Leon Festinger, James M. Carlsmith · 1957–1959

Festinger's $1 and $20 experiment

If you are talked into saying something you don't believe, does your belief quietly move to match your words, and does a small reward make that more likely than a big one?

  • Stanford University, California
  • Contested

The question

In 1957 Leon Festinger proposed a simple idea with large consequences. When two of our thoughts clash, such as “I believe X” and “I just said the opposite of X”, we feel an uncomfortable tension he called cognitive dissonance, and we are motivated to reduce it. One way is to change the belief.

The theory made a prediction that ran against common sense and against the learning theories of the day. If you are paid to say something you don’t believe, a big payment gives you an easy explanation: you did it for the money. A small payment does not. So, Festinger reasoned, the smaller the reward that is just enough to get you to say it, the more your private opinion should drift toward what you said. Festinger and his student James Carlsmith set out to test this.

How it was done

The participants were 71 male students from the introductory psychology course at Stanford, who had to take part in experiments as part of the course. They signed up for a two-hour study listed as “Measures of Performance”.

Each man first spent an hour on two deliberately tedious tasks, working with one hand at his own pace:

The experimenter appeared to time and note everything, so the tasks looked like the whole study. Then came the real manipulation. The experimenter explained that the study was about how expectations affect performance and that some participants were told in advance that the tasks were enjoyable. In the two paid conditions, he added that the assistant who usually delivered this message had not come in, and asked the participant to step in, telling the next student, who was waiting outside, that the tasks were fun and interesting.

The “waiting student” was an undergraduate woman hired by the researchers. She mentioned that a friend had found the experiment boring, and most participants then tried to reassure her. Their conversation was recorded without their knowledge.

Finally, each participant was sent to what seemed like a routine departmental interview about experiments in general. The interviewer did not know which condition he was in. Among other questions, he asked how enjoyable the tasks had been, on a scale from −5 to +5.

Eleven of the 71 men were excluded from the analysis: five suspected the real purpose, two told the woman the truth, three refused the money, and one asked her for her phone number. That left 20 per condition.

What came out

On the key question, the results went the way dissonance theory predicted:

The difference between the $1 group and each of the other two groups was statistically significant. The $1 group was also more willing to take part in a similar experiment (+1.20, against −0.62 for controls and −0.25 for the $20 group), though these differences were weaker and short of conventional significance (p = .08 against the control group, p = .15 against the $20 group). There were no meaningful differences on how much participants felt they had learned.

The researchers also checked a simpler explanation: that the $1 men had argued more persuasively and so convinced themselves. Two raters scored the recorded conversations, and if anything, the $20 men had been slightly more convincing. Festinger and Carlsmith concluded the results backed the theory.

What really happened (replications and critiques)

Methodological objections. In 1964 N. P. and A. Chapanis published a critical review of the first years of dissonance research in Psychological Bulletin. In 1966 the psychologist Marvin Dunnette singled out their long discussion of one practice: dropping participants from the analysis, which in this study meant 11 of 71. Joel Cooper, a leading dissonance researcher, notes that this wave of criticism ended up prompting new studies, many of which supported the theory.

A rival explanation. In 1967 Daryl Bem proposed self-perception theory: we often work out our own attitudes the way an outsider would, by looking at what we did and why. To test this, he had 75 students listen to a description of the study and a recorded conversation in which a student praised the dull tasks, and then guess how much that student really enjoyed them. Listeners told the student had been paid $1 guessed he liked the tasks more (+0.52) than listeners told he had been paid $20 (−1.96). The observers had no inner tension to reduce, yet they reproduced the pattern. Bem argued that no special discomfort was needed to explain the result.

Narrowing the effect. Later work found that the “less pay, more liking” effect is conditional. According to Cooper, a 1967 study by Linder, Cooper and Jones found it only when people felt free to decline the task. In a 1974 study by Zanna and Cooper, participants given a pill they were told would make them tense did not change their attitudes after writing against their views, as if they blamed the pill for their discomfort. That result is hard for pure self-perception to explain. A 1977 paper by Fazio, Zanna and Cooper suggested that both theories are right in different territory: dissonance when people act clearly against what they believe, self-perception when the step is small.

The big test. In 2024 a team led by David Vaidis and Willem Sleegers published a Registered Replication Report: the method and analysis were reviewed and accepted before the data were collected. They did not repeat the $1/$20 study. Instead, they reran a standard version of the same idea, the induced-compliance paradigm, based on a 1983 experiment by Croyle and Cooper. Students wrote an essay in favour of raising tuition fees at their own university, a position most opposed (two labs used other unpopular proposals), either after being told they were free to refuse (high choice) or simply being told to write it (low choice). A third group wrote a neutral essay.

The project recruited 4,898 participants in 39 labs across 19 countries and 14 languages. Two findings stood out:

The authors concluded that this classic paradigm may not provide the robust evidence for dissonance that textbooks assume. Their data still leave open what produced the small shift: participants who wrote against their views did report feeling more uncomfortable and conflicted, but that discomfort did not statistically explain the attitude change.

The first case. Before this experiment, Festinger had studied dissonance outside the lab. In When Prophecy Fails (1956), he, Riecken and Schachter followed a small group whose prediction of the end of the world did not come true. Cooper describes it as arguably the first published report of dissonance. It was a field observation, not an experiment.

Ethics

The study was built on layers of deception. Participants were misled about the purpose of the tasks, about the missing assistant, about the “routine” interview, and about the woman outside, who was working for the researchers. They were also asked to deceive her, so each man believed he was lying to a fellow student for money, and the conversation was secretly recorded.

On the other side, the risk was low, and Festinger and Carlsmith explained the true purpose to every participant at the end, including the reasons for each step. They also asked all paid participants to return the money, and report that all did so. Today, a study like this would likely need clearer consent wording, and asking people to give back payment for a task they had completed would likely be questioned. (This is our assessment, not an ethics committee’s.)

What it means today

The idea that we adjust our beliefs to fit what we have said or done remains famous, and the 2024 replication suggests a small version of it is real. What is less certain is the mechanism and the conditions, especially the claim that free choice and small rewards are what make the effect work.

Some everyday reflections follow, though they go beyond what these experiments tested:

Would you have passed?

Coming soon. A short reflection, not a test: picture an hour of turning pegs, then being asked to tell a stranger it was fun. Would you have agreed? For $1? For $20? At the end you’ll see how the Stanford students rated the task in each group. One honest caveat: imagining the situation is not the same as being in it, and people are often poor at predicting their own reactions.

Questions people ask

What is the $1 vs $20 experiment?

In 1959 Leon Festinger and James Carlsmith had Stanford students do an hour of dull tasks, then paid some $1 and others $20 to tell the next participant it had been enjoyable. Afterwards, the men paid $1 rated the task as more enjoyable than either the $20 group or a control group.

Why did the people paid $1 like the task more?

Festinger's explanation was cognitive dissonance: saying something you don't believe is uncomfortable, and $20 gave a good excuse while $1 did not, so the $1 group eased the tension by deciding the task wasn't so bad. Daryl Bem argued instead that people simply infer their own attitudes from their behaviour, as an outside observer would.

Has cognitive dissonance been replicated?

Partly. A 2024 Registered Replication Report across 39 labs in 19 countries found that writing an essay against one's own view shifted attitudes slightly. But it did not find the effect of free choice that the theory predicts. The original $1/$20 design itself was not repeated.

Was the Festinger and Carlsmith experiment ethical?

It relied on deception at several levels, and participants were asked to mislead someone they believed was another volunteer. Everyone was told the true purpose afterwards, and all paid participants were asked to give the money back, which they did.

Sources

  1. Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced compliance. Journal of Abnormal and Social Psychology, 58(2), 203–210. doi:10.1037/h0041593
  2. Festinger, L. (1957). A theory of cognitive dissonance. Row, Peterson.
  3. Bem, D. J. (1967). Self-perception: An alternative interpretation of cognitive dissonance phenomena. Psychological Review, 74(3), 183–200. web.mit.edu/curhan/www/docs/Articles/15341_Readings/Motivation/Bem_1967_Self_perception.pdf
  4. Chapanis, N. P., & Chapanis, A. (1964). Cognitive dissonance: Five years later. Psychological Bulletin, 61(1), 1–22. doi:10.1037/h0043457
  5. Dunnette, M. D. (1966). Fads, fashions, and folderol in psychology. American Psychologist, 21(4), 343–352. doi:10.1037/h0023535
  6. Linder, D. E., Cooper, J., & Jones, E. E. (1967). Decision freedom as a determinant of the role of incentive magnitude in attitude change. Journal of Personality and Social Psychology, 6(3), 245–254. doi:10.1037/h0021220
  7. Zanna, M. P., & Cooper, J. (1974). Dissonance and the pill: An attribution approach to studying the arousal properties of dissonance. Journal of Personality and Social Psychology, 29(5), 703–709. doi:10.1037/h0036651
  8. Fazio, R. H., Zanna, M. P., & Cooper, J. (1977). Dissonance and self-perception: An integrative view of each theory's proper domain of application. Journal of Experimental Social Psychology, 13(5), 464–479. doi:10.1016/0022-1031(77)90031-2
  9. Cooper, J. (2019). Cognitive dissonance: Where we've been and where we're going. International Review of Social Psychology, 32(1), 7. doi:10.5334/irsp.277
  10. Croyle, R. T., & Cooper, J. (1983). Dissonance arousal: Physiological evidence. Journal of Personality and Social Psychology, 45(4), 782–791. doi:10.1037/0022-3514.45.4.782
  11. Vaidis, D. C., Sleegers, W. W. A., van Leeuwen, F., DeMarree, K. G., Sætrevik, B., Ross, R. M., … Priolo, D. (2024). A multilab replication of the induced-compliance paradigm of cognitive dissonance. Advances in Methods and Practices in Psychological Science, 7(1), 1–26. doi:10.1177/25152459231213375
  12. Association for Psychological Science (2024, February 14). Multilab replication challenges long-held theories on cognitive dissonance. APS Observer. www.psychologicalscience.org/observer/multilab-replication-cognitive-dissonance