We are wired to assume that things that move or happen have someone behind them, a mind with intentions. When a bush rustles, it feels safer to guess creature than wind, so our brains guess agent first and ask questions later 1Barrett 2000, Trends in Cognitive Sciences. Introduced the hyperactive agent detection device (HADD), the proposal that human minds are tuned to over-attribute events to intentional agents. reference.
Everyday Hearing the house creak at night and instantly feeling sure someone is in the hallway, when it is just the radiator cooling.
An availability cascade is a snowball of belief: the more a claim gets repeated in public, the more plausible it feels, which prompts more people to repeat it. The belief spreads not because new evidence arrived but because the claim became easy to bring to mind 3Kuran and Sunstein 1999, Stanford Law Review. Defined availability cascades as self-reinforcing belief formation driven by rising availability in public discourse, propelled by informational and reputational motives. reference.
Everyday A rumor that a local restaurant failed a health inspection ricochets around town group chats until everyone knows it, even though no inspection ever happened.
Holding two thoughts that clash, or acting against what you believe, produces real mental discomfort. People usually relieve it not by changing their behavior but by quietly adjusting their beliefs to fit what they already did 4Festinger and Carlsmith 1959, Journal of Abnormal and Social Psychology. Participants paid $1 to describe a dull task as fun came to believe it more than those paid $20, showing attitudes shift to justify behavior. reference.
Everyday A smoker who knows the health risks decides the studies are exaggerated, rather than quitting.
When the measure of a problem and the measure of its cause both come from the same source, such as one survey answered by one person, the shared source can manufacture a correlation all by itself. Conclusions look stronger than they are because the errors travel together 7Meier and O'Toole 2013, Journal of Public Administration Research and Theory. Administrators' self-reported performance measures were biased in predictable ways and produced spurious relationships absent in independent indicators. reference.
Everyday Employees asked in a single questionnaire to rate both their own effort and their team's success tend to produce a rosy link between the two, whatever is actually happening.
What you already believe walks into the room before the evidence does. Strong initial beliefs, accurate or not, shape how new facts get weighed, so two people can read the same study and come away more convinced of opposite things 8Lord, Ross and Lepper 1979, Journal of Personality and Social Psychology. Proponents and opponents of capital punishment shown identical mixed evidence each judged it as supporting their side and became more polarized. reference.
Everyday A fan convinced the referee hates their team sees every close call as proof, and every fair call as an exception.
When two options come to mind, the one that arrives faster and more smoothly tends to be judged bigger, truer or better. The mind quietly treats ease of processing as a signal of quality 9Hertwig, Herzog, Schooler and Reimer 2008, Journal of Experimental Psychology: Learning, Memory, and Cognition. Four experiments showed people can discriminate retrieval fluency and use it as a proxy for real-world quantities, especially in fast inferences. reference.
Everyday Asked which of two cities is larger, most people confidently pick the one whose name they retrieved instantly, without knowing any population facts.
In tight, friendly groups, keeping harmony can quietly become more important than getting things right. Doubters censor themselves, the group overestimates its own virtue, and bad plans sail through unchallenged 10Janis 1972, Victims of Groupthink (Houghton Mifflin, book). Case analyses of policy fiascoes (Bay of Pigs, Vietnam escalation) attributing failures to cohesion-driven self-censorship and illusions of unanimity; later empirical tests of the full model have shown mixed support. reference.
Everyday A group of friends plans a road trip nobody privately wants, because each person assumes everyone else is excited.
Talking with like-minded people does not average out opinions; it moves the whole group further in the direction it already leaned. After discussion, a cautious group gets more cautious and a bold group gets bolder 11Myers and Lamm 1976, Psychological Bulletin. Review showing group discussion reliably shifts members toward more extreme positions in the group's initial direction, best explained by informational influence. reference.
Everyday Five friends each mildly annoyed at a landlord finish dinner together collectively furious and drafting a lawsuit.
Most of us think we understand how everyday things work, until someone asks us to explain them step by step. Rozenblit and Keil found that confidence in our explanations runs well ahead of the explanations themselves 12Rozenblit and Keil 2002, Cognitive Science. People rated their understanding of devices and phenomena far higher before attempting explanations than after, with the illusion strongest for explanatory knowledge. reference.
Everyday Feeling sure you know how a zipper works, then stalling at "well, the little teeth... interlock?"
Statements feel truer each time we meet them, whether or not they are true. Hasher and colleagues showed that mere repetition raised people's confidence in trivia statements, including the false ones 13Hasher, Goldstein and Toppino 1977, Journal of Verbal Learning and Verbal Behavior. Plausible statements repeated across sessions were rated more valid than novel ones, regardless of actual truth. reference.
Everyday Believing that people only use ten percent of their brains because you have heard it a hundred times, not because you ever checked.
When rewards arrive unpredictably, people tend to spread their choices to match the frequencies, picking the 70 percent option about 70 percent of the time, instead of always picking the better option. Matching feels balanced, but it reliably earns less than committing to the best bet 14Vulkan 2000, Journal of Economic Surveys. Survey of probability matching experiments documenting persistent sub-optimal matching behavior that violates expected utility maximization. reference.
Everyday Guessing heads on some flips and tails on others in a biased-coin game, instead of always calling the side that comes up more often.
A statement that rhymes feels more accurate than the same statement without the rhyme. McGlone and Tofighbakhsh found people rated rhyming aphorisms as truer than matched non-rhyming versions of the same idea 15McGlone and Tofighbakhsh 2000, Psychological Science. Rhyming aphorisms were judged more accurate than semantically matched non-rhyming versions, an effect attenuated when people were warned to separate form from content. reference.
Everyday "An apple a day keeps the doctor away" feels like medical advice; "eat fruit regularly for health" feels like a suggestion.
What gets measured starts to feel more real and more important than what does not, even when the unmeasured thing matters more. Judgment tilts toward whatever comes with a number attached 16Muller 2018, The Tyranny of Metrics (Princeton University Press, book). Cross-domain case studies of metric fixation, where standardized measurement displaces judgment and distorts institutional goals; conceptual rather than experimental evidence. reference.
Everyday Choosing a job for the salary figure over the unquantified daily misery of a long commute.
Attention goes to whatever is vivid, moving or emotionally striking, and judgment follows attention. Taylor and Fiske showed that even arbitrary changes in what is visually prominent shift who gets credited and blamed 17Taylor and Fiske 1978, Advances in Experimental Social Psychology. Perceptually salient people and features draw attention and are subsequently weighted more heavily in causal attribution and judgment. reference.
Everyday After a plane crash dominates the news, flying feels dangerous while the far riskier drive to the airport feels fine.
When we tailor a message to please a particular audience, we tend to remember and believe our own tailored version afterward. Higgins and Rholes found that tuning a description to a listener's attitude shifted the speaker's own later memory of the underlying facts 18Higgins and Rholes 1978, Journal of Experimental Social Psychology. Modifying a message about a person to suit a listener's attitude produced lasting changes in the communicator's own recall and evaluation of that person. reference.
Everyday Describing a mutual acquaintance kindly to their best friend, then noticing you genuinely remember them more fondly.
When a sample chooses itself, or gets chosen in a way that is tangled up with the thing being studied, conclusions drawn from it quietly stop applying to anyone else. The data are real; they are just not about the population you think 19Heckman 1979, Econometrica. Showed sample selection bias operates as a specification error and introduced a correction, foundational for analyzing non-random samples. reference.
Everyday An online poll about whether people like online polls.
Our default setting is to believe what people tell us, and we detect lies barely better than chance. In Bond and DePaulo's meta-analysis across tens of thousands of judgments, accuracy averaged 54 percent, with truths believed far more readily than lies were caught 21Bond and DePaulo 2006, Personality and Social Psychology Review. Meta-analysis of 206 studies: people judged lie versus truth at 54 percent accuracy, classifying 61 percent of truths but only 47 percent of lies correctly. reference.
Everyday Accepting "the check is in the mail" at face value, again.
People rate vague, one-size-fits-all descriptions as impressively accurate portraits of themselves, especially when the descriptions are flattering and supposedly personalized. Forer's students gave a stock horoscope-style sketch high marks as a description of their unique personalities 22Forer 1949, Journal of Abnormal and Social Psychology. Students rated an identical generic personality sketch, presented as personalized, as a highly accurate description of themselves. reference.
Everyday Reading "you have a great need for others to like you, yet you tend to be critical of yourself" and feeling deeply seen.
We judge arguments by whether we like the conclusion, not just by whether the logic holds. Evans and colleagues found people accepted invalid syllogisms with believable conclusions and rejected valid ones with unbelievable conclusions 23Evans, Barston and Pollard 1983, Memory and Cognition. Three experiments showed substantial belief bias in syllogistic reasoning, strongest on invalid arguments with believable conclusions. reference.
Everyday "All fish swim; salmon swim; therefore salmon are fish" sounds airtight to most people, though the logic is broken.
When you only look at cases that cleared some bar, like hospital admission or getting funded, two unrelated traits can appear negatively correlated. The selection itself manufactures the pattern 24Berkson 1946, Biometrics Bulletin. Demonstrated that studying only hospitalized patients creates spurious associations between diseases that are absent in the general population. reference.
Everyday Among people invited to a party for being either funny or generous, the funny ones seem stingier, only because the generous-but-unfunny ones still got invited.
Random data naturally contains streaks and clumps, but our eyes read them as patterns with causes. Gilovich, Vallone and Tversky showed that basketball's beloved hot hand was largely a misreading of ordinary chance sequences 25Gilovich, Vallone and Tversky 1985, Cognitive Psychology. Shooting records showed no evidence of streak shooting, yet players and fans confidently perceived a hot hand in random-like sequences. reference.
Everyday A slot machine pays out twice in ten minutes and the seat is declared lucky.
We search for, notice, interpret and remember evidence that fits what we already believe, and slide past what does not. Nickerson's classic review found it operating in courtrooms, medicine, science and everyday judgment 26Nickerson 1998, Review of General Psychology. Review documenting confirmation bias as the seeking and interpreting of evidence partial to existing beliefs, across many practical domains. reference.
Everyday Convinced your phone battery is failing, you notice every fast drain and forget the days it lasted fine.
When testing an idea, people run the test that would confirm it rather than tests that could rule out alternatives. In Wason's 2-4-6 task, most people only proposed number triples that fit their own hypothesis, so they never discovered the real rule 27Wason 1960, Quarterly Journal of Experimental Psychology. In the 2-4-6 task, most participants tested only hypothesis-confirming triples and failed to discover the rule. reference; Baron and colleagues later documented the same overvaluing of confirmatory questions in diagnostic reasoning 28Baron, Beattie and Hershey 1988, Organizational Behavior and Human Decision Processes. Subjects overvalued questions likely to give positive results under the favored hypothesis, a direct demonstration of congruence bias in question selection. reference.
Everyday Suspecting the WiFi is down, you refresh the same page five times instead of trying another site or another device.
The size of a problem, meaning how many individuals it touches, barely moves our feelings or our judgments; we respond to the prototype, not the count. In a famous valuation study, people offered essentially the same money to save 2,000, 20,000 or 200,000 birds 29Desvousges et al. 1993, in Hausman (ed.), Contingent Valuation: A Critical Assessment. Willingness to pay to save migratory waterfowl was nearly identical whether 2,000, 20,000 or 200,000 birds would be saved. reference.
Everyday Feeling roughly the same pull to donate whether a shelter appeal mentions eight dogs or eight hundred.
Implicit assumptions about gender shape judgments of competence, leadership and credibility, usually without anyone noticing or intending it. In a randomized experiment, science faculty rated an identical job application as more competent and more hireable when it carried a male name 30Moss-Racusin et al. 2012, Proceedings of the National Academy of Sciences. In a randomized double-blind study, faculty rated a male-named applicant more competent and hireable and offered higher salary than the identical female-named applicant. reference.
Everyday In a meeting, an idea gets a lukewarm reception, then real enthusiasm when restated minutes later by a male colleague.
We see relationships that are not there, especially between things that are individually rare or memorable, because unusual pairings stick in memory. Hamilton and Gifford showed people perceived a link between a minority group and undesirable behavior in data that contained no such link 31Hamilton and Gifford 1976, Journal of Experimental Social Psychology. Co-occurrence of two infrequent things (minority group membership and rare behaviors) produced perceived correlations absent from the data, a cognitive basis for stereotypes. reference.
Everyday "It always rains right after I wash the car."
It feels responsible to gather more information before deciding, but some information cannot change any decision you would make, and we overvalue it anyway. Baron and colleagues found people rated diagnostic questions as worth asking even when no possible answer would alter the chosen action 28Baron, Beattie and Hershey 1988, Organizational Behavior and Human Decision Processes. Subjects overvalued questions likely to give positive results under the favored hypothesis, a direct demonstration of congruence bias in question selection. reference.
Everyday Reading a fortieth review of a toaster that is already in your cart.
Researchers who expect a result tend to find it, by unconsciously cueing subjects, nudging procedures or reading ambiguous data their way. Rosenthal and Fode's students recorded better maze performance from rats labeled bright than from identical rats labeled dull 32Rosenthal and Fode 1963, Behavioral Science. Randomly labeled 'maze-bright' rats outperformed 'maze-dull' rats solely because student experimenters expected them to, demonstrating experimenter expectancy effects. reference.
Everyday A teacher told that certain students are late bloomers grades their ambiguous answers a little more generously.
People's confidence in their answers, plans and estimates routinely outruns their accuracy, though the pattern is subtle: we overestimate ourselves most on hard tasks and hold overly precise beliefs almost everywhere. Moore and Healy's review untangles these distinct kinds of overconfidence 33Moore and Healy 2008, Psychological Review. Distinguished overestimation, overplacement and overprecision; people overestimate performance on hard tasks, and overprecision is the most persistent form. reference.
Everyday Being ninety-nine percent sure of the directions and adding twenty minutes to the drive.
The brain is so eager to find meaningful patterns, especially faces, that it manufactures them from noise: clouds, toast, electrical outlets. Neuroimaging shows the same face-processing machinery that handles real faces activates for illusory ones 34Liu et al. 2014, Cortex. Face pareidolia in pure noise images engaged the right fusiform face area, with top-down expectation driving illusory face perception. reference.
Everyday The car's headlights and grille smiling at you from the driveway.
We accept a claim as true when it feels personally meaningful or when our beliefs need it to be true, rather than because the evidence supports it. The term comes from Marks and Kammann's investigation of psychic claims, where believers found personal confirmation in vague or chance results 35Marks and Kammann 1980, The Psychology of the Psychic (Prometheus Books, book). Coined subjective validation to describe accepting statements as true because they feel personally meaningful; evidence base is largely investigative rather than experimental. reference.
Everyday A horoscope says a difficult conversation awaits, and by evening you have found the conversation it must have meant.
We study what made it through some filter, planes that returned, businesses still open, and forget that the filter erased the most informative cases. Wald's wartime insight was to armor bombers where the returning planes were not hit, because planes hit there never came home 36Mangel and Samaniego 1984, Journal of the American Statistical Association. Exposition of Abraham Wald's World War II method for estimating aircraft vulnerability from surviving planes, the canonical survivorship-bias correction. reference.
Everyday "My grandfather smoked a pack a day and lived to ninety" leaves out all the grandfathers who did not.
Attitudes and stereotypes we do not endorse, and often cannot introspect on, can still tilt our snap judgments of people and groups. The Implicit Association Test made these associations measurable by timing how quickly people pair concepts 37Greenwald, McGhee and Schwartz 1998, Journal of Personality and Social Psychology. Introduced the Implicit Association Test, showing reaction-time differences reveal associations people consciously disavow. reference.
Everyday A hiring manager who sincerely values fairness still feels an unexamined sense of fit with candidates who resemble past hires.
Facing an unfamiliar quantitative problem, people grab numbers that were handed to them and answer with those, even when the right answer requires computing something new. In Bayesian reasoning studies, wrong answers overwhelmingly matched values given in the problem rather than reflecting calculation slips 39Talboy and Schneider 2022, Frontiers in Psychology. Across Bayesian reasoning problem variants, incorrect responses almost always conformed to values provided in the problem rather than computed quantities; evidence base for this bias is still narrow. reference.
Everyday Asked what an 80 dollar item costs at 25 percent off, reaching for 25 or 80 instead of doing the arithmetic to get 60.
1.Barrett 2000, Trends in Cognitive Sciences Introduced the hyperactive agent detection device (HADD), the proposal that human minds are tuned to over-attribute events to intentional agents.
3.Kuran and Sunstein 1999, Stanford Law Review Defined availability cascades as self-reinforcing belief formation driven by rising availability in public discourse, propelled by informational and reputational motives.
5.Piazza et al. 2015, Appetite The 4Ns (natural, normal, necessary, nice) captured 83 to 91 percent of the justifications people spontaneously offer for eating meat.
6.Rothgerber 2020, Appetite Conceptual framework showing meat eaters rely mainly on perceptual and cognitive strategies, rather than behavior change, to reduce meat-related dissonance.
10.Janis 1972, Victims of Groupthink (Houghton Mifflin, book) Case analyses of policy fiascoes (Bay of Pigs, Vietnam escalation) attributing failures to cohesion-driven self-censorship and illusions of unanimity; later empirical tests of the full model have shown mixed support.
11.Myers and Lamm 1976, Psychological Bulletin Review showing group discussion reliably shifts members toward more extreme positions in the group's initial direction, best explained by informational influence.
12.Rozenblit and Keil 2002, Cognitive Science People rated their understanding of devices and phenomena far higher before attempting explanations than after, with the illusion strongest for explanatory knowledge.
14.Vulkan 2000, Journal of Economic Surveys Survey of probability matching experiments documenting persistent sub-optimal matching behavior that violates expected utility maximization.
15.McGlone and Tofighbakhsh 2000, Psychological Science Rhyming aphorisms were judged more accurate than semantically matched non-rhyming versions, an effect attenuated when people were warned to separate form from content.
19.Heckman 1979, Econometrica Showed sample selection bias operates as a specification error and introduced a correction, foundational for analyzing non-random samples.
24.Berkson 1946, Biometrics Bulletin Demonstrated that studying only hospitalized patients creates spurious associations between diseases that are absent in the general population.
26.Nickerson 1998, Review of General Psychology Review documenting confirmation bias as the seeking and interpreting of evidence partial to existing beliefs, across many practical domains.
32.Rosenthal and Fode 1963, Behavioral Science Randomly labeled 'maze-bright' rats outperformed 'maze-dull' rats solely because student experimenters expected them to, demonstrating experimenter expectancy effects.
33.Moore and Healy 2008, Psychological Review Distinguished overestimation, overplacement and overprecision; people overestimate performance on hard tasks, and overprecision is the most persistent form.
34.Liu et al. 2014, Cortex Face pareidolia in pure noise images engaged the right fusiform face area, with top-down expectation driving illusory face perception.
39.Talboy and Schneider 2022, Frontiers in Psychology Across Bayesian reasoning problem variants, incorrect responses almost always conformed to values provided in the problem rather than computed quantities; evidence base for this bias is still narrow.
