필사 모드: The Experiment Where One Dollar Beat Twenty — Cognitive Dissonance and Sixty Years of Rebuttal

Introduction — The People Paid Less Said They Had More Fun

In the era when behaviorism ruled, the common sense was simple. Bigger reward, bigger attitude change. The size of the reinforcement sets the size of the learning.

In 1959, Leon Festinger and James Carlsmith at Stanford reported exactly the opposite. They had participants perform an obviously boring task, then had them lie that it had been fun, paying one group one dollar and the other group twenty dollars. The people who later rated the task as genuinely enjoyable were not the ones who had received twenty dollars but the ones who had received one dollar.

Out of this result came the concept of cognitive dissonance, and by now the idea that "people rewrite their beliefs to match their behavior" is common sense. This installment of Psychology, Straight from the Papers opens the original behind that common sense and follows how far the objections raised over sixty years have gone. To give the conclusion first: the phenomenon has largely held up, and the mechanism is still unsettled.

What the 1959 Experiment Actually Did

The participants were 71 students in the introductory psychology course at Stanford. All of them were male, and they took part as a course requirement.

The task was designed to be meaningless on purpose. For the first 30 minutes, participants used one hand to load twelve spools onto a tray, empty it, and load it again, over and over. For the next 30 minutes, they turned each of 48 square pegs on a board a quarter turn clockwise, one at a time. On reaching the end they went back to the start and turned them again. Boredom itself was the experimental stimulus.

When the hour was up, the experimenter said, with an embarrassed look, that the assistant who normally tells the next participant in advance that the experiment is interesting had not shown up today, and asked whether the participant might do it instead. Here the conditions split. One group was offered one dollar for the favor, the other twenty dollars. A control group finished the task without the request and went home.

Participants who agreed told a female student in the waiting room, posing as the next participant, that what they had just done was interesting and fun. Then, in a session staged as a separate departmental interview, they rated how enjoyable the task had been on a scale from minus five to plus five. Eleven participants were excluded from the analysis for various reasons, leaving 20 per condition, 60 in total.

How you read this matters. The twenty dollar group is effectively indistinguishable from the control group. They told the lie and their attitude did not move. The only group that moved was the one dollar group, and that difference reached a p value below 0.02 against the control group and below 0.03 against the twenty dollar group.

There are honest additions to make. Twenty people per condition is a small sample by present standards. The gap is roughly 1.4 points on an 11-point scale, and among the several dependent measures the paper reported, the one with a sharp difference between conditions was the enjoyment rating, while the differences on the remaining measures were less pronounced. The force of this paper came not from its statistics but from the direction of its prediction, because the dominant theory of the day predicted precisely the opposite.

Why a Smaller Reward Changes Attitudes More

Two mismatched sentences sit in a participant head. "That task was boring" and "I just told someone it was fun."

The twenty dollar group has a third sentence available. "Because I got twenty dollars." In 1959, twenty dollars was by no means small money for a student. That one sentence settles the mismatch. I did it for the money, that is understandable, and so my evaluation of the task needs no adjustment. When external justification is sufficient, the attitude stays put.

The one dollar group does not have that sentence. One dollar cannot account for their own behavior. Nor can they take back what they have already said. Only one card remains: touching up the memory of the task. "Come to think of it, it was not that boring." This is insufficient justification. When unchangeable behavior collides with a changeable attitude, the side that loses is always the attitude.

The extended version of the same logic is the overjustification effect. In 1973, Lepper, Greene, and Nisbett told children who enjoyed drawing that they would receive a certificate for drawing. Once the award disappeared, those children drew less than they had before. External reward had crowded out the internal account, "I do this because I like it." From here comes the practical implication that you must design rewards carefully if you want to protect intrinsic motivation.

The Observation That Came Before — When Prophecy Fails

The origin story of dissonance theory is set not in a laboratory but in a living room in Chicago. In their 1956 book When Prophecy Fails, Festinger, Riecken, and Schachter recorded what happened when they sent observers into a small group that had prophesied a great flood at midnight on December 21, 1955, with flying saucers coming to rescue the faithful. After the prophecy failed, the leader declared that the faith of the group had saved the world, and the group, which until then had avoided the press, began proselytizing in earnest. The story is that a belief grew stronger when it was disconfirmed.

It is a powerful narrative, but treating it as evidence is a problem. The limits are clear.

First, the observers made up a large share of the group. Researchers and hired observers entered a group whose core membership was about ten people, and their very participation may have read to the believers as proof of conviction. The record also includes a situation in which one observer was pressed to speak as though having had a vision. The observation, in effect, produced its object.

Second, there was no control group, no baseline measurement, and no quantitative indicator. It is a case study written as narrative.

Third, later studies reviewing comparable groups conclude that intensified proselytizing after a failed prophecy is not a general rule. Depending on the cohesion of the group, the response of the leader, and the amount of social support available, dissolution is if anything more common.

This book is the starting point of the theory and unsuitable material for testing it. Keeping that distinction when you cite it is the conscientious way to read.

The Rival Hypothesis — Self-Perception Theory and an Argument That Never Ended

In 1967, Daryl Bem offered a far simpler explanation. No internal tension is required. People do not look directly into their own attitudes; they infer them from their own behavior in exactly the way an outside observer would. "I said it was fun for a mere dollar. I must think it was pretty fun."

The way Bem argued is what impresses. He described the procedure of Festinger and Carlsmith in writing, then asked observers who had never gone through the experiment to guess what score the participants had given. The guesses of the observers reproduced the actual pattern of the one dollar and twenty dollar conditions. If a third party who experienced no dissonance at all can predict the result, dissonance as an internal state may not be necessary to explain it.

The counterattack was elaborate too. In 1974, Zanna and Cooper gave participants a placebo, telling some of them that the pill induced tension. Participants who could attribute their own arousal to the pill showed no attitude change. That result is hard to explain unless dissonance is an actually felt, unpleasant state of arousal.

The argument ended not in victory but in a partition of territory. In 1977, Fazio, Zanna, and Cooper proposed that the two theories govern different ranges. Behavior far outside your own attitude produces dissonance; behavior inside the latitude of acceptance is explained by self-perception. Plausible, but this is a ceasefire rather than a verdict. Sixty years on, the mechanism of cognitive dissonance remains undetermined.

And there is a part that genuinely collapsed. It is the free-choice paradigm, the other flagship procedure of dissonance research. Rank a set of objects by preference, have the person choose between two rated similarly, then have them rank again, and the rank of the chosen one rises. For half a century this was cited as evidence that choice changes preference. Then in 2010 Chen and Risen showed that this design produces the same result even with no attitude change whatsoever. The act of choosing reveals pre-existing preferences that the measurement failed to capture, so the effect is a statistical artifact. Later designs that block this hole do still observe a real effect, but a smaller one. A good share of the free-choice paradigm examples in popular books need rereading in light of 2010.

Effort Justification — Why We Come to Like What Was Hard to Enter

The branch of dissonance theory that survived best is effort justification.

In 1959, Aronson and Mills had 63 female college students apply to a discussion group on the psychology of sex. Under the pretext of a screening for admission, some were made to read aloud embarrassingly explicit words and passages (severe condition), some read only mild words (mild condition), and the rest went through no procedure at all. All three groups then listened to the same recording: a deliberately dull discussion of sexual behavior in animals. The group that rated that dull discussion most favorably was the one that had passed through the severe procedure.

There is also a check that the result was not about embarrassment or sexual arousal. In 1966, Gerard and Mathewson repeated the same structure with electric shocks instead, and the result was the same. Whatever the price, we come to like more the things we paid dearly for.

Once you know the structure, your surroundings look different. Long onboarding, rites of passage, communities with high barriers to entry, hobbies with expensive gear. Part of the attachment they generate comes not from the value of the object but from the cost you paid. That said, there is a rival explanation here too. As with the so-called IKEA effect, where people rate furniture they assembled themselves more highly, there is a route by which effort creates attachment without any dissonance.

It is related to sunk cost but not the same. In 1985, Arkes and Blumer randomly applied full price, a two dollar discount, or a seven dollar discount to 60 buyers of season tickets at the Ohio University theater. The group that paid full price attended significantly more performances in the first half of the season. Effort justification is an attitude change in which "you come to like the object," while sunk cost is a behavioral persistence in which "you keep spending because of what you already spent." Both arise from the pressure to justify what has already been paid, but the sunk cost side is largely explained without invoking cognitive dissonance at all. Loss aversion plus a norm against waste is enough.

What to Discard and What to Keep

Discard 1. Using cognitive dissonance as a verified law. The phenomenon is fairly robust, but the mechanism has been unsettled for sixty years, and the evidence in the free-choice paradigm family needs reinterpretation.

Discard 2. Citing When Prophecy Fails as evidence. It is a participant observation record with no control and no baseline measurement, and there are traces of observers influencing their subject. Introducing something as the origin of a theory is different from presenting it as grounds for one.

Discard 3. The management prescription that "pay them less and they will commit more". The one dollar in the laboratory rested on a very narrow set of conditions: voluntary choice, a trivial lie, immediate measurement. Carrying this result over into a rationale for low pay is an abuse of the original paper. Low pay does not produce dissonance; it produces turnover.

Keep 1. Notice the moment you are defending yourself. The signal is clear: when you are manufacturing fresh grounds for a decision you have already made, especially when those grounds occurred to you only after the decision. The voice gets particularly loud after a technology stack choice, a job change, or a large purchase.

Keep 2. Write down your reasons before deciding. This is nearly the only practical defense against dissonance. Post-decision justification is not by itself distinguishable from a real reason, but held against a pre-decision record, its identity is exposed. This is what an architecture decision record or a pre-mortem document actually does.

Keep 3. To learn what you like, watch what you do when there is no reward. Even if Bem was only half right, this observation method holds. A record of behavior gives you more accurate information about your own taste than self-report does.

Keep 4. Audit the source of your attachment. If you are rating highly an organization that was hard to get into, a project you have held on to for a long time, or equipment you bought at a high price, it is worth asking once. Did this evaluation come from the object, or from the cost I paid.

One final dividing line. Affirming what has happened and the inertia of justifying what has happened look similar from the outside, because both speak well of the past. The criterion is direction. If the narrative widens your future options, it is acceptance; if it blocks the next choice in order to protect a choice already made, it is dissonance.

Reading Guide

• The original paper: Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced compliance. Journal of Abnormal and Social Psychology, 58(2), 203-210.
• The theory book: Festinger, L. (1957). A Theory of Cognitive Dissonance. Stanford University Press.
• The observation record: Festinger, L., Riecken, H. W., & Schachter, S. (1956). When Prophecy Fails. University of Minnesota Press.
• The rival theory: Bem, D. J. (1967). Self-perception: An alternative interpretation of cognitive dissonance phenomena. Psychological Review, 74(3), 183-200.
• Evidence for arousal: 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.
• Effort justification: Aronson, E., & Mills, J. (1959). The effect of severity of initiation on liking for a group. Journal of Abnormal and Social Psychology, 59(2), 177-181.
• The hole in the free-choice paradigm: Chen, M. K., & Risen, J. L. (2010). How choice affects and reflects preferences: Revisiting the free-choice paradigm. Journal of Personality and Social Psychology, 99(4), 573-594.
• Sunk cost: Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes, 35(1), 124-140.

Reading tip: the 1959 paper runs eight pages of main text in easy prose, which makes it one of the lower barriers to entry in this series. If you are short on time, look at the results table alone. In that single table holding three conditions and four items, reading just the three numbers in the enjoyment row from left to right gives you the entire claim of the paper. The fact that the twenty dollar column is nearly identical to the control column is the real finding of this study. The next installment is the bystander effect, where we test a concept created by a single newspaper article.