Self-Determination Theory (SDT) says people have three basic psychological needs: autonomy (acting from your own values), competence (feeling effective), and relatedness (feeling connected). Satisfy them and you get self-driven, durable motivation and wellbeing; thwart them and you get the brittle, do-it-because-I-have-to kind (Ryan and Deci, 2000). The applied lesson that matters, and that most incentive schemes get wrong, is that the quality of motivation beats the quantity. Motivation that people own outlasts and outperforms motivation you impose. That is why bolting a controlling reward onto something people already enjoy can quietly poison it, the overjustification effect Deci first showed in 1971: pay people for what they were doing for free and they often stop doing it for free. The honest complication is that this undermining effect is real but bounded and hotly argued (Deci, Koestner and Ryan, 1999; Cameron and Pierce, 1994), so the popular slogan "rewards always backfire" overstates a subtler truth. The practical move is to build motivation people internalise, by supporting the three needs and giving the reason why, rather than buying compliance you have to keep paying for.

Regulatory Focus Theory splits motivation into two systems rather than one. A promotion focus chases gains, growth, and ideals, and moves by eager, go-for-it strategies; a prevention focus guards against losses, seeks security, and moves by careful, vigilant ones (Higgins, 1997). The applied payoff is a phenomenon called regulatory fit: a message feels right and persuades more when its strategy matches the audience's focus (Higgins, 2000). This is the missing piece behind one of communication's oldest puzzles, why gain-framed and loss-framed messages keep trading wins with no stable winner. Gain-framed, aspirational messages land with promotion-focused people and in growth categories, while security-framed, loss-averse messages land with prevention-focused people and in categories like insurance and pensions. So the practical question is never "is gain-framing better?" It is "which focus is my audience in, and does my message fit it?" The honest caveat is that fit is a real but moderate and variable effect, not a guaranteed multiplier, so it is a diagnosis to test rather than a rule to trust.

Self-Discrepancy Theory says we carry three versions of ourselves: the actual self we think we have, the ideal self we hope to become, and the ought self we believe we are obliged to be, usually by others' standards (Higgins, 1987). Its sharpest, most useful claim is that the two kinds of gap feel different in a predictable way. Falling short of your ideal produces dejection, the sadness and disappointment of a hope unmet; falling short of your ought produces agitation, the anxiety and guilt of a duty unfulfilled. For a communicator that runs in reverse into a diagnostic: if your message leaves people anxious or guilty, you are appealing to an ought (obligation), and if it leaves them wistful or dissatisfied, you are appealing to an ideal (aspiration). The honest limit is that the broad link between self-gaps and emotional distress is well established, but the neat one-to-one emotion mapping is weaker and more conditional than the theory first claimed (Phillips and Silvia, 2005), so it is a lens to diagnose the register of your communication, not a formula that guarantees a feeling.

Identity-Based Motivation says the identity that is active in the moment shapes what feels natural to do, and that identities are cued by the situation rather than fixed (Oyserman, 2009). Its most useful and least obvious claim is about how people read difficulty. When a task feels like "what people like me do," hitting a hard patch is taken as a sign that the task matters and is worth pushing through; when the same task feels like it belongs to some other kind of person, the identical difficulty is taken as proof that it is not for me and a reason to quit. So difficulty is not just a barrier, its meaning flips with identity. For a communicator that yields three moves: cue the relevant identity right where the action happens, make the behaviour feel like "what people like us do," and pre-frame the effort as a sign of belonging rather than of being in the wrong place. The honest limit is that most of the strong evidence comes from adolescent-education interventions, and that cueing an identity can shift how a moment feels but cannot on its own remove the structural barriers that make some identities hard to claim in the first place.

Dual-process theories say thinking comes in two broad kinds: a fast, automatic, low-effort mode (Type 1) and a slow, effortful, working-memory-hungry mode (Type 2), with the fast mode usually producing a default that the slow mode may or may not check (Stanovich and West, 2000; Evans, 2008). Popularised as "System 1 and System 2" (Kahneman, 2011), it became a household model. The honest, and less known, development is that the field has largely retreated from the crude version of it. The "two systems" turn out to be defined by lists of features (fast, emotional, automatic, and so on) that do not reliably travel together, so treating them as two little agents in the head is a convenient fiction rather than a fact of neural architecture, a point pressed even by the theory's own proponents (Melnikoff and Bargh, 2018; Pennycook et al., 2018). What survives is genuinely useful and much narrower: a real distinction between thinking that loads working memory and thinking that does not (Evans and Stanovich, 2013), plus the correction that fast, intuitive answers are often correct rather than errors waiting to be fixed. So keep the effort distinction as a design lens, and drop the story about a wise reasoner fighting a foolish gut.

Prospect Theory (Kahneman and Tversky, 1979) describes how people actually make risky choices, and it does so with a precision most behavioural frameworks never reach: a Nobel prize, a formal value function, decades of replication. It rests on three ideas. First, people judge outcomes as gains and losses against a reference point rather than as final amounts of money, so the same result can feel like a win or a loss depending on where you start counting. Second, they weight probabilities non-linearly, overweighting small chances and underweighting moderate-to-large ones, which is why the same person buys both lottery tickets and insurance. Third, losses are felt more sharply than equivalent gains, the famous "loss aversion." Here is the part the popular version leaves out: that third idea, the one everyone borrows, is exactly the one now under serious scientific challenge (Gal and Rucker, 2018; Mrkva et al., 2020). Loss aversion is real but contingent and smaller than the "losses loom twice as large" slogan suggests, and its supposed knock-on effects are contested. The reference point and the probability weighting, by contrast, are sturdy. So the practical move is to set the reference point deliberately, design for how people misweight probabilities, and treat loss framing as a sometimes-useful tactic rather than a guaranteed multiplier.

Construal Level Theory says the mind represents the same thing at different levels of abstraction depending on how psychologically distant it feels (Trope and Liberman, 2010). Distance comes in four flavours, time, physical space, social closeness, and how hypothetical something is, and they all pull the same way. When something is far off, we think about it abstractly: the big picture, the point of it, whether it is a good idea in principle. When it is near, we think concretely: the mechanics, the steps, whether we can actually pull it off. That single shift explains a lot of familiar behaviour. A distant plan is judged on how desirable it is and looks wonderful; the same plan, once it is upon us, is judged on how feasible it is and suddenly looks daunting (Liberman and Trope, 1998). The practical upshot is two levers. First, match your message to the mode the audience is in: sell the why to people thinking about something far away, and the how to people about to act. Second, expect the desirability-to-feasibility flip and plan for it rather than being surprised by it. One honest caveat carries through everything below: the direction of these effects is well supported, the size less so. Pooled across hundreds of studies the core effect is reliable but only medium-sized (Soderberg et al., 2015), and when a high-powered team re-ran one of the theory's signature experiments, it did not replicate (Calderon et al., 2020). So treat this as a reliable guide to which way to lean, not a dial you can set to a precise number.

Self-Categorisation Theory explains how a person shifts between thinking as an individual and thinking as a group member (Turner et al., 1987). The self runs at different levels: a personal identity, "me, unlike other people," and social identities, "us, a group, unlike other groups." Which level is active is not fixed; it flips with the situation. When a social identity switches on, something specific happens: you start to see yourself and other members as more or less interchangeable examples of the group, you adopt the group's norms, you favour fellow members, and you act on the group's behalf. This is why the same person is measured and moderate at their desk and roars with the crowd at a match. The practical consequence is large. If behaviour follows the norms of whichever identity is currently salient, then the lever that moves people is often not a better individual argument but a shift in which "we" feels active, after which the group's own norm does the persuading. Two honest cautions run through everything below. First, you have to get the category right, because the same message means different things depending on which identity is switched on, and appeals to a "we" that is not actually salient fall flat. Second, because the theory makes salience a product of what is accessible and what fits the immediate situation, it is far better at explaining which identity was active after the fact than at predicting which one a messy real situation will switch on in advance. So use it to design and activate a shared identity deliberately, and then test which one actually fires rather than assuming.

The Social Identity Approach takes the idea that behaviour flows from a shared "us" (the cognitive side of that is in the companion piece on self-categorisation) and turns it into a practical craft for leadership, influence, and even health. Its central claim is that leadership is a group process: people follow leaders who represent and build a sense of shared identity, not simply leaders who are charismatic or who offer the best deal (Haslam, Reicher and Platow, 2011). That craft has been broken into four moves, the "four R's": be seen as one of us, be seen to act for us rather than for yourself, actively craft a sense of who "us" is, and make that identity matter by building it into real events and structures. This is not hand-waving. The four moves have been turned into a measured instrument, the Identity Leadership Inventory, which predicts trust, commitment and satisfaction beyond what older leadership models explain, and holds up across twenty countries and fourteen languages (Steffens et al., 2014; van Dick et al., 2018). A related strand, the "social cure," shows that belonging to groups is a genuine resource for health and wellbeing (Haslam, Jetten et al., 2018). The catch, and the reason this article leads with ethics, is that all of this is content-neutral machinery. The same four moves build an inclusive, high-trust team or a demagogue's crowd. In a study across four nations, seeing a national leader as an identity leader made citizens follow COVID health measures more under one leader and less under another, because the "us" each leader built pointed in opposite directions (Frenzel et al., 2022). So the real question is never just how to do the four R's. It is what you are making "us" mean, and who that leaves out.

Intergroup Emotions Theory says that when you identify with a group and that identity is active, the group becomes part of your sense of self, so you feel emotions about things that happen to the group even when they do not touch you personally (Smith, 1993; Mackie, Devos and Smith, 2000). These group-based emotions are not just private moods. They are distinct from your individual emotions, they get stronger the more you identify, and each specific emotion pushes toward a specific action. Group-based anger pushes you to confront and move against an out-group, fear pushes you to withdraw, contempt pushes you to exclude, and group-based guilt pushes you to make amends (Doosje et al., 1998; Smith, Seger and Mackie, 2007). That single link, particular emotion to particular action, is what makes the theory so useful and so double-edged. Useful, because if you want a group to act you have to evoke the emotion that actually drives that action, and a fear appeal when you need people to mobilise, or an anger appeal when they feel powerless, simply misfires. Dangerous, because the same machinery that turns anger at an injustice into a collective-action movement (van Zomeren et al., 2004) turns manufactured contempt for an out-group into exclusion and dehumanisation. This is the framework that most demands you lead with ethics, because the honest dividing line is often the emotion itself: anger at an injustice can still grant the other side their humanity, and contempt cannot.

Pierre Bourdieu's framework rests on three linked ideas and one hard claim. The habitus is the durable set of dispositions, tastes, instincts, ways of speaking and judging, that you absorb from growing up in a particular social position, and that then generates how you act without any conscious calculation (Bourdieu, 1977). A field is a structured arena, art, politics, finance, academia, each with its own stakes, rules of the game, and pecking order. Capital is the resources that place you within a field, and Bourdieu's key move was to see it in several forms, not one: economic (money), cultural (know-how, taste, and credentials), social (who you know), and symbolic (recognised prestige), a set he built up across his work and set out most directly, for the first three, in his essay on capital (Bourdieu, 1986). The hard claim, demonstrated in his landmark study of French taste, is that these fit together into a homology: aesthetic preferences are systematically structured by the volume and type of capital people hold, so "taste" is really class position made to feel like personal choice (Bourdieu, 1984). The framework's lasting achievement is dissolving an old dichotomy, showing that practice is neither free will nor pure determination but disposition meeting situation. For a practitioner, the payoff is a reliable way to read taste as a map of social position, to see that different fields reward different currencies, and to treat cultural capital as the real, tradeable asset it is. The honest caveat, and this is a genuinely contested framework, is that it leans hard toward social reproduction and can underplay agency (King, 2000), that people carry plural and even contradictory dispositions rather than one tidy habitus (Lahire, 2003), and that the specific map was drawn in 1960s and 1970s France and has loosened since (Bennett et al., 2009). Treat it as one of the most powerful lenses in social science, not as an iron law.

Practice Theory makes one radical move: it shifts the unit of analysis away from the individual, with their attitudes, intentions and values, and onto the practice, a routinised, recognisable way of doing something like commuting, showering or eating dinner (Reckwitz, 2002; Warde, 2005). Practices exist beyond any single person; they recruit people into them and carry on when those people leave, so a practitioner is better understood as a carrier of a practice than as the free author of a choice. Each practice is held together by three kinds of element: materials (the objects, technologies and infrastructures involved, and the body itself), competences (the skills and know-how it takes), and meanings (what the activity signifies and why it matters). A practice lasts only while those elements keep being linked together in performance, which is why practices emerge, spread, mutate and die as their elements and links change (Shove, Pantzar and Watson, 2012). The practical consequence is large and counterintuitive: if you want to change what people do, the lever is usually not changing minds one at a time but reconfiguring the practice, its materials, its competences, its meanings. This is the sharpest available challenge to attitude-change and nudge approaches, captured in Shove's (2010) argument to move "beyond the ABC" of attitude, behaviour and choice. The honest caveat is that the framework is a powerful corrective and diagnostic but is deliberately silent on individual psychology and better at explaining a practice after the fact than at predicting exactly which lever will shift it, so it works best alongside the individual-level frameworks rather than as a replacement for them.

Consumer Culture Theory is not a single theory but an umbrella for a whole tradition of interpretive, qualitative research that treats consumption as a cultural act rather than a rational calculation (Arnould and Thompson, 2005). Where mainstream models ask what features and benefits drive a purchase, CCT asks what the consumption means and what the person is doing with it, and it draws on anthropology, sociology and semiotics to answer. Four big ideas anchor it. Cultural meaning moves through a society, travelling from the wider culture into goods and then into the individual who uses them, carried by advertising, fashion and everyday rituals (McCracken, 1986). Our possessions become part of who we are, an "extended self" rather than mere property (Belk, 1988). Consumers gather into genuine communities around brands, with shared rituals and a sense of mutual obligation (Muñiz and O'Guinn, 2001). And people do not passively absorb the meanings marketers send; they rework, resist and co-produce them (Holt, 2002). For a practitioner, the payoff is a different and often sharper question, "what does this mean?", and a research toolkit, ethnography, netnography and semiotics, built to answer it. The honest caveat, and it matters, is that CCT is an interpretive paradigm rather than a predictive theory. It is superb at telling you what is going on and why it matters, and weak at telling you how big the effect is or which lever to pull, its methodological looseness is a real risk, and its early enthusiasm for consumer agency tended to underplay the structural forces bearing down on people (Askegaard and Linnet, 2011). It works best paired with quantitative work, not as a substitute for it.

Cultural Consensus Theory is a formal statistical model, built in cognitive anthropology, that recovers shared cultural knowledge from people's answers to a set of questions (Romney, Weller and Batchelder, 1986). Its cleverness is that it does two things at once, and needs no answer key to do them: from the pattern of who agrees with whom, it estimates both the culturally "correct" or shared answer to each question and how much each person knows that shared culture, their "competence". The logic underneath is simple. If two people agree far more than chance would predict, they must be drawing on some shared signal, a common culture, so the agreement pattern alone can reveal what that shared knowledge is and who holds it most fully. That lets you do something researchers usually just assume they can skip: actually test whether a shared view exists, measure how strong it is, and see who holds it and who diverges. For a practitioner the uses are concrete. You can test whether a market segment genuinely shares a mental model or merely shares demographics, find the real experts in a room, spot hidden sub-groups, and, when agreement is high, get a reliable read from surprisingly few people. The honest limits are worth stating plainly: it is descriptive, not explanatory, so it tells you what is shared but not how it got that way or how it is changing; the classic version assumes a single shared culture, and you need its later multi-group extensions if your sample actually contains sub-cultures (Anders and Batchelder, 2012); and it measures that something is shared and how strongly, never what the shared thing means, which is a different job entirely.

Bronfenbrenner's ecological systems theory places the developing person inside a set of nested contexts, and its lasting value is as a corrective to explaining behaviour by the individual alone (Bronfenbrenner, 1979). The layers run from the microsystem, the face-to-face settings a person is actually in, like family, school and work; to the mesosystem, the connections between those settings; to the exosystem, settings the person never enters but which still shape them, such as a parent's workplace or a local policy; to the macrosystem, the surrounding culture, economy and law; and, added later, the chronosystem, change over time, both a person's own life transitions and the historical moment (Bronfenbrenner, 1986). That much is the famous diagram. What far fewer people know is that Bronfenbrenner spent his later career revising it into the "bioecological" model, which puts proximal processes, the sustained, close-up, back-and-forth interactions in the microsystem such as a parent playing with a child or a mentor working with a protege, at the centre as the actual engine of development, organised as Process-Person-Context-Time (Bronfenbrenner and Morris, 2006). The distal circles matter, on this mature view, precisely because they feed or starve those close-up processes. And this is the catch worth carrying: a review of studies claiming to use the theory found that almost all of them used only the old concentric-circles version and ignored the revision, sometimes citing the wrong source (Tudge et al., 2009). So the honest way to use Bronfenbrenner is to keep the map of nested contexts, but never forget the motor at its centre, and to treat the whole thing as a rich organising frame rather than a precise predictive theory.

The Transtheoretical Model, universally known as the "Stages of Change", says that people change through a sequence of stages: precontemplation (not even thinking about it), contemplation (considering it), preparation (planning to act soon), action (doing it), and maintenance (keeping it up), with relapse a normal part of the loop (Prochaska, DiClemente and Norcross, 1992). Its central instruction is to match your intervention to the person's stage rather than pushing everyone straight to action. The model has one genuinely valuable insight at its heart, which is that "not ready to change" is a real and distinct state, so treating someone who has not even begun to consider change as if they were ready to act simply fails, and that reframing did real good in a field that used to assume everyone was ready. But here is the part that matters most: the specific claims that turn that insight into a "model" have not held up under scrutiny, and this is one of the most heavily criticised frameworks in behaviour change. The five stages are not real, distinct states; they are arbitrary lines drawn across what is really a continuous variable, so that someone planning to change in thirty days counts as "preparation" while thirty-one days makes them merely "contemplation" (Sutton, 2001). People do not move through the stages in tidy order. And the model's central promise, that matching your intervention to someone's stage works better than not, has weak and mixed-to-negative evidence behind it (West, 2005; Riemsma et al., 2003), to the point that one prominent critic called for putting the model "to rest" (West, 2005). So the honest way to use it is to keep the readiness insight and the shared vocabulary as a rough, useful heuristic, and to stop treating the stages as though they were real or stage-matching as though it were a proven lever.

Social Cognitive Theory frames behaviour as the joint product of three things that continuously influence one another: the person (their thoughts, beliefs and feelings), their behaviour, and their environment. Bandura called this reciprocal determinism (Bandura, 1986). Two ideas inside it did most of the work in the world. The first is observational learning: we pick up behaviour by watching models, shown vividly in the Bobo doll studies where children copied an adult's aggression (Bandura, Ross and Ross, 1961). The second, and the most influential single construct, is self-efficacy, the belief in your own capability to carry out a specific action, which Bandura argued is built from four sources: mastering the task yourself, watching people like you succeed, being credibly persuaded, and reading your own body's signals of calm or panic (Bandura, 1977). Across many studies, people with higher self-efficacy do perform better, with a meta-analysis putting the average correlation with work performance at around .38 (Stajković and Luthans, 1998). Here is the honest complication, and it is worth knowing before you build anything on it. When you track the same person over time rather than comparing different people, the story flips: self-efficacy looks less like the engine of performance and more like a readout of how you have done so far. A within-person meta-analysis found the self-efficacy to later-performance link shrinks close to zero once you account for past performance, while past performance strongly predicts your later confidence (Sitzmann and Yeo, 2013), and some studies find high confidence can even slightly hurt (Vancouver, Thompson and Williams, 2001). So the practical reading is: self-efficacy is real and worth building, but you build it mostly by engineering genuine success and credible models, not by pep talks, and you should not assume that pumping up someone's stated confidence, on its own, will drive their results.

The life-course perspective, developed by the sociologist Glen Elder, says we should read a life as a trajectory embedded in historical time and social structure: when you were born, the institutions you passed through, and the lives you are linked to all shape your path, and people exercise agency within those constraints rather than outside them (Elder, 1994). Bolted onto that is a more mechanical claim, the developmental cascade: competence or trouble in one area spills into others over time, so that effects accumulate rather than fade, and a small early difference can snowball (Masten and Cicchetti, 2010). Economists tell a compatible story about skills, where early investment raises the payoff to all later investment, so skills beget skills (Cunha and Heckman, 2007). The core insight is real and matters enormously: development compounds, early conditions echo, and inequality can widen through ordinary developmental processes rather than through any single dramatic event. But three honest cautions have to travel with it. First, proving a genuine cascade is hard, because a long-running correlation between early and late outcomes can just reflect a stable environment rather than one stage causing the next, so many cascade claims are softer than they sound. Second, the popular policy version, the steep curve on which returns to investment fall sharply with age so that early is everything and later is nearly hopeless, does not hold up well: a study of a large set of real programmes found no support for early-life programmes reliably having the biggest returns (Rea and Burton, 2019). Third, early gains frequently fade unless something keeps feeding them, so a one-off early boost is not a lasting fix (Bailey et al., 2016). So the usable version is: think in trajectories and compounding, invest early by all means, but budget for the environment that sustains the gain, and never read "early matters" as "the die is cast."

The Means-End Chain, set out by Jonathan Gutman (1982), says a product is a means to an end: a concrete attribute produces a consequence, and that consequence serves a value. A low-fat yoghurt (attribute) means fewer calories (a functional consequence), which means I can keep my weight in check (a psychosocial consequence), which serves self-respect and health (values). The practical power is that it takes you from a feature nobody cares about on its own to the deep end-state that actually motivates, which is gold for positioning and creative work. Laddering is the interview technique that gets you there, developed by Reynolds and Gutman (1988): you ask "why is that important to you?" over and over, climbing from attributes to values, then aggregate the answers into a hierarchical value map showing the main routes from features to values across a group. Here is the honest part, and it matters as much as the method. The clean ladder is partly a product of the technique rather than a pure readout of someone's mind: pushing "why?" repeatedly can march a person up a hierarchy they would never have built on their own, and the tidy tree you get out is partly enforced by how you asked (Grunert and Grunert, 1995). It is also a deliberative, "cool" method that invites people to rationalise, and it is largely silent on the fast, emotional, automatic side of choice. So use laddering as one of the best structured ways to generate hypotheses about the values your product serves and the language to talk about them, but treat the map as a hypothesis to test against real behaviour, not as proof of what is in the consumer's head.

Jobs-to-be-Done says a customer does not buy a product for its own sake; they hire it to make progress on a job, the functional, emotional and social outcome they are trying to reach in a particular situation, and your true competitive set is everything else that could be hired for the same job, often from categories you never considered. The canonical story is the milkshake that turned out to be hired by commuters as a filling, one-handed companion for a dull morning drive, competing less with other milkshakes than with bagels, bananas and boredom (Christensen, Hall, Dillon and Duncan, 2016). Anthony Ulwick's more measured variant, Outcome-Driven Innovation, turns jobs into "desired outcome" statements that customers rate on importance and satisfaction, so you can spot outcomes that matter a lot and are served badly (Ulwick, 2005). Used as a diagnostic, this is genuinely useful: it forces situational, goal-level thinking instead of demographic or feature thinking, and it reframes who you are really competing with. Here is the honest part, and it matters as much as the idea. Jobs-to-be-Done is largely a practitioner framework with very little independent, peer-reviewed validation; most of its evidence is consulting case studies, and academic attempts to pin it down note that its scientific foundations are thin (Lucassen et al., 2018). It has a circularity problem, since the "job" is usually inferred from the very purchase it is meant to explain, and a definitional one, since a "job" is sometimes a goal, sometimes an outcome, sometimes a context. And it travels closely with Christensen's disruption theory, which fared badly when its own case examples were checked (King and Baatartogtokh, 2015). So use it as a strong prompt for reframing and generating hypotheses, not as a validated engine that predicts what will work, and always test the job you think you have found against what people actually do and pay for.

The Ehrenberg-Bass tradition, popularised by Byron Sharp's How Brands Grow (2010), argues that brands grow chiefly through two things: mental availability, the probability that a brand comes to mind in a buying situation, and physical availability, how easy it is to actually find and buy, rather than through loyalty, differentiation or a deep emotional bond. Growth comes mostly from increasing penetration, reaching more buyers, especially the many light buyers, not from squeezing more loyalty out of existing ones. Mental availability is made concrete through category entry points, the cues people use to call brands to mind: the occasions, needs, locations, moods and motives that trigger a category, so a brand grows by being linked to more of those cues, more strongly, among more people. What gives this weight is the empirical record. The double jeopardy law, that smaller brands have both fewer buyers and slightly less loyal ones, and the Dirichlet model, which reproduces a whole category's buying patterns from little more than market shares, replicate across categories and decades and are among the most robust regularities in marketing science (Ehrenberg, Goodhardt and Barwise, 1990; Ehrenberg, Uncles and Goodhardt, 2004). That robustness makes the framework a bracing corrective to the unfalsifiable "purpose" and "emotional bond" stories the industry runs on. The honest limit is scope: the laws are close to iron in mature, frequently-bought packaged-goods markets, and they show systematic, well-catalogued exceptions elsewhere (Scriven, Bound and Graham, 2017), while luxury, identity-loaded, business-to-business and subscription or platform categories, where loyalty and switching costs genuinely bite, are exactly where the tradition is weakest. So take the core seriously as a default, grow by reach and availability, track penetration and category-entry-point coverage rather than love, but do not cargo-cult "penetration only" into markets built differently.

Decision-journey models say a buyer or voter moves through identifiable stages, and the best-known version, McKinsey's consumer decision journey, sets out four: an initial consideration set of brands you start with, an active evaluation phase where you add and drop options as you research, the moment of purchase, and the post-purchase experience, with a "loyalty loop" that lets satisfied repeat buyers skip straight back to buying (Court et al., 2009). Later work extended this to the tangle of digital touchpoints, search, social, reviews, that a modern journey runs through (Edelman, 2010), and the most rigorous academic synthesis pulls the customer-experience and journey research together across the pre-purchase, purchase and post-purchase stages (Lemon and Verhoef, 2016). Used where it fits, the journey lens is genuinely valuable: it forces you to think touchpoint by touchpoint, aligns your measurement and creative across the whole path, and surfaces influence points people forget, like peer reviews and what happens after the sale. The honest catch, and it is a big one, is that the elaborate journey often describes nothing real. For the many low-involvement, frequently-bought categories, people do not run an "active evaluation"; they buy what comes to mind, triggered by a cue from memory (Romaniuk and Sharp, 2016), so the journey the marketer maps is mostly imaginary. The reconciliation is simple and practical: genuine multi-stage journeys are real for high-involvement, considered, expensive or rare purchases, and for business-to-business and complex services, while low-involvement habitual buying is better understood through mental availability. So the first move is always to diagnose whether the category actually has a journey before you spend a penny mapping one. This also quietly settles one of marketing's oldest arguments, whether B2B marketing is fundamentally different from B2C: mostly it is not, because what decides the approach is not who the buyer is but what is being bought. A big, complex, high-stakes purchase is a journey whether it is a household choosing a house or a firm choosing an enterprise system, and a quick, cheap, repeated purchase is grabbed from memory whether it is a shopper's chocolate bar or a procurement team's usual printer paper. The real axis is journey versus no-journey, and it runs straight across the business-versus-consumer line.

VUCA stands for volatility, uncertainty, complexity and ambiguity, and it is a vocabulary for describing a turbulent environment rather than a theory that explains or predicts anything. The four words point at genuinely different conditions: volatility is fast, large change that you may understand perfectly well but cannot smooth out; uncertainty is not knowing what is coming or whether an event even matters; complexity is having many interconnected moving parts; and ambiguity is not understanding cause and effect at all, even when you have the information in front of you. The acronym is usually traced to US Army War College leadership doctrine in the late 1980s, though its exact origin is genuinely murky; it moved into corporate strategy in the 2000s and became inescapable in leadership writing through the 2010s, spawning a prescriptive cousin, VUCA-Prime (vision, understanding, clarity, agility), from Bob Johansen. The one genuinely useful idea, and it is easy to miss, is that because the four conditions are different, they call for different responses: the most rigorous treatment argues you meet volatility by building in slack and buffers, uncertainty by investing in information and sensing, complexity by restructuring and bringing in the right expertise, and ambiguity by experimenting to learn what actually causes what (Bennett and Lemoine, 2014). The honest catch is that VUCA is not an empirical theory, has no predictive content, and has been stretched to cover any modern difficulty at all, so that it frequently means nothing more precise than "the world is hard now"; if everything is VUCA, then nothing is. It also stays completely silent on who actually bears the volatility, treating a turbulent world as a single condition everyone shares rather than one whose shocks land very unevenly. So use it as a way to open a conversation and force people to say which of the four they actually mean, never as a diagnosis that ends one.

BANI stands for brittle, anxious, nonlinear and incomprehensible, and it was proposed by the futurist Jamais Cascio in 2020 as a successor to VUCA for a world of pandemic, climate breakdown and fragile infrastructure (Cascio, 2020). Brittleness is systems that look solid but shatter all at once under stress instead of bending; anxiety is the pervasive low-grade dread of living through cascading crises; nonlinearity is the disconnection of cause and effect in scale and timing, where a small input produces a huge output or a long delay hides the consequence; and incomprehensibility is the sense that events now exceed our capacity to make sense of them even when we have the data. It resonates because it names the emotional weather that VUCA ignored and because "brittle" captures something true about hyper-optimised, tightly-coupled modern systems. Here is the honest status: BANI has no empirical research programme, no measurement, and no way to tell a genuinely brittle system from a merely complex one, so it is a vocabulary and, more bluntly, a brand, not a theory. Its real value is that each of its four words points at a serious body of research you can actually use: brittleness at the science of resilience and antifragility (Taleb, 2012), nonlinearity at complex-systems and tipping-point research with its early-warning signals (Scheffer et al., 2009), anxiety at the real epidemiology of anxiety disorders, which did surge globally in 2020 (COVID-19 Mental Disorders Collaborators, 2021), and incomprehensibility at the old, well-founded idea of bounded rationality, the limits of human information and computation (Simon, 1955). So use BANI to start a conversation and to feel the shape of the moment, then translate each letter into its grounded question and go to the real literature for the actual tools, rather than treating the acronym itself as knowledge.