Although it might not have surprised you to hear that we can often predict people’s behaviors if we know their thoughts and their feelings about the attitude object, you might be surprised to find that our actions also have an influence on our thoughts and feelings. It makes sense that if I like strawberry jam, I’ll buy it, because my thoughts and feelings about a product influence my behavior. But will my attitudes toward orange marmalade become more positive if I decide—for whatever reason—to buy it instead of jam?

It turns out that if we engage in a behavior, and particularly one that we had not expected that we would have, our thoughts and feelings toward that behavior are likely to change. This might not seem intuitive, but it represents another example of how the principles of social psychology—in this case, the principle of attitude consistency—lead us to make predictions that wouldn’t otherwise be that obvious.

Imagine that one Tuesday evening in the middle of the semester you see your friend Joachim. He’s just finished his dinner and tells you that he’s planning to head home to study and work on a term paper. When you see him the next day, however, he seems a bit shaken. It turns out that instead of going home to study, Joachim spent the entire evening listening to music at a rock club in town. He says that he had a great time, stayed up late to watch the last set, and didn’t get home until the crack of dawn. And he woke up so late this morning that he missed his first two classes.

You might imagine that Joachim might be feeling some uncertainty and perhaps some regret about his unexpected behavior the night before. Although he knows that it is important to study and to get to his classes on time, he nevertheless realizes that, at least in this case, he neglected his schoolwork in favor of another activity. Joachim seems to be wondering why he, who knows how important school is, engaged in this behavior after he promised himself that he was going home to study. Let’s see if we can use the principles of attitude consistency to help us understand how Joachim might respond to his unexpected behavior and how his attitudes toward listening to music and studying might follow from it.

Self-Perception Involves Inferring Our Beliefs from Our Behaviors

People have an avid interest in understanding the causes of behavior, both theirs and others, and doing so helps us meet the important goals of other-concern and self-concern. If we can better understand how and why the other people around us act the way they do, then we will have a better chance of avoiding harm from others and a better chance of getting those other people to cooperate with and like us. And if we have a better idea of understanding the causes of our own behavior, we can better work to keep that behavior in line with our preferred plans and goals.

In some cases, people may be unsure about their attitudes toward different attitude objects. For instance, perhaps Joachim is a bit unsure about his attitude toward schoolwork versus listening to music (and this uncertainty certainly seems to be increasing in light of his recent behavior). Might Joachim look at his own behavior to help him determine his thoughts and feelings, just as he might look at the behavior of others to understand why they act the way that they do? Self-perception occurs when we use our own behavior as a guide to help us determine our own thoughts and feelings (Bem, 1972; Olson & Stone, 2005).

We can use the principles of self-perception to help understand how Joachim is interpreting his behavior of staying out all night at the club rather than studying. When Joachim looks at this behavior, he may start to wonder why he engaged in it. One answer is that the social situation caused the behavior; that is, he might decide that the band he heard last night was so fantastic that he simply had to go hear them and could not possibly have left the club early. Blaming the situation for the behavior allows him to avoid blaming himself for it and to avoid facing the fact that he found listening to music more important than his schoolwork. But the fact that Joachim is a bit worried about his unusual behavior suggests that he, at least in part, might be starting to wonder about his own motivations.

Perhaps you have experienced the effects of self-perception. Have you ever found yourself becoming more convinced about an argument you were making as you heard yourself making it? Or did you ever realize how thirsty you must have been as you quickly drank a big glass of water? Research has shown that self-perception occurs regularly and in many different domains. For instance, Gary Wells and Richard Petty (1980) found that people who were asked to shake their heads up and down rather than sideways while reading arguments favoring or opposing tuition increases at their school ended up agreeing with the arguments more, and Daryl Bem (1965) found that when people were told by the experimenter to say that certain cartoons were funny, they ended up actually finding those cartoons funnier. It appears in these cases that people looked at their own behavior: if they moved their head up and down or said that the cartoons were funny, they figured that they must agree with the arguments and like the cartoon.

Creating Insufficient Justification and Overjustification

You may recall that one common finding in social psychology is that people frequently do not realize the extent to which behavior is influenced by the social situation. Although this is particularly true for the behavior of others, in some cases it may apply to understanding our own behavior as well. This means that, at least in some cases, we may believe that we have chosen to engage in a behavior for personal reasons, even though external, situational factors have actually led us to it. Consider again the children who did not play with the forbidden toy in the Aronson and Carlsmith study, even though they were given only a mild reason for not doing so. Although these children were actually led to avoid the toy by the power of the situation (they certainly would have played with it if the experimenter hadn’t told them not to), they frequently concluded that the decision was a personal choice and ended up believing that the toy was not that fun after all. When the social situation actually causes our behavior, but we do not realize that the social situation was the cause, we call the phenomenon insufficient justification. Insufficient justification occurs when the threat or reward is actually sufficient to get the person to engage in or to avoid a behavior, but the threat or reward is insufficient to allow the person to conclude that the situation caused the behavior.

Although insufficient justification leads people to like something less because they (incorrectly) infer that they did not engage in a behavior due to internal reasons, it is also possible that the opposite may occur. People may in some cases come to like a task less when they perceive that they did engage in it for external reasons. Overjustification occurs when we view our behavior as caused by the situation, leading us to discount the extent to which our behavior was actually caused by our own interest in it (Deci, Koestner, & Ryan, 1999; Lepper & Greene, 1978).

Mark Lepper and his colleagues (Lepper, Greene, & Nisbett, 1973) studied the overjustification phenomenon by leading some children to think that they engaged in an activity for a reward rather than because they simply enjoyed it. First, they placed some fun felt-tipped markers into the classroom of the children they were studying. The children loved the markers and played with them right away. Then, the markers were taken out of the classroom and the children were given a chance to play with the markers individually at an experimental session with the researcher. At the research session, the children were randomly assigned to one of three experimental groups. One group of children (the expected reward condition) was told that if they played with the markers they would receive a good-drawing award. A second group (the unexpected reward condition) also played with the markers and got the award—but they were not told ahead of time that they would be receiving the award (it came as a surprise after the session). The third group (the no reward condition) played with the markers too but got no award.

Then, the researchers placed the markers back in the classroom and observed how much the children in each of the three groups played with them. The results are shown in Figure 4.9, “Undermining Initial Interest in an Activity.” The fascinating result was that the children who had been led to expect a reward for playing with the markers during the experimental session played with the markers less at the second session than they had at the first session. Expecting to receive the award at the session had undermined their initial interest in the markers.

Figure 4.9 Undermining Initial Interest in an Activity. Children who had been expecting to receive a reward when they played with the fun markers played less with them in their free play period than did children who received no reward or an unexpected reward—their initial interest had been undermined by the expected reward. Data are from Lepper, Greene, and Nisbett (1973). Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Undermining children’s intrinsic interest with extrinsic reward: A test of the “overjustification” hypothesis. Journal of Personality and Social Psychology, 28, 129–137.

Although this might not seem logical at first, it is exactly what is expected on the basis of the principle of overjustification. When the children had to choose whether to play with the markers when the markers reappeared in the classroom, they based their decision on their own prior behavior. The children in the no reward condition group and the children in the unexpected reward condition group realized that they played with the markers because they liked them. Children in the expected award condition group, however, remembered that they were promised a reward for the activity before they played with the markers the last time. These children were more likely to infer that they play with the markers mostly for the external reward, and because they did not expect to get any reward for playing with the markers in the classroom, they discounted the possibility that they enjoyed playing the markers because they liked them. As a result, they played less frequently with the markers compared with the children in the other groups.

The experimental manipulation involved the amount of money the students were paid to tell the lie. Half of the students were offered a large payment ($20) for telling the lie, whereas the other half were offered only a small payment ($1) for telling the lie. After the participants had told the lie, an interviewer asked each of them how much they had enjoyed the task they had performed earlier in the experiment. As you can see in Figure 4.10, “Employment of Task,” Festinger and Carlsmith found that the students who had been paid $20 for saying the tasks had been enjoyable rated the task as very boring, which indeed it was. In contrast, the students who were paid only $1 for telling the lie changed their attitude toward the task and rated it as significantly more interesting.

Festinger explained the results of this study in terms of consistency and inconsistency among cognitions. He hypothesized that some thoughts might be dissonant, in the sense that they made us feel uncomfortable, while other thoughts were more consonant, in the sense that they made us feel good. He argued that people may feel an uncomfortable state (which he called cognitive dissonance) when they have many dissonant thoughts—for instance, between the idea that (a) they are smart and decent people and (b) they nevertheless told a lie to another student for only a small payment.

Festinger argued that the people in his experiment who had been induced to lie for only $1 experienced more cognitive dissonance than the people who were paid $20 because the latter group had a strong external justification for having done it whereas the former did not. The people in the $1 condition, Festinger argued, needed to convince themselves that that the task was actually interesting to reduce the dissonance they were experiencing.

Figure 4.10 Festinger and Carlsmith. Participants who had engaged in a boring task and then told another student it was interesting experienced cognitive dissonance, leading them to rate the task more positively in comparison to those who were paid $20 to do the same. Data are from Festinger and Carlsmith (1959). Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced compliance. Journal of Abnormal and Social Psychology, 58, 203–210.

Although originally considered in terms of the inconsistency among different cognitions, Festinger’s theory has also been applied to the negative feelings that we experience when there is inconsistency between our attitudes and our behavior, and particularly when the behavior threatens our perceptions of ourselves as good people (Aronson, 1969). Thus Joachim is likely feeling cognitive dissonance because he has acted against his better judgment and these behaviors are having some real consequences for him. The dissonant thoughts involve (a) his perception of himself as a hardworking student, compared with (b) his recent behaviors that do not support that idea. Our expectation is that Joachim will not enjoy these negative feelings and will attempt to get rid of them.

We Reduce Dissonance by Decreasing Dissonant or by Increasing Consonant Cognitions

Because Joachim’s perception of himself as a hardworking student is now in jeopardy, he is feeling cognitive dissonance and will naturally try to reduce these negative emotions. He can do so in a number of ways. One possibility is that Joachim could simply change his behavior by starting to study more and go out less. If he is successful in doing this, his dissonance will clearly be reduced and he can again feel good about himself. But it seems that he has not been very successful in this regard—over the past weeks he has continually put off studying for listening to music. A second option is to attempt to reduce his dissonant cognitions—those that threaten his self-esteem. Perhaps he might try to convince himself that he has failed only one test and that he didn’t expect to do very well in biology anyway. If he can make the negative behaviors seem less important, dissonance will be reduced.

But Joachim has a third option: even if he cannot change his behavior and even if he knows that what he’s doing has negative consequences, he can create new consonant cognitions to counteract the dissonant cognitions. For instance, Joachim might try to convince himself that he is going to become an important record producer some day and that it is therefore essential that he attend many concerts. When Joachim takes this route he changes his beliefs to be more in line with his behavior, and the outcome is that he has now restored attitude consistency. His behaviors no longer seem as discrepant from his attitudes as they were before, and when consistency is restored, dissonance is reduced. What the principles of cognitive dissonance suggest, then, is that we may frequently spend more energy convincing ourselves that we are good people than we do thinking of ourselves accurately. Of course we do this because viewing ourselves negatively is painful.

Cognitive Dissonance in Everyday Life

Cognitive dissonance is an important social psychological principle that can explain how attitudes follow behavior in many domains of our everyday life. For instance, people who try but fail to quit smoking cigarettes naturally suffer lowered self-esteem (Gibbons, Eggleston, & Benthin, 1997). But rather than accepting this negative feeling, they frequently attempt to engage in behaviors that reduce dissonance. They may try to convince themselves that smoking is not that bad: “My grandmother smoked but lived to be 93 years old!” “I’m going to quit next year!” Or they may try to add new consonant cognitions: “Smoking is fun; it relaxes me.” You can see that these processes, although making us feel better about ourselves at least in the short run, may nevertheless have some long-term negative outcomes.

Elliot Aronson and Judson Mills (1959) studied whether the cognitive dissonance created by an initiation process could explain how much commitment students felt to a group they were part of. In their experiment, female college students volunteered to join a group that would be meeting regularly to discuss various aspects of the psychology of sex. According to random assignment, some of the women were told that they would be required to perform an embarrassing procedure before they could join the group (they were asked to read some obscene words and some sexually oriented passages from a novel in public), whereas other women did not have to go through this initiation. Then all the women got a chance to listen to the group’s conversation, which turned out to be very boring.

Aronson and Mills found that the women who had gone through the embarrassing experience subsequently reported more liking for the group than those who had not, and Gerard and Matthewson (1966) found that having to take some electrical shocks as part of an initiation process had the same effect. Aronson and Mills argued that the more effort an individual expends to become a member of the group (e.g., a severe initiation), the more he or she will become committed to the group in order to justify the effort put in during the initiation. The idea is that the effort creates dissonant cognitions (e.g., “I did all this work to join the group”), which are then justified by creating more consonant ones (e.g., “Okay, this group is really pretty fun”). The women who spent little effort to get into the group were able to see the group as the dull and boring conversation that it was. The women who went through the more severe initiation, however, succeeded in convincing themselves that the same discussion was a worthwhile experience. When we put in effort for something—an initiation, a big purchase price, or even some of our precious time—we will likely end up liking the activity more than we would have if the effort had been less. Even the effort of having to fill out a purchase agreement for a product, rather than having the salesperson do it for you, creates commitment to the purchase and a greater likelihood of staying in the deal (Cialdini, 2001).

Another time you may have experienced the negative affective state of cognitive dissonance is after you have made an important and irrevocable decision. Imagine that you are about to buy a new car and you have narrowed your search to a small new car and a larger (but much cheaper) used car. The problem is that you can see advantages and disadvantages to each. For instance, the smaller car would get better gas mileage, but the larger car—because it is used—is cheaper. Imagine, however, that you finally decide to buy the larger car because you feel that you really don’t have enough money for the new car.

That night, you’re lying in bed and wondering about your decision. Although you’ve enjoyed driving the big car that you have just purchased, you’re worried about rising gas costs, the negative impact of the big car on the environment, and the possibility that the car might need a lot of repairs. Have you made the right decision? This “buyer’s remorse” can be interpreted in terms of postdecisional dissonance—the feeling of regret that may occur after we make an important decision (Brehm, 1956). However, the principles of dissonance predict that once you make the decision—and regardless of which car you choose—you will convince yourself that you made the right choice. Since you have chosen the larger car, you will likely begin to think more about the positive aspects of the choice that you have made (what you are going to be able to do with the money you saved, rather than how much more it is going to cost to fill up the gas tank), and at the same time you will likely downplay the values of the smaller car.

Jack Brehm (1956) posed as a representative of a consumer testing service and asked women to rate the attractiveness and desirability of several kinds of appliances, such as toasters and electric coffee makers. Each woman was told that as a reward for having participated in the survey, she could have one of the appliances as a gift. She was given a choice between two of the products she had rated as being about equally attractive. After she made her decision, her appliance was wrapped up and given to her. Then, 20 minutes later, each woman was asked to re-rate all the products. As you can see in Figure 4.11, “Postdecisional Dissonance,” Brehm found that the women rated the appliance that they had chosen and been given as a gift higher than they had the first time. And the women also lowered their rating of the appliance they might have chosen but decided to reject. These results are of course consistent with the principles of cognitive dissonance—postdecisional dissonance is reduced by focusing on the positive aspects of the chosen product and the negative aspects of the rejected product.

Figure 4.11 Postdecisional Dissonance. As predicted by the desire to reduce postdecisional dissonance, participants increased the perceived desirability of a product they had chosen and decreased the perceived desirability of a product they did not choose. Data are from Brehm (1956). Brehm, J. W. (1956). Postdecision changes in the desirability of alternatives. Journal of Abnormal and Social Psychology, 52(3), 384–389.

What research on cognitive dissonance suggests, then, is that people who are experiencing dissonance will generally try to reduce it. If we fail to lose the weight we wanted to lose, we decide that we look good anyway. If we cheat on an exam, we decide that cheating is okay or common. If we hurt other people’s feelings, we may even decide that they are bad people who deserve our negative behavior. To escape from feeling poorly about themselves, people will engage in quite extraordinary rationalizing. No wonder that most of us believe the statement, “If I had it all to do over again, I would not change anything important.”

As you can see in Figure 4.12, “Spread of Alternatives by Culture and Feedback Condition,” the researchers found a significant interaction between culture and personality feedback. The pattern of means showed that the feedback mattered for the Canadian participants—the difference in the ratings of the chosen versus the rejected CD (the “spread of alternatives”) increased from the positive to the control to the negative feedback conditions. However, there was no significant simple effect of feedback for the Japanese students, nor did they show a significant spread of alternatives in any feedback condition.

Figure 4.12 Spread of Alternatives by Culture and Feedback Condition. The Canadian participants showed a greater spread of alternatives when their self-esteem was threatened, but Japanese participants did not. Data are from Heine and Lehman (1997). Heine, S. J., & Lehman, D. R. (1997). Culture, dissonance, and self-affirmation. Personality and Social Psychology Bulletin, 23, 389-400. doi:10.1177/0146167297234005.

However, other researchers have found that individuals from collectivistic cultures do show dissonance effects when they are focused on their relationships with others. For instance, Kitayama, Snibbe, Markus, and Suzuki (2004) found that East Asian participants experienced dissonance particularly when they were asked to think about a close friend who had made a dissonance-creating decision. Such a result would be expected because behaviors that involve more other-oriented, collectivistic outcomes should be more important for these people. Indeed, research has found that advertisements that are framed in terms of personal benefits (e.g., “Use this breath mint!”) are more persuasive in individualistic cultures, whereas ads that emphasize family or ingroup benefits (e.g., “Share this breath mint with your friends!”) are more persuasive in collectivistic cultures (Han & Shavitt, 1994).