The Bias Archipelago: Why the Mind’s Errors Are Not Islands

The availability heuristic estimates frequency or probability from how easily relevant examples come to mind. Retrieval ease can be informative, but vividness, recency and rehearsal can make memorable cases unrepresentative.

Example: After several reports about dengue fill the Calcutta news, a family judges it more likely than a quieter but statistically common illness.

Recipe: associative-retrieval · salience-weighting · attribute-substitution

Evidence: Broadly supported as a family of judgements based on retrieval, although ease, amount and content of recall can have different effects.

Also called: availability bias

The recency effect gives the latest items or experiences disproportionate accessibility in memory and judgement. It is a robust part of serial-position memory, but its size and interpretation depend on delay, task and retrieval conditions.

Example: A manager lets one bad week dominate an appraisal despite eleven months of solid work.

Recipe: associative-retrieval · finite-attention · reconstructive-memory

Evidence: Well established in list memory; extending the label to every overweighting of recent real-world evidence requires care.

Also called: recency bias

The primacy effect gives early items or first impressions greater influence on recall or later judgement. Extra rehearsal, interpretation of later evidence through an initial frame and task order can each contribute.

Example: The first confident speaker in a hiring panel sets a favourable frame that later reservations struggle to dislodge.

Recipe: finite-attention · associative-retrieval · prior-belief-preservation

Evidence: Serial-position primacy is well established, while first-impression effects vary with task structure and opportunities to revise.

Also called: first-impression advantage

The vividness effect occurs when concrete, emotionally engaging or imageable information exerts more influence than pallid statistical information. Vividness can aid comprehension, so the term denotes bias only when prominence outruns diagnostic value.

Example: One neighbour's dramatic story about a failed operation outweighs a hospital's much larger outcome table.

Recipe: salience-weighting · affect-as-information · associative-retrieval

Evidence: Concrete anecdotes often carry extra weight, but reviews find substantial dependence on relevance, presentation and the judgement being made.

Also called: concreteness effect in judgement

Attentional bias is a systematic tendency for certain kinds of stimuli to capture or hold attention more readily than others. The content of the bias depends on goals, learning and emotional state rather than reflecting one universal direction.

Example: An anxious patient scanning a test report notices every alarming phrase while repeatedly overlooking the reassuring qualifiers.

Recipe: finite-attention · salience-weighting · affect-as-information

Evidence: Supported across dot-probe, eye-tracking and related paradigms, with reliability and clinical specificity varying by measure.

Also called: selective attention bias

Salience bias overweights information that stands out perceptually or socially relative to its context. It overlaps with attention and availability, but concerns prominence at encounter rather than ease of later retrieval alone.

Example: During a post-mortem, the loudest production outage dominates discussion while numerous quieter process failures receive little weight.

Recipe: salience-weighting · finite-attention · causal-attribution

Evidence: Salience effects are broad and well supported, but the umbrella label overlaps heavily with several more specific attention and attribution constructs.

Also called: perceptual salience effect

The isolation effect improves memory for an item that differs from otherwise similar surrounding items. Distinctiveness is relational, so an unusual item is not automatically memorable outside the comparison set that makes it unusual.

Example: A single handwritten expense in a page of printed figures is the one item an auditor later recalls.

Recipe: salience-weighting · finite-attention · associative-retrieval

Evidence: A durable laboratory finding, although accounts based on attention, encoding distinctiveness and retrieval cues remain distinguishable.

Also called: von Restorff effect

The focusing illusion exaggerates the effect of whichever aspect of life is currently drawing attention on an overall evaluation. Other determinants fade from view during judgement even though they will still shape actual experience.

Example: While considering a job in Bengaluru, someone treats the salary difference as if it will determine all future happiness and underweights commute, friends and daily routine.

Recipe: finite-attention · salience-weighting · attribute-substitution

Evidence: Supported in well-being forecasts and related judgements; the size of the illusion depends on the focal question and available comparison information.

Also called: focus illusion

An availability cascade is a collective process in which repeated public attention makes a claim easier to recall and increasingly plausible. Its growth combines informational repetition with social and reputational pressures rather than proving that the claim is true or false.

Example: A rumour repeated across neighbourhood WhatsApp groups starts to feel independently corroborated even when every post traces back to one message.

Recipe: associative-retrieval · social-imitation · norm-coordination · processing-fluency

Evidence: An influential explanatory model of public belief formation rather than a single standardized laboratory effect.

Also called: social availability cascade

The mere-exposure effect is an increase in liking after repeated encounter with a stimulus, even without new substantive information. The relationship is not unlimited, because boredom, prior dislike and exposure pattern can weaken or reverse it.

Example: A song that seemed unremarkable on its first play becomes a favourite after appearing repeatedly during the commute.

Recipe: processing-fluency · associative-retrieval · affect-as-information

Evidence: Meta-analytic evidence supports a generally positive, often curved relation between exposure and liking with important stimulus and timing moderators.

Also called: familiarity preference effect

The illusory-truth effect makes repeated statements feel more true than comparable new statements. Familiarity and processing fluency can influence truth judgement even when repetition supplies no independent evidence.

Example: A false claim seen on several different news feeds begins to feel verified although each feed copied the same wire story.

Recipe: processing-fluency · associative-retrieval · attribute-substitution

Evidence: Robust across many statement types, though prior knowledge, source recollection and exposure conditions moderate the effect.

Also called: repetition-induced truth effect, truth effect

The affect heuristic uses a rapid positive or negative feeling as a shortcut for judging risk, benefit or value. Feelings can summarize genuine experience, but they can also substitute for separate analysis of the attributes being judged.

Example: Because a familiar company feels trustworthy, an investor rates both its expected return as high and its risk as low without examining the balance sheet.

Recipe: affect-as-information · attribute-substitution · processing-fluency

Evidence: Well supported in risk-benefit judgements, especially when time or analytic resources are limited.

Also called: affect-as-information shortcut

Processing fluency is the experienced ease with which information is perceived or understood, and that ease can alter evaluation. People may misattribute fluency to truth, familiarity, beauty or safety rather than to typography, repetition or prior exposure.

Example: A proposal in clear type and familiar language feels better reasoned than an equivalent proposal with awkward formatting.

Recipe: processing-fluency · affect-as-information · attribute-substitution

Evidence: A broad, well-supported mechanism whose behavioural direction depends on what people believe the experienced ease signifies.

Also called: fluency heuristic, ease-of-processing effect

The halo effect lets a salient favourable impression of one attribute spill into judgements of other, weakly evidenced attributes. It is an impression-formation pattern, not proof that every correlated rating is irrational or caused by attractiveness.

Example: A charismatic founder is assumed to be a careful financial manager even though the pitch provides no evidence about accounting discipline.

Recipe: affect-as-information · causal-attribution · attribute-substitution

Evidence: A long-established rating and impression effect, with multiple mechanisms and substantial dependence on information structure.

Also called: halo error

Mood-congruent judgement makes information matching a person's current mood easier to retrieve or more influential. It can operate through memory accessibility and through treating mood as evidence about the object being judged.

Example: After a miserable commute, a team lead recalls mainly recent frustrations and judges the whole project more harshly.

Recipe: affect-as-information · associative-retrieval · reconstructive-memory

Evidence: Supported across memory and judgement tasks, but effects vary with mood intensity, awareness and the material being evaluated.

Also called: mood congruence effect

Negativity bias gives adverse events or information greater psychological weight than comparably positive material. The umbrella term covers attention, learning, impression and memory effects that need not share a single mechanism.

Example: One cutting remark in an otherwise warm performance review occupies the employee's thoughts for the rest of the week.

Recipe: salience-weighting · affect-as-information · associative-retrieval

Evidence: Broad evidence supports many bad-is-stronger effects, but the label joins several tasks and exceptions rather than naming one uniform law.

Also called: negative information bias

The identifiable-victim effect increases concern or giving for a specific known person relative to statistically described people at comparable risk. Identifiability can engage affect and perceived responsibility, but its effects vary with group size, blame and presentation.

Example: A fundraiser for one named child attracts more donations than a spreadsheet describing many children who need the same treatment.

Recipe: affect-as-information · salience-weighting · attribute-substitution

Evidence: Supported in charitable and risk judgements, with notable moderators and no guarantee that a named victim always increases helping.

Also called: identifiability effect

Hindsight bias makes an outcome seem more foreseeable after the outcome is known. Knowledge of what happened can reshape remembered predictions, perceived inevitability and the story told about earlier evidence.

Example: After a product launch fails, colleagues insist the warning signs were obvious although their dated forecasts were mixed.

Recipe: reconstructive-memory · pattern-completion · causal-attribution

Evidence: A well-replicated family of retrospective distortions with separable memory, inevitability and foreseeability components.

Also called: knew-it-all-along effect

Outcome bias evaluates the quality of an earlier decision partly from how the outcome turned out. It confuses information available after the fact with the evidence and process that were available when the choice was made.

Example: A careful medical decision is called foolish after a rare complication, while the same reasoning is praised when the patient recovers.

Recipe: causal-attribution · salience-weighting · reconstructive-memory

Evidence: The classic effect has been reproduced across decision-evaluation settings, although responsibility and outcome type moderate its size.

Also called: resulting

Consistency bias reconstructs past attitudes, traits or behaviour as more similar to the present than they were. The distortion follows beliefs about personal stability and can reverse when people instead expect a particular kind of change.

Example: After changing political views, someone remembers having held roughly the same position years earlier despite old messages showing otherwise.

Recipe: reconstructive-memory · prior-belief-preservation · identity-protection

Evidence: Supported for autobiographical attitudes and traits, with direction governed by implicit theories about stability and change.

Also called: reconstructive consistency effect

The misinformation effect occurs when misleading information encountered after an event alters later reports of that event. It can reflect source confusion, response processes or changes in accessible memory rather than a simple erasure of the original trace.

Example: After witnesses discuss a collision, one later remembers a red taxi mentioned by another witness but absent from the scene.

Recipe: reconstructive-memory · source-monitoring · social-imitation

Evidence: A robust eyewitness-memory phenomenon, while the relative contributions of trace impairment, source error and test demand depend on method.

Also called: post-event misinformation effect

A source-monitoring error assigns a memory, idea or feeling to the wrong origin. Familiar content may be remembered while the distinction between seeing, hearing, inferring or imagining it is lost.

Example: A designer confidently repeats a suggestion as their own, forgetting that it came from a client's earlier email.

Recipe: source-monitoring · reconstructive-memory · processing-fluency

Evidence: A central and well-supported framework for several memory errors rather than a single narrow experimental effect.

Also called: source misattribution

Choice-supportive bias remembers a selected option more favourably and rejected options less favourably than the original information warrants. Misattributing features between options has stronger support than some broader claims about invented or selectively forgotten details.

Example: Months after buying a laptop, its owner assigns the chosen model features that actually belonged to the rejected model.

Recipe: source-monitoring · reconstructive-memory · commitment-consistency

Evidence: Solid evidence supports source misattribution favouring chosen options; other proposed forms of choice-supportive memory are less consistent.

Also called: choice-supportive misremembering

Rosy retrospection evaluates a past experience more positively than it was evaluated while it unfolded. Transient annoyances lose prominence, leaving a summary that can differ from moment-to-moment experience.

Example: A traveller remembers a monsoon holiday as delightful while forgetting the daily complaints recorded in messages sent during the trip.

Recipe: reconstructive-memory · affect-as-information · salience-weighting

Evidence: Demonstrated for meaningful experiences, with timing and the balance of anticipated, experienced and remembered affect shaping the pattern.

Also called: rosy view effect

The fading-affect bias describes unpleasant emotion attached to autobiographical memories fading faster than pleasant emotion. It concerns the emotional intensity linked to remembered events, not necessarily forgetting the negative facts themselves.

Example: The embarrassment of a failed presentation weakens faster than the pleasure attached to a successful one from the same year.

Recipe: reconstructive-memory · affect-as-information · self-enhancement

Evidence: Repeatedly observed in autobiographical memory, with reduced or altered patterns under dysphoria and in some event domains.

Also called: fading affect effect

The telescoping effect misplaces events in time, often pulling older events forward into a recent reporting window. Boundary cues, retention and increasing dating error can produce the pattern without a general compression of subjective time.

Example: Asked about the last six months, a patient reports a fall that actually happened nine months ago.

Recipe: reconstructive-memory · source-monitoring · probability-simplification

Evidence: Well documented in autobiographical dating and survey recall, with forward and backward errors shaped by the reporting interval.

Also called: temporal displacement

Memory conformity occurs when another person's account changes what someone later reports remembering. The report can shift through informational influence, social compliance or source confusion, which should not be treated as identical mechanisms.

Example: Two colleagues compare recollections of a tense meeting, and one later remembers a remark heard only in the other's version.

Recipe: social-imitation · source-monitoring · reconstructive-memory

Evidence: Supported in co-witness and collaborative-memory studies, with effects moderated by confidence, discussion and source access.

Also called: social contagion of memory

The representativeness heuristic judges probability from how much a case resembles a prototype or an expected process. Resemblance can be useful, but it can displace base rates, sample size and the logic of how evidence was generated.

Example: A quiet, meticulous applicant is judged more likely to be an archivist than a salesperson without considering how many sales jobs exist.

Recipe: representativeness · pattern-completion · attribute-substitution

Evidence: A foundational and extensively studied heuristic, though resemblance can be normatively relevant in some environments.

Also called: similarity heuristic

Base-rate neglect gives too little weight to the prior frequency of a category when interpreting case-specific evidence. People do not always ignore priors, and use improves when rates, likelihoods and reference classes are made meaningful.

Example: A rare disease is treated as likely after a positive test without combining the result with how uncommon the disease is.

Recipe: representativeness · probability-simplification · finite-attention

Evidence: Well established, but strongly affected by information format, causal relevance and how the reference class is understood.

Also called: base-rate fallacy

The conjunction fallacy judges a detailed combination of events more probable than one of its constituent events. A coherent, representative story can feel more plausible even though adding a condition cannot increase mathematical probability.

Example: A detailed description makes colleagues rate it more likely that a new hire is both a designer and an environmental activist than that the person is a designer.

Recipe: representativeness · pattern-completion · probability-simplification

Evidence: Replicated across many tasks, while wording, pragmatics and frequency formats can reduce or alter responses.

Also called: conjunction error

The gambler's fallacy expects a run in an independent random process to be followed by compensating reversal. It treats local balance as if a fair process must quickly repair short sequences.

Example: After six heads in a row, a bettor insists the next independent toss is more likely to be tails.

Recipe: representativeness · sample-size-insensitivity · pattern-completion

Evidence: A robust judgement pattern for processes believed to be random; beliefs about the generating mechanism strongly moderate it.

Also called: negative recency in random sequences

Hot-hand belief expects recent success to continue because it is read as evidence of a temporarily elevated success rate. Real performance can contain serial dependence, so the error is unwarranted inference from a short run rather than belief in every possible streak.

Example: A sales team assumes a representative who closed three deals today is now intrinsically more likely to close the next independent lead.

Recipe: representativeness · sample-size-insensitivity · pattern-completion

Evidence: Belief in continuation is well documented, but the classic claim that observed sports streaks are wholly illusory has been substantially revised.

Also called: hot-hand fallacy, belief in streak shooting

Belief in the law of small numbers expects small samples to resemble their parent population more closely than sampling variation allows. It produces overconfidence in patterns, differences and replications drawn from limited observations.

Example: After two customers from one district cancel, a shopkeeper concludes that the whole neighbourhood has lost interest.

Recipe: sample-size-insensitivity · representativeness · confidence-miscalibration

Evidence: A foundational account of sample-size insensitivity with continuing relevance to intuitive statistics and research practice.

Also called: small-sample representativeness

The clustering illusion perceives meaningful bunches or streaks in data that can plausibly arise from random variation. Randomness itself contains runs, so visual irregularity is not sufficient evidence for a changing process.

Example: A support lead sees three outages on Fridays and declares a weekly pattern before checking the full multi-year distribution.

Recipe: pattern-completion · sample-size-insensitivity · representativeness

Evidence: The tendency to underestimate natural clustering is well supported, but any particular observed cluster may still reflect a real process.

Also called: illusory pattern perception in random data

Regression neglect fails to expect extreme observations to be followed, on average, by less extreme observations when measurements are imperfectly correlated. The resulting change is often given a causal story even when ordinary statistical regression is sufficient.

Example: A student sent to tutoring after an unusually poor test improves next time, and the entire improvement is credited to the intervention without a comparison group.

Recipe: representativeness · causal-attribution · probability-simplification

Evidence: A well-established statistical reasoning error, especially when extreme cases are selected for attention or intervention.

Also called: regression-to-the-mean neglect

Denominator neglect gives insufficient weight to the size of the reference set when comparing ratios or risks. Larger numerators can therefore feel more compelling even when the corresponding proportion is smaller.

Example: A patient prefers a treatment described as helping 20 of 200 people over one helping 2 of 10, although the second rate is higher.

Recipe: probability-simplification · finite-attention · attribute-substitution

Evidence: Supported across ratio and risk tasks, with numeracy, visual format and gist-based instruction affecting susceptibility.

Also called: ratio bias, numerator bias

Confirmation bias seeks, interprets or remembers evidence in ways that favour an existing belief, expectation or hypothesis. It is an umbrella construct spanning several processes and does not always require a desired conclusion or strong identity stake.

Example: A developer convinced that the database is at fault repeatedly inspects slow queries and overlooks logs pointing to a network timeout.

Recipe: prior-belief-preservation · finite-attention · associative-retrieval

Evidence: Broadly supported across many paradigms, but the label joins positive testing, selective search, biased evaluation and memory effects that should not be treated as one process.

Also called: confirmatory bias

Myside bias evaluates or generates arguments in a way that favours one's own prior position. It is closely related to confirmation bias but is defined around a person's stated side of an issue rather than any hypothesis currently under test.

Example: Two readers apply strict sourcing rules to an opposing editorial and accept a supportive editorial with only a quick glance.

Recipe: prior-belief-preservation · identity-protection · directional-motivation

Evidence: Observed across argument generation and evaluation tasks and often only weakly related to conventional measures of intelligence.

Also called: own-side bias

Selective exposure is a preference for encountering information that supports an existing attitude or choice over information that challenges it. Accuracy goals, usefulness and curiosity can counter the preference, so avoidance of disagreement is not automatic.

Example: An investor subscribes to channels praising a stock already owned and quietly removes analysts who question it.

Recipe: prior-belief-preservation · directional-motivation · finite-attention

Evidence: Meta-analysis supports a moderate preference for congenial information with substantial variation by commitment, quality and utility.

Also called: congenial information selection

Belief perseverance is the survival of a belief after the evidence that originally supported it has been discredited or removed. Explanations generated around the belief can provide new support even when the initiating evidence is gone.

Example: A hiring manager keeps thinking an applicant is unreliable after learning that the damaging reference was sent for the wrong person.

Recipe: prior-belief-preservation · pattern-completion · commitment-consistency

Evidence: Established in debriefing and correction paradigms, although persistence depends on the belief, explanation and corrective information supplied.

Also called: belief persistence

Motivated reasoning describes goals influencing which memories, arguments and standards of proof are accessed, constructed or accepted. Directional motivation favours a desired conclusion, whereas accuracy motivation can instead increase careful reasoning.

Example: A founder who needs a funding round to succeed treats optimistic market assumptions as realistic and identical pessimistic assumptions as absurd.

Recipe: directional-motivation · identity-protection · prior-belief-preservation

Evidence: A well-supported theoretical framework spanning many domains; motivation changes reasoning through multiple constrained processes rather than unlimited wishful thinking.

Also called: directional reasoning

Attitude polarization is movement towards more extreme positions after considering evidence or discussion. Biased assimilation can produce it when opposing sides evaluate the same mixed evidence by different standards, but polarization also has other social mechanisms.

Example: Supporters and critics of a policy read the same inconclusive report and each leaves more certain that their original position is correct.

Recipe: prior-belief-preservation · directional-motivation · identity-protection

Evidence: Polarization occurs reliably in some attitude and group contexts, but identical mixed evidence does not always increase extremity.

Also called: belief polarization

Disconfirmation bias subjects belief-incongruent evidence to more counterargument and stricter scrutiny than congenial evidence. It concerns unequal evaluation after exposure, unlike selective exposure, which concerns what evidence is encountered.

Example: A manager demands flawless methodology from a survey criticising a favoured initiative but accepts a supportive survey without checking its sample.

Recipe: directional-motivation · prior-belief-preservation · identity-protection

Evidence: Demonstrated in politically charged evaluation tasks; generalisation depends on attitude strength, sophistication and domain.

Also called: motivated scepticism

Belief bias makes the believability of a conclusion influence judgement of whether an argument is logically valid. A believable conclusion can be accepted from invalid premises, while an unbelievable conclusion can make a valid argument harder to endorse.

Example: A plausible claim in a board memo is accepted as logically entailed even though the stated evidence does not actually support it.

Recipe: prior-belief-preservation · attribute-substitution · finite-attention

Evidence: A robust reasoning effect with an interaction between logical validity and conclusion believability.

Also called: belief-bias effect

The continued-influence effect is the persistence of misinformation in reasoning after a correction has been accepted. Retractions may leave a gap in the event model or fail to make the corrective account as retrievable as the original claim.

Example: Residents accept that an electrical fault did not start a fire but still cite the fault when explaining why the building burned.

Recipe: prior-belief-preservation · reconstructive-memory · associative-retrieval

Evidence: Replicated and widely reviewed, although effect sizes and successful correction strategies vary across materials and designs.

Also called: continued influence of misinformation

The anchoring effect pulls a numerical estimate or judgement towards an initially encountered value or position. Insufficient adjustment is one account, but selective accessibility, conversational inference and scale construction can also produce anchoring.

Example: A flat first advertised at ₹80,000 a month makes a later quote of ₹60,000 feel reasonable before comparable rents are checked.

Recipe: reference-dependence · prior-belief-preservation · finite-attention

Evidence: A robust effect across many estimation tasks, with several mechanisms and sensitivity to expertise, relevance and incentives.

Also called: anchoring-and-adjustment

A framing effect changes judgement or choice when equivalent or closely matched information is described in different ways. Risky-choice, attribute and goal framing are related paradigms rather than one unitary mechanism.

Example: Patients respond differently when the same operation is described with a ninety per cent survival rate instead of a ten per cent mortality rate.

Recipe: reference-dependence · salience-weighting · affect-as-information

Evidence: Meta-analytic evidence supports reliable framing effects with marked variation across frame type, task and response mode.

Also called: choice framing effect

Loss aversion gives losses relative to a reference point greater weight than comparable gains. It is a property of choices and valuations under specified conditions, not a fixed personal coefficient or a claim that every loss always dominates every gain.

Example: The prospect of losing ₹1,000 from an existing account feels more consequential than the prospect of gaining ₹1,000 feels attractive.

Recipe: loss-asymmetry · reference-dependence · affect-as-information

Evidence: Influential and supported in many designs, while magnitude, boundary conditions and some common demonstrations remain actively debated.

Also called: loss-gain asymmetry

The endowment effect raises the valuation of an object after it becomes part of a person's endowment or is treated as owned. Loss aversion is one explanation, while attachment, transaction framing and market experience can also matter.

Example: Someone demands much more to sell a concert ticket already held than they would have paid to buy the same ticket that morning.

Recipe: ownership-attachment · reference-dependence · loss-asymmetry

Evidence: Observed in many ownership and exchange tasks, with substantial effects of design, experience, goods and transaction framing.

Also called: ownership valuation effect

Status-quo bias favours an existing state over changing to an alternative beyond what switching costs or evidence alone justify. The current state acts as a reference and change can evoke uncertainty, anticipated regret or loss.

Example: A company renews an expensive vendor contract because replacing it requires a meeting, even though a reviewed alternative is clearly better.

Recipe: default-inertia · reference-dependence · loss-asymmetry

Evidence: Well established in choice experiments and field settings, with rational switching costs needing separate measurement.

Also called: status-quo preference

The sunk-cost effect lets irrecoverable past investment influence whether to continue an action. Past cost can be psychologically relevant through waste aversion or self-justification even when only future costs and benefits should guide the current choice.

Example: A family sits through a dreadful film because the tickets were expensive, although leaving would not recover the money.

Recipe: commitment-consistency · loss-asymmetry · mental-accounting

Evidence: A well-replicated tendency with effect size varying by responsibility, framing, experience and whether future benefits remain genuinely uncertain.

Also called: sunk-cost fallacy

Escalation of commitment is repeated investment in a failing or uncertain course after negative feedback. It overlaps with sunk cost but adds dynamics such as personal responsibility, reputation, public defence and organizational incentives.

Example: A manager who publicly championed a delayed software project secures another budget rather than reopen the decision from current evidence.

Recipe: commitment-consistency · identity-protection · loss-asymmetry

Evidence: Well established across organizational and experimental research, but no single mechanism explains every escalation episode.

Also called: commitment escalation, self-justifying escalation

The default effect increases selection of an option that is preselected or takes effect without active choice. Defaults can work through effort, implied endorsement, reference dependence and procrastination rather than through inertia alone.

Example: Employees are much more likely to join a pension plan when enrolment is automatic and opting out requires a form.

Recipe: default-inertia · authority-deference · reference-dependence

Evidence: Strong laboratory and field evidence shows defaults matter, while mechanisms and ethical interpretation depend on transparency and context.

Also called: default bias

Mental accounting assigns money, costs and outcomes to separate psychological accounts rather than treating all fungible resources alike. Those accounts can support useful budgeting while also creating inconsistent spending and valuation across frames.

Example: Someone refuses to use savings for an urgent repair but pays costly credit-card interest because the savings are mentally reserved for a holiday.

Recipe: mental-accounting · reference-dependence · commitment-consistency

Evidence: A broad, influential framework with many supported manifestations rather than one uniform error.

Also called: mental budgeting

Self-serving bias attributes favourable outcomes more to one's own qualities and unfavourable outcomes more to external circumstances. Motivational self-protection and differences in information or expectations can both contribute to the asymmetry.

Example: A manager credits a successful release to leadership but blames a failed release entirely on the market and suppliers.

Recipe: self-enhancement · identity-protection · causal-attribution

Evidence: A widely supported attributional pattern, with cultural, informational and motivational moderators.

Also called: self-serving attributional bias

The better-than-average effect rates one's abilities or traits more favourably than those of an average peer. Ambiguous trait definitions and selective comparison can produce the pattern, which can reverse for difficult tasks or unfavourable attributes.

Example: Nearly everyone in a team rates their communication as better than the team's average communication.

Recipe: self-enhancement · attribute-substitution · social-imitation

Evidence: Well replicated, with task difficulty, trait ambiguity, comparison group and culture shaping direction and magnitude.

Also called: illusory superiority, above-average effect

Optimism bias expects personally favourable outcomes to be more likely, or adverse outcomes less likely, than evidence supports. Absolute optimism and comparative optimism are related forms and should not be collapsed into one personal optimism score.

Example: A household assumes it is less likely than similar households to face a financial emergency and keeps no contingency fund.

Recipe: self-enhancement · control-inference · directional-motivation

Evidence: A robust comparative judgement pattern with marked variation by controllability, experience, event valence and reference group.

Also called: unrealistic optimism, comparative optimism

Overconfidence is excessive confidence relative to an appropriate accuracy or uncertainty benchmark. Overestimation, overplacement and overprecision are distinct forms that can move differently across easy and difficult tasks.

Example: An engineer gives a one-day estimate with near certainty even though similar integrations have varied from one to ten days.

Recipe: confidence-miscalibration · self-enhancement · finite-attention

Evidence: Extensively studied, but the umbrella label must preserve the distinction among overestimation, overplacement and overprecision.

Also called: confidence miscalibration

The illusion of control overestimates personal influence over outcomes substantially governed by chance or other agents. Choice, competition, familiarity and active involvement can make random outcomes feel more controllable.

Example: A trader believes personally choosing a lottery-like asset makes its random return more manageable than an equivalent asset assigned by a system.

Recipe: control-inference · self-enhancement · causal-attribution

Evidence: Well established in chance and control paradigms, while perceived control can be calibrated or even underestimated in other settings.

Also called: perceived-control illusion

The planning fallacy underestimates how long or how much a future task will require despite relevant experience with past overruns. People often focus on the current plan's imagined steps and underuse the distribution of comparable completed cases.

Example: A software team predicts a six-week migration while treating three previous six-month migrations as exceptional.

Recipe: pattern-completion · self-enhancement · prior-belief-preservation

Evidence: A well-supported forecasting error, especially for own projects; strategic misrepresentation and organizational incentives can coexist with cognitive mechanisms.

Also called: planning optimism

The bias blind spot sees cognitive and motivational bias more readily in other people than in oneself. People often inspect their own intentions while judging others from behaviour, and the absence of a felt motive is mistaken for evidence of objectivity.

Example: During a family argument, each person identifies the other's loyalty and emotion while describing their own position as the plain reading of facts.

Recipe: self-enhancement · causal-attribution · finite-attention

Evidence: Replicated in self-other bias ratings, although measurement partly overlaps better-than-average judgement and introspection effects.

Also called: bias asymmetry

Naive realism is the tendency to experience one's own view as a direct, objective reading of reality. Disagreement is therefore easily explained by the other person's ignorance, irrationality or bias rather than by different information and construal.

Example: Two neighbours interpret the same housing dispute differently, and each concludes that only personal interest could explain the other's reading.

Recipe: causal-attribution · self-enhancement · prior-belief-preservation

Evidence: An influential framework connecting several self-other asymmetries, not one standardized task effect with a single magnitude.

Also called: naive realism in social perception

The Dunning–Kruger effect describes lower performers in a task showing especially poor calibration of their relative performance. Popular claims that novices are always more confident than experts overstate findings that also reflect task difficulty, measurement error and regression towards the mean.

Example: Novices on a spreadsheet test substantially overrank their performance, while experienced users estimate their relative rank more accurately.

Recipe: confidence-miscalibration · self-enhancement · finite-attention

Evidence: Task-specific miscalibration is supported, but the size, mechanism and familiar internet caricature are contested and partly sensitive to statistical artefacts.

Also called: unskilled-and-unaware effect

In-group bias evaluates or allocates resources more favourably towards people categorised as members of one's own group. Minimal categorisation can be sufficient in experiments, but real intergroup behaviour also reflects institutions, norms, status and material interests.

Example: A committee gives familiar alumni the benefit of the doubt while reading identical gaps in an outsider's application as warning signs.

Recipe: in-group-protection · identity-protection · social-imitation

Evidence: Well established from minimal-group studies through field research, with culture, status, norms and competition moderating expression.

Also called: in-group favouritism, ingroup bias

The out-group homogeneity effect perceives members of another group as more alike than members of one's own group. Familiarity, comparative context and identity can weaken, remove or sometimes reverse the asymmetry.

Example: Members of one department describe another department as all thinking alike while easily listing differences within their own team.

Recipe: in-group-protection · finite-attention · pattern-completion

Evidence: Supported across intergroup paradigms, with important familiarity, size, status and comparative-context moderators.

Also called: outgroup homogeneity bias

Authority bias gives an authority figure's claim or instruction extra weight because of perceived status, legitimacy or expertise. Classic obedience findings demonstrate situational influence, but they do not establish a fixed personal tendency to obey every authority.

Example: A junior clinician notices a dosage inconsistency but stays silent because the prescription came from the most senior consultant.

Recipe: authority-deference · norm-coordination · finite-attention

Evidence: Authority and legitimacy reliably influence judgement and action, but the popular label combines distinct obedience, expertise and endorsement processes.

Also called: authority influence, obedience to authority

The bandwagon effect increases support for an option as it appears more popular or widely adopted. Popularity may be read as information about quality or as a cue to social belonging, making it related to but broader than direct conformity pressure.

Example: A restaurant queue attracts more customers because the queue itself is taken as evidence that the food must be good.

Recipe: social-imitation · norm-coordination · associative-retrieval

Evidence: Popularity effects occur in markets, politics and social choice, while the label often overlaps loosely with conformity and informational cascades.

Also called: popularity effect

Conformity changes a judgement or action towards a group norm or majority position. Informational uncertainty and normative concern about acceptance are distinct motives, and effects depend strongly on unanimity, task and culture.

Example: A meeting participant privately doubts a deadline but endorses it after every earlier speaker says it is realistic.

Recipe: norm-coordination · social-imitation · finite-attention

Evidence: A foundational and widely replicated social-influence effect with substantial task, unanimity and cultural moderation.

Also called: majority conformity, Asch conformity effect

Groupthink is a proposed pattern in which pressures for concurrence within a cohesive decision group impair critical evaluation of alternatives. Popular usage often means any bad group decision, while the original model specifies antecedents and symptoms that empirical research supports unevenly.

Example: A leadership team suppresses objections to a risky launch because questioning the shared plan is treated as disloyalty.

Recipe: norm-coordination · identity-protection · commitment-consistency

Evidence: Influential but debated: reviews find heuristic value alongside inconsistent support for the original model's complete causal sequence.

Also called: groupthink model

Pluralistic ignorance occurs when people privately reject a norm but incorrectly believe that most others accept it. Each person's public compliance then becomes misleading evidence that sustains the perceived consensus.

Example: Most employees dislike an always-on messaging norm, yet everyone replies at midnight because each assumes colleagues expect it.

Recipe: norm-coordination · social-imitation · sample-size-insensitivity

Evidence: Well supported in norm-perception research; the construct requires a documented gap between private attitudes and perceived group attitudes.

Also called: collective misperception of norms

Group polarization moves a group's post-discussion position towards a more extreme version of its members' initial tendency. Persuasive arguments, social comparison and identity processes can all contribute, so the direction follows the starting distribution rather than always becoming riskier.

Example: A mildly enthusiastic investment club becomes strongly committed to a speculative stock after members mainly exchange supporting reasons.

Recipe: norm-coordination · prior-belief-preservation · in-group-protection

Evidence: A well-established group-discussion effect whose mechanisms and magnitude vary with argument pools, identity and communication structure.

Also called: risky-shift and cautious-shift generalization

The false-consensus effect overestimates how common one's own choices, beliefs or behaviours are among other people. Selective exposure, ambiguity and using the self as a sample can each contribute without implying that every estimate from personal experience is unreasonable.

Example: A commuter who cycles assumes most colleagues would prefer secure bicycle parking over improved bus service.

Recipe: sample-size-insensitivity · self-enhancement · pattern-completion

Evidence: A robust social-estimation effect with domain, ambiguity and available comparison information influencing its size.

Also called: consensus overestimation

The fundamental attribution error describes overinferring stable dispositions from another person's behaviour or underweighting situational constraints. The correspondence-bias tradition substantially overlaps it, but effects are neither universal across cultures and tasks nor necessarily fundamental.

Example: A late colleague is labelled careless without considering the cancelled suburban trains that morning.

Recipe: causal-attribution · salience-weighting · finite-attention

Evidence: Situational underweighting is well documented, while the broad label covers several paradigms with meaningful cultural and methodological boundaries.

Also called: correspondence bias, disposition bias

The traditional actor–observer hypothesis says people explain their own behaviour more situationally and other people's behaviour more dispositionally. A large meta-analysis found little overall effect, although narrower asymmetries appear under particular valence, relationship and explanation conditions.

Example: A driver explains their own sharp turn by a hidden pothole but initially explains another driver's turn as recklessness.

Recipe: causal-attribution · self-enhancement · finite-attention

Evidence: The famous global formulation is not supported on average; moderator-specific self-other differences remain an active and more defensible research target.

Also called: actor-observer bias

The ultimate attribution error extends self-serving attribution to groups by explaining out-group harm dispositionally and in-group harm situationally. Favourable behaviour can receive the reverse treatment, but the exact pattern depends on identity, status and conflict.

Example: A team's own missed deadline is blamed on impossible requirements while a rival team's identical delay is blamed on incompetence.

Recipe: causal-attribution · in-group-protection · identity-protection

Evidence: An influential intergroup attribution model with support for several predicted patterns but no claim of universal group-serving attribution.

Also called: group-serving attribution bias

Illusory correlation perceives an association between variables that is absent or weaker than believed. Distinctive pairings, prior expectations and uneven memory can make rare group-behaviour combinations seem especially diagnostic.

Example: A supervisor remembers the few remote workers who missed deadlines and overlooks the many remote workers who delivered reliably.

Recipe: associative-retrieval · salience-weighting · pattern-completion

Evidence: A well-established covariation error with both expectancy-based and distinctiveness-based forms.

Also called: spurious association perception

Hostile attribution bias interprets ambiguous interpersonal behaviour as intentionally hostile. It is especially studied in developmental and aggression research and predicts reactive responses without implying that every perceived threat is mistaken.

Example: A child whose classmate bumps a desk accidentally assumes the act was deliberate and retaliates.

Recipe: causal-attribution · affect-as-information · prior-belief-preservation

Evidence: Associated reliably with reactive aggression across diverse samples, with developmental history and context affecting interpretation.

Also called: hostile intent attribution

Belief in a just world assumes that people generally get outcomes they deserve and deserve the outcomes they get. When confronted with innocent suffering, it can support victim derogation or invented responsibility that restores a sense of order.

Example: After hearing about a fraud victim, a listener searches for careless choices that would make the loss feel preventable and deserved.

Recipe: identity-protection · causal-attribution · control-inference

Evidence: A well-established belief and attribution research programme, not merely a momentary shortcut or an inevitable response to every misfortune.

Also called: belief in a just world, just-world bias

Stereotype-based judgement applies group-linked expectations to an individual when individuating evidence is absent, ambiguous or underweighted. This cognitive process cannot by itself explain racism, sexism, caste discrimination or institutional inequity, which also involve history, law, incentives, procedures, resources and power.

Example: A recruiter treats an applicant's neighbourhood as evidence about reliability instead of evaluating the person's record.

Recipe: pattern-completion · finite-attention · prior-belief-preservation

Evidence: A large research domain rather than one unitary bias; category effects vary with goals, information, control and social structure.

Also called: stereotyping, category-based social judgment

Naive cynicism overestimates how much other people's judgements are distorted by self-interest or strategic bias. It is a counterpart to naive realism because one's own account feels principled while another person's account is discounted as motivated.

Example: Two project leads each regard the other's budget forecast as political while treating their own forecast as a neutral reading of need.

Recipe: causal-attribution · self-enhancement · identity-protection

Evidence: Supported in responsibility and interpersonal-judgement studies, though the construct has a smaller independent literature than attribution's major families.

Also called: biased assumptions of bias

The illusion of asymmetric insight makes people feel they know others better than others know them. Observable behaviour is treated as revealing of the other person, while one's own private complexity is treated as hidden from outside view.

Example: Each partner in a negotiation believes they can read the other's motives while insisting their own real priorities remain invisible.

Recipe: causal-attribution · self-enhancement · finite-attention

Evidence: Demonstrated across interpersonal and intergroup studies, with conceptual overlap with naive realism and introspection-based self-other asymmetry.

Also called: asymmetric insight illusion

Present bias gives outcomes available now disproportionate weight relative to outcomes at later dates. It can create dynamically inconsistent plans, while hyperbolic or quasi-hyperbolic discounting are models of the pattern rather than exact synonyms for every impatient choice.

Example: Someone plans all week to save a bonus but spends it on a same-day sale when the money arrives.

Recipe: temporal-discounting · affect-as-information · finite-attention

Evidence: Strong experimental and field evidence supports present-focused discounting, with domain, stakes and commitment devices shaping behaviour.

Also called: immediacy bias, quasi-hyperbolic discounting

Omission bias judges harmful inaction as less blameworthy or risky than comparably harmful action. Responsibility, regret and perceived causation can differ between action and inaction even when expected outcomes are matched.

Example: A parent rejects a low-risk vaccination because harm from the injection would feel more blameworthy than the same harm from leaving the child unprotected.

Recipe: action-omission-asymmetry · causal-attribution · affect-as-information

Evidence: Reliable in many moral and health scenarios, with intent, regret, causality and baseline norms moderating the action-inaction asymmetry.

Also called: inaction preference in harm

Ambiguity aversion prefers options with known probabilities over options whose probabilities are uncertain or underspecified. Ambiguity seeking also occurs in some domains, so the construct describes a common choice pattern rather than a universal dislike of missing information.

Example: An investor chooses a lower-yield deposit with known terms over a potentially better product whose risk model is opaque.

Recipe: probability-simplification · affect-as-information · finite-attention

Evidence: The Ellsberg pattern is foundational, while ambiguity attitudes vary with domain, competence, framing and whether outcomes are gains or losses.

Also called: Ellsberg preference

The certainty effect gives the change from probable to certain outcomes more weight than a comparable probability change among uncertain outcomes. It is a prospect-theory choice regularity, not a claim that certainty is preferred in every context.

Example: A shopper gives up a much larger possible voucher to secure a modest voucher guaranteed today.

Recipe: probability-simplification · reference-dependence · affect-as-information

Evidence: A central, well-supported risky-choice pattern with boundaries set by framing, outcome domain and probability representation.

Also called: certainty overweighting

The disposition effect sells assets showing gains more readily than comparable assets showing losses. Reference prices, regret, tax rules, beliefs about reversal and portfolio rebalancing can all contribute, so loss aversion is not its sole engine.

Example: An investor quickly sells a share that has risen but keeps a deteriorating share to avoid recording the loss.

Recipe: reference-dependence · loss-asymmetry · mental-accounting

Evidence: Well documented in investor records, with multiple behavioural and institutional mechanisms and meaningful market variation.

Also called: sell-winners-hold-losers pattern

The decoy effect increases preference for an option when a clearly inferior alternative is added near it in attribute space. The decoy changes the comparison context even though it is unlikely to be chosen itself.

Example: A café adds an almost-as-expensive smaller drink, making the large drink appear like the obvious bargain.

Recipe: reference-dependence · attribute-substitution · finite-attention

Evidence: A well-established context effect, with magnitude sensitive to attribute structure, expertise, incentives and presentation.

Also called: attraction effect, asymmetric-dominance effect

The compromise effect increases preference for an option when it becomes the middle alternative on salient attributes. The same option can therefore gain appeal after a more extreme alternative is added to the choice set.

Example: A middle-priced phone sells better after the shop adds a premium model above it, even though the middle phone is unchanged.

Recipe: reference-dependence · attribute-substitution · finite-attention

Evidence: A replicated context effect whose size and reliability vary with attribute trade-offs, presentation and real versus hypothetical choice.

Also called: extremeness aversion

Choice overload occurs when a larger option set reduces selection, satisfaction or decision quality because comparison becomes difficult. More choice is not uniformly harmful, and meta-analysis finds strong heterogeneity rather than a universal overload threshold.

Example: A shopper leaves without choosing mobile data after trying to compare dozens of nearly indistinguishable plans.

Recipe: finite-attention · attribute-substitution · affect-as-information

Evidence: The original effect is influential, while meta-analysis finds a near-zero average with important moderators and substantial heterogeneity.

Also called: overchoice

Action bias favours doing something over waiting when action provides a sense of agency or signals effort despite no improvement in expected outcome. The effect is domain- and framing-dependent and is not the opposite of omission bias in every setting.

Example: After one quiet sales week, an executive launches a disruptive reorganisation mainly to demonstrate that management is responding.

Recipe: control-inference · affect-as-information · action-omission-asymmetry

Evidence: Demonstrated in several applied tasks, but the direction depends on norms, perceived agency and whether action or inaction is the salient default.

Also called: preference for action