Abstract

Intuition is a type of cognition: knowing or understanding something without consciously working it out. Psychology approaches it from several directions, including the systematic errors of fast judgment, the recognition skills of experienced professionals, the feelings that accompany an immediate impression, and the finding that people sometimes reach logically correct answers intuitively. Dual-process accounts frame intuition as a fast mode of thought set against slower deliberation, although whether the two modes form distinct systems is disputed. The central practical question is when intuition can be trusted. Current work suggests the answer depends on whether the environment offers learnable regularities, and that a confident intuition is not reliably an accurate one.

Keywords: intuition, dual processing, heuristics, intuitive expertise, logical intuition

Two research traditions have shaped what psychologists say about intuition. One set out to catalogue the predictable mistakes that quick judgments make; the other went into the field to study how experienced professionals decide under pressure. For decades they seemed to disagree about whether intuition deserves trust. When Kahneman and Klein (2009) reviewed the two programmes together, they concluded that intuitive expertise can develop only where the environment is predictable enough to supply regularities a person can learn, and that the subjective confidence attached to an intuition is a poor guide to whether it is correct.

Underneath both traditions sits the dual-process picture, in which a fast, effortless mode of thought is contrasted with a slow, deliberate one. The title of Kahneman's (2011) book, Thinking, fast and slow, has become a common shorthand for that picture. The sections below take the main lines of research in turn, state where each theory now stands, and close with what is settled and what is not.

Key Takeaways

  • Heuristics, mental shortcuts that usually serve well and economically, also produce systematic, predictable errors in quick judgment, such as the conjunction fallacy, rating a combination of events as more probable than one of its parts.
  • Dual-process theory, the idea that thought runs in a fast intuitive mode and a slow deliberate mode, organises much of the field but is contested; a major review found that the proposed attributes do not sort cleanly into two systems.
  • Recognition-primed decision, sizing up a situation as a familiar type and acting without weighing alternatives, was documented in interviews with 26 fireground commanders, a single domain; expert intuition can be trusted only where the environment has learnable regularities, and confidence is not a reliable sign of accuracy.
  • Logical intuitions, correct answers reached without slow correction, turn up often in two-response studies, although the most extensive evidence comes from one item, the bat-and-ball problem, and not every correct answer arrives intuitively.
  • The somatic marker hypothesis, the claim that bodily signals steer choices before a person is aware of the reason, is contested: a study of 40 undergraduates found that participants knew more than had been assumed, without proving the hypothesis wrong.
  • Unconscious thought theory, the claim that a distracted mind makes better complex choices, was not supported by a meta-analysis and a replication with 399 students, and a registered replication across 21 laboratories found no cooperative boost from forcing fast decisions, though a meta-analysis of 67 studies linked intuition to more pure cooperation.

What Intuition Is

Knowing or understanding without conscious use of reasoning.

Scope note for the MeSH descriptor Intuition (D019545), U.S. National Library of Medicine.

The MeSH scope note captures the core of the concept: intuition delivers a conclusion without the steps that produced it being available to awareness. Within the wider study of thinking, that makes intuition a matter of outcome rather than procedure, since the person has an answer, a preference or an impression but cannot say how it was reached.

Researchers have tried to sharpen this into a working definition. In a theoretical review aimed at managerial decision making, Dane and Pratt (2007) reconceptualised intuition and marked it off from two neighbours, insight and rational decision making. They also proposed conditions under which intuition should prove effective, involving the decision maker's knowledge of the domain, the way that knowledge was acquired, and the nature of the task, though these were offered as propositions and left untested.

A second strand traces intuition to how it is learned. Reviewing behavioural and neuroscientific evidence, Lieberman (2000) argued that social intuition rests on implicit learning, the gradual and unreported pickup of regularities, and noted that deliberate attempts to learn explicitly can interfere with it. His case was built by inference from patient and imaging studies rather than from a single decisive test.

A third strand treats intuition as the product of a whole system of mind. Epstein's (1994) cognitive-experiential self-theory posits two interacting systems, a rational one and an experiential one, and places heuristic reasoning on the experiential side. Intuition in this account is what the experiential system delivers. The theory's current standing is taken up in the next section.

Intuition is not confined to everyday judgment. Bowers and colleagues examined it in the context of discovery, the subject announced by the title of their paper (Bowers et al., 1990).

Dual-Process Accounts

Most modern accounts of intuition are versions of dual-process theory, and that theory is contested. Its long history is now under explicit challenge: De Neys (2023) argued that two of its working assumptions, that a given response comes from one mode or the other and that thinking switches from the fast mode to the slow one, rest on weak foundations. His critique sketches the groundwork for a new architecture without yet testing it.

The modern case began with reasoning. Sloman (1996) reviewed evidence for two reasoning systems, one associative, driven by similarity and by things occurring close together in time, and one rule-based, operating on symbolic structures. He observed that the two can arrive at conflicting answers and that the rule-based system does not fully suppress the associative one.

Epstein's two-system scheme produced one of the best-known demonstrations of that conflict. In a laboratory task, participants could draw from a bowl holding 1 in 10 winning jelly beans or from one holding 7 in 100, and many chose the larger bowl with the worse odds even though they knew the odds were worse (Denes-Raj & Epstein, 1994). The effect was weaker when the draws concerned losing rather than winning. Pacini and Epstein (1999) later measured rational and experiential thinking styles with the Rational-Experiential Inventory across two studies and found them to be independent dimensions with distinct correlates. On the ratio-bias task, a stronger rational style went with fewer nonoptimal choices, whereas experiential style showed no relation to them.

Cognitive-experiential self-theory's two thinking styles have since been tested by meta-analysis. A meta-analysis of 89 samples and 17,704 participants found that a reflective thinking style was linked to better decision performance and experience, but the associations were small, varied across tasks, ages and time pressure, and depended on context (Phillips et al., 2016). Thinking styles predict outcomes, in other words, but only weakly on average; an intuitive style went with slightly worse performance but a better decision experience, and effects were strongest where a task suited a style's theoretical strengths, a pattern the authors found consistent with the two systems' characteristics.

The broader dual-process family was soon showing strain. Reviewing work across reasoning, judgment and social cognition, Evans (2008) found that the attributes usually assigned to the fast and slow systems do not line up neatly onto two systems, that there seem to be several kinds of implicit process, and that the relevant literatures had developed largely in isolation from one another. Evans and Stanovich (2013) responded by recasting the distinction as one between Type 1 and Type 2 processing, tying the slower type to working memory and to hypothetical thinking, and treating fast responses as defaults. They answered five main lines of criticism, conceding that some objections hold for particular theories while arguing that critics had wrongly treated a varied family of theories as a single position.

The critics were not persuaded. Keren and Schul (2009) judged that two-system theories had added little scientific understanding, questioned whether the notion of a system was conceptually coherent, found the empirical evidence for separate systems weak, and called for stricter definitions and tests. Melnikoff and Bargh (2018) went further, rejecting the dual-process typology as a myth; their argument, which presents no new data, is that properties such as awareness, control and effort do not travel together in two fixed packages, and that assuming they do holds research back. Kruglanski and Gigerenzer (2011) offered an alternative: intuitive and deliberate judgments follow the same rules, selected through a two-step process, and deliberate judgment is not in general the more accurate. Rule selection, they argued, is the important open question, and the two-step process is their proposal for answering it.

Heuristics and Biases

The heuristics and biases program made intuitive error its central topic. Tversky and Kahneman (1974) described three heuristics that people use to judge under uncertainty, representativeness, availability, and adjustment from an anchor, and showed that each leads to systematic, predictable biases. They were careful to add that these shortcuts are usually effective and economical; the errors reveal how the shortcuts work, not that they are generally poor. The availability heuristic, for example, judges frequency by how readily instances come to mind.

The program's reading of intuition as a source of bias remains contested. From the ecological rationality side, a review of formal models of heuristic inference concluded that strategies which ignore information can outperform the strategy of weighting and adding every cue, so that a shortcut is not inherently a defect (Gigerenzer & Gaissmaier, 2011). The disagreement between the two readings has not been resolved.

Across a range of contexts, and among experts as well as lay people, participants judged a conjunction of two events as more probable than one of its components, violating the conjunction rule of probability, an error known as the conjunction fallacy; Tversky and Kahneman (1983) attributed the violations to the representativeness and availability heuristics and set intuitive reasoning against extensional reasoning, while also discussing alternative interpretations of the results. Hertwig and Gigerenzer (1999) challenged this reading in a series of experiments, arguing that conjunction errors could reflect reasonable semantic and pragmatic inferences about what the many-sensed word probability means, not a breakdown in reasoning, and that violations fell when the context or a request for frequency judgments led people to read probability in its mathematical sense, an effect they credited to that reading rather than to extensional cues.

Demo 2: does adding a detail make a story more likely?

Sam is 34, cycles to work every day and spends weekends volunteering at a wildlife charity. Which is more probable?

An invented description written for this article. The diagram shows the rule behind the conjunction fallacy: the probability of two things together can be no greater than the probability of either one alone.

The program later restated its core idea. Kahneman and Frederick (2002) traced its history back to a 1969 survey of 84 participants whose respondents placed too much confidence in small samples, a pattern Tversky and Kahneman explained by the expectation that a sample resembles its population, and they reformulated the representativeness heuristic in terms of attribute substitution, answering an easier question in place of the one asked. Kahneman (2003) then organised the whole account around accessibility, the ease with which a thought comes to mind, which he used to explain both intuitive biases and framing effects. In the same review he argued that the study of biases is consistent with a view of intuition as generally skilled and successful, and he traced the occasional corrections of intuitive judgments to how accessible the relevant rules are.

Individual differences turned the question from what people do on average to who errs. Frederick (2005) introduced the Cognitive Reflection Test, a three-item measure, and found that scores predicted preferences about time and risk; he also reported sex differences in scores. Stanovich and West (2000) used variation in performance on classic heuristics and biases tasks to weigh rival explanations for the gap between normative and actual answers: that the experimenter had applied the wrong standard, that participants construed the task differently, that capacity limits got in the way, or that errors were random slips. Random slips turned out to matter little, capacity limits accounted for only some tasks, and they argued that attention to the most common response alone had been too narrow.

Stanovich (2018) later refined this analysis. Distinguishing five processing states and three kinds of processing defect, he held that whether a person detects the need to override a miserly response, and succeeds in doing so, is strongly related to how fully the relevant mindware, the learned rules and knowledge a problem calls for, is in place. He also cautioned that a correct answer need not reflect costly thought, and that earlier inferences from these tasks had been too simple.

Ecological Rationality

Ecological rationality asks how well a judgment strategy fits the environment in which it is used, and it too is contested. Even its main review named a general theory of what heuristics are built from as a challenge still ahead, and gave low predictability and small samples only as examples of conditions in which ignoring information can make judgments more accurate (Gigerenzer & Gaissmaier, 2011). Kahneman's (2003) accessibility account, which treats intuitive biases as a product of what comes to mind most easily, represents the competing reading.

The approach shares its starting point with bounded rationality: minds have limited time and information, so the question is which strategies work given those limits. An early result concerned how information is presented. Gigerenzer and Hoffrage (1995) found that statistically naive participants reached Bayesian answers more often when problems were framed in frequency formats, the counts a person would meet through natural sampling of cases one at a time, than when the same problems were framed as probabilities. The authors analysed several thousand solutions to Bayesian problems.

The question was posed directly in the title of a paper by Cosmides and Tooby, which asked whether humans are good intuitive statisticians after all (Cosmides & Tooby, 1996).

The later review generalised the argument. Gigerenzer and Gaissmaier (2011) described an adaptive toolbox of heuristics and documented less-is-more effects, in which using less information produces more accurate inferences. They noted that the assumptions of standard rational models are often not met in real environments, which is where frugal strategies can do well.

Intuitive Expertise

Naturalistic decision making approached intuition from the opposite direction, starting with skilled people who seemed to decide well without deliberation. In critical-incident interviews, Klein and colleagues asked 26 experienced fireground commanders about past incidents and analysed 156 decision points; the commanders typically recognised a situation as an instance of a familiar type and acted on it, rather than generating and comparing options (Klein et al., 2010). The study was carried out in 1985, long went unpublished, appeared with a later postscript, and is confined to firefighting. Its account became the recognition-primed decision model, which matters for decision making research because it describes choice without a choice set.

The recognition-primed decision model is supported under conditions. Kahneman and Klein (2009), writing jointly from the heuristics and biases and naturalistic traditions, agreed that intuitive expertise requires an environment predictable enough to offer regularities that can be learned. They also agreed that subjective confidence is an unreliable indicator of whether an intuition is accurate. The demonstration below adds a second dial, the speed of feedback, as an illustrative stand-in for the second condition Kahneman and Klein name, the individual's opportunity to learn those regularities; its values are the article's own, not measurements from that paper.

Demo 3: when can intuition be learned?

Both conditions are met: there are regular cues to learn and feedback to learn them from. Intuitions formed here can become skilled, though confidence alone still does not show they are right.

Illustrative only: the 0 to 1 scales and the 0.6 cut-off are this article's own, not measured quantities. They express the idea that skilled intuition needs both a learnable environment and the chance to learn it.

Applied reviews carried the idea into workplaces. Salas and colleagues reviewed the literature on expertise-based intuition in organisations, organised the review around four goals, and argued that intuition aids expert decisions and should be studied with its multilevel nature in mind (Salas et al., 2010). Their review presents no new data.

Medicine supplies a parallel case. In a narrative review, Stolper and colleagues proposed that the gut feelings of general practitioners form a third track in diagnostic reasoning, arising from non-analytical processing, beside two established diagnostic tracks, one concerned with making medical decisions and one with solving clinical problems, and distinguished two kinds: a sense of alarm that something is wrong and a sense of reassurance that nothing is (Stolper et al., 2011). The model was not tested against data; the authors called for research on how well gut feelings predict outcomes and noted that a questionnaire to measure them was still being validated.

Feelings and Intuition

A recurring idea is that intuition is felt before it is understood, which ties the topic to emotion. The strongest early evidence came from a gambling task. Bechara and colleagues compared healthy participants with patients who had prefrontal damage and found that healthy participants began to show anticipatory skin conductance responses before risky choices, and to choose advantageously, before they could state the advantageous strategy (Bechara et al., 1997). The patients never developed the anticipatory responses, and knowing the strategy was not enough for them to act on it. The authors read this as a nonconscious bias that precedes conscious knowledge.

That reading underpins the somatic marker hypothesis, which is contested; the exchange described here ran from 1997 to 2004, and this article cites no later evaluation of it. Maia and McClelland (2004) reran the Iowa Gambling Task with 40 undergraduates, 20 in each of two conditions, one probing knowledge of the decks with a more detailed structured questionnaire and the other repeating the original questions. Participants' conscious knowledge of the decks kept pace with their advantageous choices, so the behaviour could be explained without invoking somatic markers. They stressed the limits of their own result: a nonconscious contribution was not ruled out, the causal role of awareness remained unresolved, and the hypothesis had not been shown to be wrong.

Mood shapes intuitive judgment in more ordinary settings too. In two experiments with word triads, some of which shared a remote associate and some of which did not, Bolte and colleagues found that participants in a positive mood judged coherence intuitively better than others; those in a negative mood performed at chance, while those in a neutral mood were above chance (Bolte et al., 2003). The authors attributed the effect to spreading activation in semantic memory, but that mechanism was inferred rather than directly tested.

Topolinski and Strack (2009) proposed a chain from fluency to feeling to judgment. Across 11 experiments using word triads, visual coherence and artificial grammar learning, they manipulated how easily material could be processed and induced affect in several ways. The authors argued that easier processing triggers a brief positive affect; in their experiments, high fluency and positive affect each raised judgments of coherence, and the effects extended to visual coherence and grammaticality. Judgments shifted with induced affect regardless of whether the material was actually coherent, and participants could not correct for that influence.

Logical Intuitions

The standard dual-process story says that intuition yields the biased answer and only slow thinking can correct it. A body of work on logical intuitions challenges that order. In a theoretical review, De Neys (2012) argued that even people who give the biased answer show sensitivity to the conflict between that answer and normative considerations, and that this implicit conflict detection implies some intuitive grasp of the principle itself. He presented the properties of these logical intuitions only in outline.

The two-response paradigm tested the claim directly. Participants give a first answer quickly, under a deadline and often while holding a memory load, and then a second answer with unlimited time. Across four experiments, Bago and De Neys (2017) found that correct logical responses were frequently already present at the fast first stage. In seven studies of the bat-and-ball problem using time pressure and cognitive load, the same authors found that people who answered correctly in the end had often done so from the start, though not every correct answer arose that way (Bago & De Neys, 2019). That second paper tested a single problem.

Demo 1: answer fast, then think again

A kettle and a mug cost £1.20 together. The kettle costs £1.00 more than the mug. How much does the mug cost?

An item written for this article in the style of a reflection test. It illustrates the two-response method; one item says nothing about any person's reasoning ability.

Cognitive capacity appears to work through intuition more than through correction. In two two-response studies, Raoelison and colleagues found that capacity was more strongly associated with being correct at the intuitive stage than with switching from a wrong first answer to a right one, though the link with correction was weaker rather than absent (Raoelison et al., 2020). Thompson and colleagues reached a compatible result with a reasoning task modelled on the Stroop paradigm in two experiments with 112 and 224 reasoners: for people of higher capacity, logical and statistical principles appeared to be more intuitive, the pattern held regardless of a strict deadline, and the authors suggested that this may account for at least part of the link between capacity and accuracy (Thompson et al., 2018).

Earlier, Pennycook and colleagues had proposed a three-stage account of when analytic thinking engages: intuitive responses arise first, conflict between them may or may not be detected, and detection can lead either to decoupling from the first answer or to rationalising it (Pennycook et al., 2015). Base-rate experiments supported the account, and showed that conflict detection sometimes fails and that errors can occur at more than one stage.

Reviewing these developments, De Neys and Pennycook (2019) concluded that intuitive logic undermines the classic view in which System 1 errs and System 2 corrects, while noting that the revised models meant to replace it are still being developed.

Intuition and Deliberation

If intuitions are sometimes right, something must decide when to check them. Thompson and colleagues proposed that a metacognition signal, the feeling of rightness that accompanies a first answer, does this job (Thompson et al., 2011). In four two-response experiments with 300 reasoners across four tasks, a weaker feeling of rightness went with longer rethinking and more frequent changes of answer. How fluently the first answer came predicted the feeling, but that evidence was correlational. Thompson's later review with Ackerman set out a broader account of meta-reasoning, in which feelings of certainty and of uncertainty track how reasoning is going and govern how much effort a person invests in it (Ackerman & Thompson, 2017). The review cautioned that the heuristic cues feeding these feelings need not be trustworthy guides, and that reasoning and problem solving had so far attracted little study from this angle (Ackerman & Thompson, 2017).

Difficulty itself can prompt analysis. Across four experiments, Alter and colleagues found that making material harder to process, for instance by printing it in a degraded, hard-to-read font or having people furrow their brows, increased analytic reasoning on syllogisms and in persuasion tasks and reduced reliance on heuristics and defaults (Alter et al., 2007). In their account, metacognitive difficulty is what moves processing from System 1 toward System 2. When 16 replication attempts were pooled with the original study, the disfluent fonts conferred no advantage on counterintuitive mathematics problems in any condition tested (Meyer et al., 2015).

The opposite claim, that deliberation can be bettered by not deliberating, was stated in unconscious thought theory, also called the deliberation-without-attention effect, and it is contested. Dijksterhuis and colleagues reported four studies, in the laboratory and with shoppers, in which people made better complex choices after a period of distraction than after conscious thought, while simple choices favoured conscious thought; the outcome was measured as how favourably people later viewed their choice (Dijksterhuis et al., 2006). The claim did not survive closer testing. A meta-analysis of 61 studies found a small pooled advantage (g = 0.15) with funnel asymmetry pointing to bias among small studies, an advantage that fell to a non-significant g = 0.018 once the asymmetry was corrected, and a replication with 399 University of Amsterdam undergraduates found no evidence for it (Nieuwenstein et al., 2015).

Reflection does not automatically win either. In the meta-analysis of thinking styles described above, a reflective style was associated with slightly better decision performance and experience, but the effects were small and depended on context, and time pressure weakened the link between reflection and performance but not the link for intuition (Phillips et al., 2016).

Applied work suggests that a lack of reflection matters for misinformation. In two online studies with 3446 Mechanical Turk workers, people who scored higher on the Cognitive Reflection Test were better at telling fake headlines from real ones, a pattern the authors read as fitting a lack of reasoning better than motivated reasoning (Pennycook & Rand, 2019). The sample was a convenience sample and the evidence correlational.

Social and Moral Intuition

Moral judgment offers the boldest claim for intuition's priority. Haidt's (2001) social intuitionist model holds that moral judgments arise intuitively and that moral reasoning usually follows afterward to justify them. He set the model explicitly against rationalist models of moral judgment, which place reasoning first, and gave four reasons for preferring it. A competing account came from two fMRI studies, in which the degree of emotional engagement varied from one moral dilemma to another and influenced the judgments people reached, and the authors held that reason and emotion are both likely to play a part in moral judgment (Greene et al., 2001). The model was contested from the start, and this article cites no later direct test of it.

Cooperation became a test case for a related idea, the social heuristics hypothesis, which is also contested. Across ten studies with economic games, Rand and colleagues found that faster or intuition-induced decisions were more cooperative and that reflection undermined cooperation (Rand et al., 2012). A registered replication run in 21 laboratories in 12 countries with 3,596 participants, using a one-shot four-person public goods game, did not support the effect of time pressure on cooperation (Bouwmeester et al., 2017). In the intent-to-treat analysis the difference was −0.37 percentage points. Restricting the analysis to participants who complied with the time limits yielded a 10.37 percentage point difference, against 15.31 points in the original study, but noncompliance was high under time pressure and that analysis is exposed to selection bias; the authors also noted that individual laboratories were underpowered and that the intent-to-treat estimate may understate the manipulation's potency. Across a meta-analysis of 67 studies (N = 17,647), intuition went with 17.3% more pure cooperation than deliberation did, whereas strategic cooperation showed no significant difference, and the bias tests gave no indication of publication bias (Rand, 2016).

Intuition in the Brain

Neuroscientific work on intuition is sparse and its conclusions modest. Lieberman (2000) linked the implicit learning that he saw as the basis of social intuition to the basal ganglia, especially the caudate and putamen, drawing on patient and imaging data.

Imaging studies have looked at intuitive judgments directly. Using a modified version of the Waterloo Gestalt Closure Task, in which fragmented images may or may not depict a coherent object, Volz and von Cramon (2006) found that intuitive judgments of coherence were associated with activity in the median orbitofrontal cortex, the amygdala and the insula. They acknowledged that what exactly is recognised in such judgments is unclear, and offered their definition of intuition only as a promising starting point.

Expertise leaves its own trace. Comparing professional and amateur shogi players with functional magnetic resonance imaging, Wan and colleagues found activation in the precuneus and the caudate that was specific to professionals, and linked this circuit to the quick perception of board patterns and the generation of a best next move (Wan et al., 2011).

Worked Example

The conditions for trustworthy intuition can be made concrete with a deliberately simple illustration. The ratings below are this article's own invented values, not data from any study, and the rule applied to them is a teaching device. The first rating follows the emphasis on learnable regularities in Kahneman and Klein (2009): how regular each setting's environment is. Both ratings use this article's own 0 to 1 scale. The second, how quickly feedback arrives, is an illustrative stand-in for the individual's opportunity to learn those regularities, the second condition in Kahneman and Klein (2009). The illustrative rule trusts intuition only when both ratings exceed 0.6.

A weather forecaster works in an environment rated 0.7 for regularity, because weather follows physical patterns that repeat, and 0.9 for feedback, because each forecast is checked within a day. Both values exceed 0.6, so the setting passes. An emergency nurse is rated 0.8 for regularity and 0.8 for feedback, since patterns of deterioration recur and a patient's response to care is visible within hours; this setting passes as well.

A stock picker is rated 0.2 for regularity and 0.7 for feedback. Prices move daily, so feedback is quick, but the movements contain little regularity that an individual could learn. The setting fails on regularity alone, which shows that fast feedback cannot rescue an environment without learnable structure. A long-range political pundit is rated 0.3 for regularity and 0.1 for feedback, failing on both counts: political events are poorly predictable and forecasts years ahead are rarely scored.

Two of the four settings pass. The exercise mirrors a point made in the Intuitive Expertise section, that confidence is not the test. A stock picker or a pundit may feel just as sure as a nurse; the illustrative rule asks instead whether the environment could ever have taught the intuition.

Table 1

Illustrative Ratings of Four Settings for Learning Intuitive Skill

SettingRegularity of the environmentSpeed of clear feedbackBoth above 0.6?Reading
Weather forecaster0.70.9YesRegular cues and daily feedback: intuition can become skilled
Emergency nurse0.80.8YesRecurring patterns and quick outcomes: intuition can become skilled
Stock picker0.20.7NoFeedback arrives, but prices hold few stable cues
Long-range political pundit0.30.1NoIrregular events and feedback years later

Note. This article's own invented ratings on a 0 to 1 scale and an illustrative cut-off, not data. A setting passes only when both ratings exceed 0.6.

Discussion

Figure 1

The Two-Response Method

A reasoning problem answered twice, first fast and then freely, giving four patterns of answers A flow diagram. A reasoning problem leads to a first answer given quickly, under time pressure or a memory load. That leads to a second, final answer given with unlimited time. Four outcome boxes below show the possible patterns: right then right, wrong then right, wrong then wrong, and right then wrong. The right-then-right box is highlighted because it shows a correct answer reached without slow correction. reasoning problem first answerfast, under pressure final answerno time limit right, then rightcorrect from the start wrong, then rightcorrected by thought wrong, then wrongerror kept right, then wrongchanged for the worse Comparing the two answers shows whether a correct final answer needed slow correction.
Note. An original schematic of the method, not data; it shows no proportions.

Some conclusions about intuition are secure. Quick judgments that rely on heuristics depart from normative standards in systematic, predictable ways (Tversky & Kahneman, 1974). Expert intuition can be accurate, but only in environments regular enough to be learned, and a person's confidence does not reveal whether that condition holds (Kahneman & Klein, 2009). Correct answers to reasoning problems often arrive early rather than through later correction, at least in two-response studies of the bat-and-ball problem (Bago & De Neys, 2019).

These findings sit uneasily together, and that is where the open questions lie. The first is architectural: whether intuition and deliberation belong to two distinct systems, to two types of processing, or to a single set of rules applied under different conditions. Defenders and critics of the dual-process view have each refined their positions (Evans & Stanovich, 2013; Melnikoff & Bargh, 2018), and the matter is not settled. The second concerns feeling. Evidence that fluency and affect shape intuitive judgments comes from a series of 11 experiments (Topolinski & Strack, 2009), but the claim that bodily signals guide choices ahead of awareness is disputed (Maia & McClelland, 2004). The third concerns replication. Two influential effects have fared differently under scrutiny: the advantage of unconscious thought did not hold up in a meta-analysis with a large replication (Nieuwenstein et al., 2015), while for the cooperative tilt of fast decisions a registered replication found no effect of time pressure but a meta-analysis linked intuition to more pure cooperation, leaving that question open (Bouwmeester et al., 2017; Rand, 2016).

What would settle these questions is easier to describe than to run. Studies that track the same problems across fast and slow responses, in populations beyond students, with effects estimated in advance and tested across laboratories, would show which claims about intuition generalise and which belong to particular tasks.

Current Directions

Much current work uses the two-response paradigm to ask what intuition can do on its own. Studies have found that correct logical answers often appear at the fast stage (Bago & De Neys, 2017, 2019), and that cognitive capacity predicts the accuracy of first answers more strongly than successful correction (Raoelison et al., 2020). Higher-capacity reasoners seem to find logical and statistical principles more intuitive in the first place (Thompson et al., 2018).

Those results are driving theory revision. Reviews argue that findings of logical intuition have cast doubt on the classic corrective picture and are prompting revised models in which intuition can be logical (De Neys & Pennycook, 2019), and that the assumptions of exclusive modes and a switch between them need rethinking (De Neys, 2023). Others reject the dual-process typology outright (Melnikoff & Bargh, 2018) or refine the analysis of when people override a miserly response (Stanovich, 2018).

Replication and synthesis form a third strand. A registered replication across 21 laboratories failed to confirm that time pressure increases cooperation (Bouwmeester et al., 2017), and a meta-analysis found only small, context-dependent links between thinking styles and decision outcomes (Phillips et al., 2016). In applied research, reflective thinking has been associated with better discernment of fake news in correlational online studies (Pennycook & Rand, 2019).

Common Misconceptions

Intuition is always worse than deliberate thinking.
A meta-analysis of 89 samples found that a reflective thinking style was linked to slightly better decisions and an intuitive style to slightly worse performance but a better decision experience, with small, context-dependent effects (Phillips et al., 2016). The belief grows out of vivid demonstrations of intuitive error, which show that intuition can mislead, not that deliberation reliably does better.
Setting a complex choice aside and letting the unconscious work produces better decisions.
A meta-analysis of 61 studies found only a small pooled advantage with signs of small-study bias, and a replication with 399 students found no advantage, the difference falling short of significance (Nieuwenstein et al., 2015). The idea persists because the original studies reported exactly that advantage for complex choices.
An expert's confidence shows whether their intuition is right.
Subjective confidence is an unreliable indicator of accuracy; what matters is whether the expert's environment is predictable enough to offer learnable regularities and whether the expert has had the opportunity to learn them (Kahneman & Klein, 2009). Confident intuitions feel the same whether or not they were earned.
Correct answers to tricky reasoning problems always require slow correction.
In seven two-response studies of the bat-and-ball problem, people who answered correctly often did so from their first, fast response, although not every correct answer arose that way (Bago & De Neys, 2019). The belief reflects the classic dual-process account, in which intuition errs and deliberation repairs.

Glossary

Attribute substitution.
Answering a hard judgment question by unknowingly answering an easier, related one in its place.
Availability heuristic.
Judging how frequent or likely something is by how easily examples come to mind.
Cognitive Reflection Test.
A short set of word problems used to measure the tendency to check a first response rather than accept it.
Conjunction fallacy.
Rating the joint occurrence of two events as more probable than one of those events alone.
Dual-process theory.
A family of accounts that divides thinking into a fast, effortless mode and a slow, deliberate one; its core assumptions are contested.
Ecological rationality.
The view that a judgment strategy should be evaluated by how well it fits the structure of the environment where it is used.
Feeling of rightness.
The sense of confidence that accompanies an initial answer and influences whether a person thinks further about it.
Heuristic.
A mental shortcut that reaches a judgment with little effort, usually serving well but open to systematic error.
Iowa Gambling Task.
A card-choice game in which some decks are better choices than others over many draws, used to study how people learn to choose advantageously.
Logical intuition.
A fast, effortless sensitivity to a logical or probabilistic principle, shown when a correct answer is reached without slow reasoning.
Processing fluency.
The ease with which information is perceived or understood, which can itself colour judgments about that information.
Recognition-primed decision.
A choice made by recognising a situation as a familiar type and acting on that recognition, without comparing several options.
Representativeness heuristic.
Judging probability by how closely a case resembles a typical example of a category.
Somatic marker.
A bodily signal proposed to tag options with emotional value and so bias choice; whether such signals guide decisions ahead of awareness is contested.
Two-response paradigm.
A method in which people give a fast first answer under time pressure or load and then a second answer with unlimited time, so that intuitive and deliberate responses can be compared.
Type 1 processing.
Fast, default processing that does not depend on working memory.
Type 2 processing.
Slower, effortful processing that draws on working memory and supports hypothetical thinking.

Key Researchers

Jonathan St. B. T. Evans

(University of Plymouth). Cognitive psychologist who helped develop dual-process accounts of reasoning. Faculty page · Wikipedia · Wikidata · Google Scholar

Gerd Gigerenzer

(Harding Center for Risk Literacy, University of Potsdam). Psychologist who leads the research programme on simple heuristics and ecological rationality. ORCID · Wikipedia · Wikidata

Daniel Kahneman

(Princeton University). Psychologist (1934–2024) whose work on heuristics and biases traced systematic errors in quick judgment, and who later set out with Gary Klein when expert intuition can be trusted. Wikipedia · Wikidata · Google Scholar

Gary Klein

(MacroCognition LLC). Research psychologist who founded the study of naturalistic decision making and proposed the recognition-primed decision model. ORCID · Wikipedia · Wikidata

Wim De Neys

(CNRS / Université Paris Cité). Cognitive scientist whose work on logical intuitions questions whether fast thinking must be corrected by slow thinking. ORCID

Gordon Pennycook

(Cornell University). Psychologist who studies analytic thinking, the Cognitive Reflection Test and susceptibility to misinformation. ORCID · Wikipedia · Wikidata

Valerie A. Thompson

(University of Saskatchewan). Cognitive psychologist who studies the feeling of rightness that accompanies a first answer and decides whether people think further. Faculty page

Frequently Asked Questions

What is intuition in psychology?

Psychologists treat intuition as reaching a judgment or conclusion without conscious reasoning. One influential review set it apart from both insight and rational decision making and proposed, without testing, conditions under which it should be effective (Dane & Pratt, 2007).

When can intuition be trusted?

Expert intuition can be trusted only in environments predictable enough to offer regularities, and only when the person has had the opportunity to learn them. In such settings it can be accurate, while the confidence that accompanies it is not a reliable guide (Kahneman & Klein, 2009).

Is intuition the same as System 1?

Intuition is usually equated with the fast mode in dual-process accounts, but that framework is contested. A recent critique argues that its assumptions about separate modes and a switch between them are poorly founded (De Neys, 2023).

Why do intuitive judgments go wrong?

Quick judgments rely on heuristics such as representativeness, availability and anchoring. These shortcuts are usually effective and economical, yet they produce systematic, predictable errors (Tversky & Kahneman, 1974).

Can intuition get the logically correct answer?

Often it can. In seven two-response studies of the bat-and-ball problem, many people who ended with the correct answer had already given it at the fast first stage, although not all correct answers arose that way (Bago & De Neys, 2019).

Do gut feelings guide decisions before people know why?

Early gambling task studies suggested so, but a study of 40 undergraduates found that conscious knowledge kept pace with good choices. Its authors noted that a nonconscious role was not ruled out and the somatic marker hypothesis was not proven wrong (Maia & McClelland, 2004).

Does a brief diversion of thought improve a complex decision?

The evidence does not support an unconscious thought advantage. A meta-analysis found only a small effect with signs of publication bias, and a replication with 399 students found none (Nieuwenstein et al., 2015).

Are people who think more reflectively better decision makers?

Slightly, on average. A meta-analysis of 17,704 participants found small positive links between a reflective style and decision outcomes, and the links varied with context (Phillips et al., 2016).

References

Ackerman, R., & Thompson, V. A. (2017). Meta-reasoning: Monitoring and control of thinking and reasoning. Trends in Cognitive Sciences, 21(8), 607–617. https://doi.org/10.1016/j.tics.2017.05.004

Alter, A. L., Oppenheimer, D. M., Epley, N., & Eyre, R. N. (2007). Overcoming intuition: Metacognitive difficulty activates analytic reasoning. Journal of Experimental Psychology: General, 136(4), 569–576. https://doi.org/10.1037/0096-3445.136.4.569

Bago, B., & De Neys, W. (2017). Fast logic?: Examining the time course assumption of dual process theory. Cognition, 158, 90–109. https://doi.org/10.1016/j.cognition.2016.10.014

Bago, B., & De Neys, W. (2019). The smart System 1: Evidence for the intuitive nature of correct responding on the bat-and-ball problem. Thinking & Reasoning, 25(3), 257–299. https://doi.org/10.1080/13546783.2018.1507949

Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R. (1997). Deciding advantageously before knowing the advantageous strategy. Science, 275(5304), 1293-1295. https://doi.org/10.1126/science.275.5304.1293

Bolte, A., Goschke, T., & Kuhl, J. (2003). Emotion and intuition: Effects of positive and negative mood on implicit judgments of semantic coherence. Psychological Science, 14(5), 416-421. https://doi.org/10.1111/1467-9280.01456

Bouwmeester, S., Verkoeijen, P. P. J. L., Aczel, B., Barbosa, F., Bègue, L., Brañas-Garza, P., Chmura, T. G. H., Cornelissen, G., Døssing, F. S., Espín, A. M., Evans, A. M., Ferreira-Santos, F., Fiedler, S., Flegr, J., Ghaffari, M., Glöckner, A., Goeschl, T., Guo, L., Hauser, O. P., … Wollbrant, C. E. (2017). Registered Replication Report: Rand, Greene, and Nowak (2012). Perspectives on Psychological Science, 12(3), 527–542. https://doi.org/10.1177/1745691617693624

Bowers, K. S., Regehr, G., Balthazard, C., & Parker, K. (1990). Intuition in the context of discovery. Cognitive Psychology, 22(1), 72-110. https://doi.org/10.1016/0010-0285(90)90004-N

Cosmides, L., & Tooby, J. (1996). Are humans good intuitive statisticians after all? Rethinking some conclusions from the literature on judgment under uncertainty. Cognition, 58(1), 1–73. https://doi.org/10.1016/0010-0277(95)00664-8

Dane, E., & Pratt, M. G. (2007). Exploring intuition and its role in managerial decision making. Academy of Management Review, 32(1), 33-54. https://doi.org/10.5465/amr.2007.23463682

Denes-Raj, V., & Epstein, S. (1994). Conflict between intuitive and rational processing: When people behave against their better judgment. Journal of Personality and Social Psychology, 66(5), 819-829. https://doi.org/10.1037/0022-3514.66.5.819

De Neys, W. (2023). Advancing theorizing about fast-and-slow thinking. Behavioral and Brain Sciences, 46, Article e111. https://doi.org/10.1017/S0140525X2200142X

De Neys, W. (2012). Bias and conflict: A case for logical intuitions. Perspectives on Psychological Science, 7(1), 28–38. https://doi.org/10.1177/1745691611429354

De Neys, W., & Pennycook, G. (2019). Logic, fast and slow: Advances in dual-process theorizing. Current Directions in Psychological Science, 28(5), 503-509. https://doi.org/10.1177/0963721419855658

Dijksterhuis, A., Bos, M. W., Nordgren, L. F., & van Baaren, R. B. (2006). On making the right choice: The deliberation-without-attention effect. Science, 311(5763), 1005-1007. https://doi.org/10.1126/science.1121629

Epstein, S. (1994). Integration of the cognitive and the psychodynamic unconscious. American Psychologist, 49(8), 709-724. https://doi.org/10.1037/0003-066x.49.8.709

Evans, J. St. B. T. (2008). Dual-processing accounts of reasoning, judgment, and social cognition. Annual Review of Psychology, 59, 255–278. https://doi.org/10.1146/annurev.psych.59.103006.093629

Evans, J. St. B. T., & Stanovich, K. E. (2013). Dual-process theories of higher cognition: Advancing the debate. Perspectives on Psychological Science, 8(3), 223–241. https://doi.org/10.1177/1745691612460685

Frederick, S. (2005). Cognitive reflection and decision making. Journal of Economic Perspectives, 19(4), 25-42. https://doi.org/10.1257/089533005775196732

Gigerenzer, G., & Gaissmaier, W. (2011). Heuristic decision making. Annual Review of Psychology, 62, 451-482. https://doi.org/10.1146/annurev-psych-120709-145346

Gigerenzer, G., & Hoffrage, U. (1995). How to improve Bayesian reasoning without instruction: Frequency formats. Psychological Review, 102(4), 684–704. https://doi.org/10.1037/0033-295X.102.4.684

Greene, J. D., Sommerville, R. B., Nystrom, L. E., Darley, J. M., & Cohen, J. D. (2001). An fMRI investigation of emotional engagement in moral judgment. Science, 293(5537), 2105–2108. https://doi.org/10.1126/science.1062872

Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychological Review, 108(4), 814-834. https://doi.org/10.1037/0033-295X.108.4.814

Hertwig, R., & Gigerenzer, G. (1999). The 'conjunction fallacy' revisited: How intelligent inferences look like reasoning errors. Journal of Behavioral Decision Making, 12(4), 275–305.

Kahneman, D. (2003). A perspective on judgment and choice: Mapping bounded rationality. American Psychologist, 58(9), 697–720. https://doi.org/10.1037/0003-066X.58.9.697

Kahneman, D., & Frederick, S. (2002). Representativeness revisited: Attribute substitution in intuitive judgment. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and biases: The psychology of intuitive judgment (pp. 49–81). Cambridge University Press. https://doi.org/10.1017/CBO9780511808098.004

Kahneman, D., & Klein, G. (2009). Conditions for intuitive expertise: A failure to disagree. American Psychologist, 64(6), 515-526. https://doi.org/10.1037/a0016755

Kahneman, D. (2011). Thinking, fast and slow. Farrar, Straus and Giroux.

Keren, G., & Schul, Y. (2009). Two is not always better than one: A critical evaluation of two-system theories. Perspectives on Psychological Science, 4(6), 533-550. https://doi.org/10.1111/j.1745-6924.2009.01164.x

Klein, G., Calderwood, R., & Clinton-Cirocco, A. (2010). Rapid decision making on the fire ground: The original study plus a postscript. Journal of Cognitive Engineering and Decision Making, 4(3), 186-209. https://doi.org/10.1518/155534310X12844000801203

Kruglanski, A. W., & Gigerenzer, G. (2011). Intuitive and deliberate judgments are based on common principles. Psychological Review, 118(1), 97-109. https://doi.org/10.1037/a0020762

Lieberman, M. D. (2000). Intuition: A social cognitive neuroscience approach. Psychological Bulletin, 126(1), 109-137. https://doi.org/10.1037/0033-2909.126.1.109

Maia, T. V., & McClelland, J. L. (2004). A reexamination of the evidence for the somatic marker hypothesis: What participants really know in the Iowa gambling task. Proceedings of the National Academy of Sciences, 101(45), 16075-16080. https://doi.org/10.1073/pnas.0406666101

Melnikoff, D. E., & Bargh, J. A. (2018). The mythical number two. Trends in Cognitive Sciences, 22(4), 280-293. https://doi.org/10.1016/j.tics.2018.02.001

Meyer, A., Frederick, S., Burnham, T. C., Guevara Pinto, J. D., Boyer, T. W., Ball, L. J., Pennycook, G., Ackerman, R., Thompson, V. A., & Schuldt, J. P. (2015). Disfluent fonts don't help people solve math problems. Journal of Experimental Psychology: General, 144(2), e16–e30. https://doi.org/10.1037/xge0000049

Nieuwenstein, M. R., Wierenga, T., Morey, R. D., Wicherts, J. M., Blom, T. N., Wagenmakers, E.-J., & van Rijn, H. (2015). On making the right choice: A meta-analysis and large-scale replication attempt of the unconscious thought advantage. Judgment and Decision Making, 10(1), 1-17. https://doi.org/10.1017/s1930297500003144

Pacini, R., & Epstein, S. (1999). The relation of rational and experiential information processing styles to personality, basic beliefs, and the ratio-bias phenomenon. Journal of Personality and Social Psychology, 76(6), 972-987. https://doi.org/10.1037/0022-3514.76.6.972

Pennycook, G., Fugelsang, J. A., & Koehler, D. J. (2015). What makes us think? A three-stage dual-process model of analytic engagement. Cognitive Psychology, 80, 34–72. https://doi.org/10.1016/j.cogpsych.2015.05.001

Pennycook, G., & Rand, D. G. (2019). Lazy, not biased: Susceptibility to partisan fake news is better explained by lack of reasoning than by motivated reasoning. Cognition, 188, 39-50. https://doi.org/10.1016/j.cognition.2018.06.011

Phillips, W. J., Fletcher, J. M., Marks, A. D. G., & Hine, D. W. (2016). Thinking styles and decision making: A meta-analysis. Psychological Bulletin, 142(3), 260-290. https://doi.org/10.1037/bul0000027

Rand, D. G. (2016). Cooperation, fast and slow: Meta-analytic evidence for a theory of social heuristics and self-interested deliberation. Psychological Science, 27(9), 1192–1206. https://doi.org/10.1177/0956797616654455

Rand, D. G., Greene, J. D., & Nowak, M. A. (2012). Spontaneous giving and calculated greed. Nature, 489(7416), 427-430. https://doi.org/10.1038/nature11467

Raoelison, M., Thompson, V. A., & De Neys, W. (2020). The smart intuitor: Cognitive capacity predicts intuitive rather than deliberate thinking. Cognition, 204, Article 104381. https://doi.org/10.1016/j.cognition.2020.104381

Salas, E., Rosen, M. A., & DiazGranados, D. (2010). Expertise-based intuition and decision making in organizations. Journal of Management, 36(4), 941-973. https://doi.org/10.1177/0149206309350084

Sloman, S. A. (1996). The empirical case for two systems of reasoning. Psychological Bulletin, 119(1), 3-22. https://doi.org/10.1037/0033-2909.119.1.3

Stanovich, K. E. (2018). Miserliness in human cognition: The interaction of detection, override and mindware. Thinking & Reasoning, 24(4), 423–444. https://doi.org/10.1080/13546783.2018.1459314

Stanovich, K. E., & West, R. F. (2000). Individual differences in reasoning: Implications for the rationality debate? Behavioral and Brain Sciences, 23(5), 645–665. https://doi.org/10.1017/S0140525X00003435

Stolper, E., Van de Wiel, M., Van Royen, P., Van Bokhoven, M., Van der Weijden, T., & Dinant, G. J. (2011). Gut feelings as a third track in general practitioners' diagnostic reasoning. Journal of General Internal Medicine, 26(2), 197–203. https://doi.org/10.1007/s11606-010-1524-5

Thompson, V. A., Pennycook, G., Trippas, D., & Evans, J. St. B. T. (2018). Do smart people have better intuitions? Journal of Experimental Psychology: General, 147(7), 945–961. https://doi.org/10.1037/xge0000457

Thompson, V. A., Prowse Turner, J. A., & Pennycook, G. (2011). Intuition, reason, and metacognition. Cognitive Psychology, 63(3), 107–140. https://doi.org/10.1016/j.cogpsych.2011.06.001

Topolinski, S., & Strack, F. (2009). The architecture of intuition: Fluency and affect determine intuitive judgments of semantic and visual coherence and judgments of grammaticality in artificial grammar learning. Journal of Experimental Psychology: General, 138(1), 39–63. https://doi.org/10.1037/a0014678

Tversky, A., & Kahneman, D. (1983). Extensional versus intuitive reasoning: The conjunction fallacy in probability judgment. Psychological Review, 90(4), 293-315. https://doi.org/10.1037/0033-295X.90.4.293

Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124-1131. https://doi.org/10.1126/science.185.4157.1124

Volz, K. G., & von Cramon, D. Y. (2006). What neuroscience can tell about intuitive processes in the context of perceptual discovery. Journal of Cognitive Neuroscience, 18(12), 2077-2087. https://doi.org/10.1162/jocn.2006.18.12.2077

Wan, X., Nakatani, H., Ueno, K., Asamizuya, T., Cheng, K., & Tanaka, K. (2011). The neural basis of intuitive best next-move generation in board game experts. Science, 331(6015), 341-346. https://doi.org/10.1126/science.1194732