The big ideas
Heuristics and biases
Mixed evidencePeople estimate probabilities and quantities with a few shortcuts, such as representativeness, availability and anchoring. They usually work, but they produce errors that are systematic rather than random.
Prospect theory
SupportedPeople judge outcomes as gains or losses from a reference point, feel losses more than equal gains, and give too much weight to small probabilities. Developed with Tversky in 1979.
Framing effects
SupportedSmall changes in how the same options are described can shift what people choose, even when the facts are identical.
Two systems of thinking
Mixed evidenceIn Thinking, Fast and Slow (2011), fast intuitive thinking (System 1) and slow effortful thinking (System 2) are presented as two characters in the mind; Kahneman said they were a figure of speech, not brain structures.
Noise
Mixed evidenceWith Olivier Sibony and Cass R. Sunstein (2021), he argued that unwanted variability between judgments that should agree is a source of error as important as bias, and often ignored.
In brief
Daniel Kahneman spent his career on a simple-sounding question: when people judge and choose under uncertainty, how do they actually do it? His answer, built mostly with his friend Amos Tversky, was that the mind uses a few quick shortcuts. They usually serve us well, but they fail in ways that are predictable, so that many people make the same mistake in the same direction.
He was born in Tel Aviv in 1934, while his mother was visiting relatives, and grew up in Paris, where his Lithuanian-born parents had settled. During the German occupation his father was held for a time in the Drancy internment camp; after his release the family fled south and kept moving to stay ahead of the Germans. His father died in 1944 of diabetes that could not be properly treated. After the war the family moved to Palestine. Kahneman studied psychology and mathematics at the Hebrew University, then served in the Israeli army’s psychology unit, where he designed a structured interview system for assessing recruits to combat units. He took his doctorate at Berkeley in 1961 and taught at the Hebrew University, the University of British Columbia and Berkeley before moving to Princeton in 1993.
Amos Tversky (1937–1996) was born in Haifa, served as a paratroop officer and was decorated for bravery, and earned his doctorate at the University of Michigan before settling at Stanford. The two began working together in Jerusalem in the late 1960s and wrote their best-known papers jointly. Tversky died of metastatic melanoma in 1996, aged 59. When Kahneman received the Nobel Memorial Prize in Economic Sciences in 2002, shared with Vernon L. Smith, the prize could not go to Tversky, because it is not awarded after death. Accepting another award in 2011, he stressed that the honoured work was “done in collaboration with Amos Tversky.”
Kahneman died on 27 March 2024, aged 90. The Wall Street Journal later reported that he had chosen an assisted death in Switzerland.
The big ideas
Heuristics and biases. In their 1974 paper in Science, Tversky and Kahneman described three shortcuts. With representativeness, we judge how likely something is by how much it resembles a typical case. With availability, we judge frequency by how easily examples come to mind. With anchoring, we start from a number at hand and adjust too little (see the anchoring experiments). Each shortcut predicts a specific, testable error.
Prospect theory. Economics assumed that people weigh final wealth and probabilities sensibly. Kahneman and Tversky’s 1979 paper in Econometrica proposed instead that people value changes, gains and losses measured from a reference point; that a loss hurts more than an equal gain pleases; that people tend to avoid risk for gains but seek it to avoid losses; and that small probabilities get too much weight.
Framing. Tversky and Kahneman also showed that small changes in how the same facts are presented shift decisions. Working with Barbara McNeil of Harvard, Tversky found that people judged the same medical procedure differently depending on whether its risk was stated as a chance of dying or a chance of living. According to McNeil and the Toronto physician Donald Redelmeier, the framing work influenced how doctors think about informed consent.
Two systems. Thinking, Fast and Slow (2011) brought this work to a general audience through two characters: System 1, fast and automatic, and System 2, slow and effortful. Noise (2021), written with Olivier Sibony and Cass R. Sunstein, turned to a different kind of error: the scatter between judgments that ought to agree. Its core point is statistical: the overall error of a set of judgments combines their shared bias and their random spread, so reducing the spread can matter as much as correcting the bias.
What holds up, and what doesn’t
Prospect theory holds up well. In 2020, a team led by Kai Ruggeri re-ran the 1979 questions with about 4,100 people in 19 countries and 13 languages. In the pooled data, 16 of the 17 original choice patterns replicated, although most effects were smaller than in 1979. Loss aversion in particular is debated: David Gal and Derek Rucker argued in 2018 that the evidence does not show losses generally weigh more than gains, while Kellen Mrkva, Eric Johnson and colleagues, with more than 17,000 participants, found loss aversion at all levels of knowledge and experience, though weaker in people who knew more about what they were choosing.
Heuristics and biases: solid effects, disputed meaning. Some individual effects replicate very strongly, anchoring among them. The interpretation is another matter. Gerd Gigerenzer argued that the programme judged people against overly narrow rules of logic and that its heuristics were too vague to predict anything in advance; Kahneman and Tversky replied in 1996 that he misrepresented their position and that the errors were real.
The priming chapter did not hold up. Chapter 4 of Thinking, Fast and Slow presented social priming, the idea that subtle cues such as words about old age change behaviour, as something readers had to accept. By 2012 several labs reported that they could not reproduce well-known priming effects, and that year Kahneman wrote to priming researchers warning of a “train wreck looming” and proposed a chain of labs replicating one another. In 2017, Ulrich Schimmack, Moritz Heene and Kamini Kesavan analysed the 31 studies behind the chapter and concluded that their evidence was very weak. Kahneman answered on their blog: “I placed too much faith in underpowered studies.” He added that he still thought actions could be primed, but that the effects could not be as large or robust as his chapter suggested. This public correction is often cited in discussions of psychology’s replication crisis.
Two systems: useful shorthand. Kahneman presented System 1 and System 2 as a way of talking, not as a finding about brain anatomy (see the myth below). Noise has been received as sound but not wholly new: the statistician David Aldous found most of the book uncontroversial and familiar from earlier research, and its proposals for organisations its most original part.
The myth
In popular retellings, the two systems are often described as two parts of the brain, a fast one and a slow one. Kahneman did not claim this. In a 2011 lecture he described them as invented characters, not separate structures with a place of their own in the brain. He kept them because people, himself included, find characters easier to think about than abstract categories of processing.
Connections
Kahneman’s most important influence was Amos Tversky, a partner rather than a teacher: their best-known work was jointly written, and Tversky’s own studies of framing and of how people see patterns in random sequences belong to the same programme. Its widest influence was on economics: the 2002 prize was given for bringing insights from psychology into economic science. Critics shaped the field too, above all Gerd Gigerenzer, who pressed for precise, testable models of how heuristics work instead of broad labels.
On this site, the anchoring experiments are Tversky and Kahneman’s best-known demonstration, and their replication record mirrors the pattern above: the classic effect is very strong, some later extensions are not.
Why it matters today
The lasting lesson is that predictable errors can be planned for. Framing, anchoring and the extra pull of losses are now discussed wherever decisions are designed, from how doctors present medical odds to public policy. Kahneman’s own record adds a second lesson: even the scientist who described overconfidence could put too much faith in small studies, and could say so in public.
Two practical habits follow from his work, though they go beyond what any single experiment tested: when a decision matters, gather independent judgments before anyone shares a number, and treat a striking finding as provisional until it has been repeated.
Influence map
Learned from
- Amos Tversky
Experiments on this site
Questions people ask
What did Daniel Kahneman win the Nobel Prize for?
In 2002 he received the Nobel Memorial Prize in Economic Sciences for bringing insights from psychology into economics, above all about judgment and decision-making under uncertainty. He shared it with the economist Vernon L. Smith.
Why did Amos Tversky not share the Nobel Prize?
Tversky, Kahneman's partner in most of the work the prize recognised, died of cancer in 1996, and the prize is not awarded after death. Kahneman repeatedly credited Tversky as his partner in the work.
Did the priming studies in Thinking, Fast and Slow replicate?
Several well-known ones did not. A 2017 statistical analysis of the studies cited in the book's priming chapter judged their evidence weak, and Kahneman replied publicly that he had trusted underpowered studies too much. He still thought some priming of actions was real, but much smaller than the chapter suggested.
Is loss aversion real?
It is debated. The response patterns of the 1979 prospect theory paper replicated in 19 countries in 2020. Some researchers argue that losses do not, in general, weigh more than gains; others, with large samples, find loss aversion is real but varies with knowledge and experience.
Sources
- Princeton University, Office of Communications. (2024, March 28). Daniel Kahneman, pioneering behavioral psychologist, Nobel laureate and 'giant in the field,' dies at 90. www.princeton.edu/news/2024/03/28/daniel-kahneman-pioneering-behavioral-psychologist-nobel-laureate-and-giant-field
- Kahneman, D. (2003). Autobiography. In T. Frängsmyr (Ed.), Les Prix Nobel: The Nobel Prizes 2002. Stockholm: Nobel Foundation. www.nobelprize.org/prizes/economics/2002/kahneman/auto-biography/
- The Royal Swedish Academy of Sciences. (2002, October 9). Press release: The Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel 2002. www.nobelprize.org/prizes/economic-sciences/2002/press-release/
- Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131. doi:10.1126/science.185.4157.1124
- Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–292. doi:10.2307/1914185
- Stanford University News Service. (1996, June 5). Amos Tversky, leading decision researcher, dies at 59. news.stanford.edu/pr/96/960605tversky.html
- Kahneman, D. (2012). Two systems in the mind. Bulletin of the American Academy of Arts & Sciences (Winter 2012). Edited text of the Talcott Parsons Prize lecture, 9 November 2011. www.amacad.org/news/two-systems-mind
- Yong, E. (2012). Nobel laureate challenges psychologists to clean up their act. Nature. doi:10.1038/nature.2012.11535
- Schimmack, U., Heene, M., & Kesavan, K. (2017, February 2). Reconstruction of a train wreck: How priming research went off the rails. Replicability-Index. With a comment by D. Kahneman (2017, February 14). replicationindex.com/2017/02/02/reconstruction-of-a-train-wreck-how-priming-research-went-of-the-rails/
- Gigerenzer, G. (1996). On narrow norms and vague heuristics: A reply to Kahneman and Tversky (1996). Psychological Review, 103(3), 592–596. Reply to: Kahneman, D., & Tversky, A. (1996). On the reality of cognitive illusions. Psychological Review, 103(3), 582–591.
- Ruggeri, K., Alí, S., Berge, M. L., et al. (2020). Replicating patterns of prospect theory for decision under risk. Nature Human Behaviour, 4(6), 622–633. doi:10.1038/s41562-020-0886-x
- Gal, D., & Rucker, D. D. (2018). The loss of loss aversion: Will it loom larger than its gain? Journal of Consumer Psychology, 28(3), 497–516. Response: Mrkva, K., Johnson, E. J., Gächter, S., & Herrmann, A. (2020). Moderating loss aversion: Loss aversion has moderators, but reports of its death are greatly exaggerated. Journal of Consumer Psychology, 30(3), 407–428 (https://doi.org/10.1002/jcpy.1156). doi:10.1002/jcpy.1047
- Aldous, D. (2022). Review of Noise: A flaw in human judgment, by D. Kahneman, O. Sibony & C. R. Sunstein (Little, Brown Spark, 2021). The Mathematical Intelligencer. https://doi.org/10.1007/s00283-022-10207-9 www.stat.berkeley.edu/~aldous/Blog/Noise_review.pdf
- Zweig, J. (2025, March 14). The last decision by the world's leading thinker on decisions. The Wall Street Journal. www.wsj.com/arts-culture/books/daniel-kahneman-assisted-suicide-9fb16124