Effects of Self-Reflection on Attitudes and Consumer Decisions

Timothy D. Wilson, University of Virginia
Douglas J. Lisle, University of Virginia
Dolores Kraft, University of Virginia
ABSTRACT - Asking people to explain their attitudes has been found to lead to temporary attitude change. In the first part of this paper, the implications of this finding for attitude theory are discussed. We argue that attitudes are best viewed as the result of a constructive process, dependent on what is currently salient to people, rather than as unitary, stable constructs. In the second part of the paper, the implications of our findings for consumer decisions are discussed. We report the results of studies showing that under some circumstances, asking people to explain why they feel the way they do about the alternatives reduces the quality of their decisions.
[ to cite ]:
Timothy D. Wilson, Douglas J. Lisle, and Dolores Kraft (1990) ,"Effects of Self-Reflection on Attitudes and Consumer Decisions", in NA - Advances in Consumer Research Volume 17, eds. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT : Association for Consumer Research, Pages: 79-85.

Advances in Consumer Research Volume 17, 1990      Pages 79-85

EFFECTS OF SELF-REFLECTION ON ATTITUDES AND CONSUMER DECISIONS

Timothy D. Wilson, University of Virginia

Douglas J. Lisle, University of Virginia

Dolores Kraft, University of Virginia

[This work was supported by grant MH41841 from the National Institute of Mental Health.]

ABSTRACT -

Asking people to explain their attitudes has been found to lead to temporary attitude change. In the first part of this paper, the implications of this finding for attitude theory are discussed. We argue that attitudes are best viewed as the result of a constructive process, dependent on what is currently salient to people, rather than as unitary, stable constructs. In the second part of the paper, the implications of our findings for consumer decisions are discussed. We report the results of studies showing that under some circumstances, asking people to explain why they feel the way they do about the alternatives reduces the quality of their decisions.

Traditionally, attitudes have been viewed as evaluations that are relatively stable over time. For example, Petty and Cacioppo (1981) defined an attitude as an "enduring positive or negative feeling about some person, object, or issue" (p. 7). This can be referred to as the file drawer analogy of attitudes. That is, when people are asked how they feel about something, such as George Bush, a brand of laundry detergent, or the city of New Orleans, presumably they consult some sort of mental file. After examining the contents of the file marked, "Bush" or "Tide" or "New Orleans," they report how they feel. The contents of these "files" can be changed by personal experiences, persuasive messages, and the like, but for the most part they are enduring evaluations that remain unchanged.

In contrast, some have argued that people do not have a single, fixed attitude. Instead, people's views are said to vary depending on such factors as the context in which they are asked, the situation they are in, what else they are thinking about, and the values that are currently salient to them (e.g. Shavitt 1989; Tesser 1978; Zanna and Rempel 1988). We might refer to this as the "slide rule" analogy of attitudes, where people's views slide up and down an attitude scale from time to time. This is not to say that people are like chameleons, changing their attitudinal colors from one extreme to the other at the spur of the moment. People are unlikely to say that they love George Bush on one occasion, only to say they hate him 10 minutes later. Instead, the "slide rule" view is similar to one espoused by social judgment theory nearly 30 years ago (Sherif and Hovland 1961). People are said to have a range of positions that they view as acceptable. People will rarely, if ever endorse positions outside of this latitude of acceptance. Their attitude within this latitude of acceptance, however, might vary from time to time.

We will discuss a program of research that is consistent with this "slide rule" view of attitudes. Our view is similar to one proposed by Tesser (1978):

The standard paradigm . . . assumed that individuals have a single attitude toward the object. We have argued, however, that one's attitude at the moment is a constructive process generated from the particular schema tuned in. . . A different schema will make different cognitions salient and hence result in a different attitude. (p. 324; emphasis in original)

Viewing attitudes as a constructive process raises a key question: What determines the attitude people will have at any given point in time? Following Tesser (1978), we suggest that this depends largely on the thoughts about the attitude object that are salient at the time people are asked for their opinion (for a similar argument, see Salancik 1974 and Seligman, Fazio, and Zanna 1980). If someone has just seen a news conference in which George Bush was particularly effective, their attitude is likely to be at the positive end of their latitude of acceptance. If they have just seen him call September 7th Pearl Harbor Day, it is likely to be at the negative end.

Interestingly, it is not only external events such as news conferences that influence the availability of our thoughts about an attitude object. Sometimes our own self-reflections can influence this as well. For example, Tesser (1978) has demonstrated that simply asking people to think about the attitude object causes them to report a more extreme attitude (in the direction of their initial position) than they normally would. This polarization effect is especially likely to occur if people have a well-developed set of beliefs about the attitude object that are evaluatively consistent (Chaiken and Yates 1985; Millar and Tesser 1986). The reason for this appears to be that when such people spend time thinking about an attitude object the thoughts that are likely to come to mind are those that are consistent with their initial position. Due to the increased salience of these thoughts, people's attitudes become more extreme.

It may be, however, that other kinds of thought do not necessarily bring to mind beliefs that are consistent with one's initial position. For the past several years we have been investigating a form of self-reflection that seems to fit this bill: Asking people to think about why they feel the way they do about an issue or object. We have found that asking people to explain their attitudes can produce attitude change, sometimes in the direction of the initial attitude and sometimes not.

We have developed the following set of arguments to explain these findings. (For a more detailed discussion, see Wilson in press and Wilson, Dunn, Kraft, and Lisle 1989.) When asked to explain an attitude, people come up with reasons that are available in memory and which seem like plausible causes of their attitude. Sometimes, however, those reasons that are available and plausible are incorrect or incomplete. One reason for this is that rational cognitions about the attitude object are, in our culture at least, viewed as the most plausible causes of attitudes, and hence are most available in memory. For example, when explaining why we like various political candidates, we are more likely to call upon such factors as their stance on the issues than such seemingly implausible things as the number of times we have seen their ads on television or whether their party affiliation is the same as our parents', even though these latter factors have been shown to influence people's attitudes. Other factors might be available in memory because they are easy to verbalize or were encountered recently, as in our earlier example of seeing George Bush look particularly effective in a news conference.

As a result of these availability effects, the reasons people bring to mind might imply a somewhat different attitude than they previously held. For example, suppose that a person has a generally favorable attitude toward Candidate X. When asked to explain why she likes this candidate, however, what comes to mind is that he looked worn and ineffective during a recent debate, and that his position on the death penalty is at variance with her own. What will happen? We have found in several studies that people adopt the attitude implied by their reasons (see Wilson, Dunn, Kraft, and Lisle 1989). Thus, if we asked this person how she felt about Candidate X she would be likely to report a somewhat negative attitude, at least more negative than if she had not focused on why she felt the way she did.

Two important points should be noted about the attitude change that results from analyzing reasons. First, unlike the results of mere thought about the attitude object (Tesser 1978), it can be difficult to predict the direction of this change. The kinds of reasons that are available in memory for one person might be primarily negative, leading to change in a negative direction. The reasons that are available to another person might be primarily positive, resulting in change in the opposite direction. We found this to be the case in a recent study of political attitudes (Wilson, Kraft, and Dunn 1989). Subjects who analyzed reasons were more likely to change their attitudes toward political candidates, but this change was not in a common direction. Subjects who listed negative reasons tended to change their attitudes in a negative direction, whereas subjects who listed positive reasons tended to change in a positive direction. (Incidentally, subjects in this study, as in most of the ones we have conducted, were asked to list their reasons privately and anonymously to "organize their thoughts." Subjects believed that no one would ever see what they wrote, which reduces the plausibility of self-presentational or demand characteristic interpretations of the results.)

The second point is that analyzing reasons is most likely to cause attitude change in subjects who are relatively unknowledgeable about the attitude object (Wilson, Kraft, and Dunn 1989) or who have relatively inaccessible attitudes (Wilson and Pollack 1989). One reason for this, we suspect, is that unknowledgeable people often have more conflicting beliefs about the attitude object than do knowledgeable people (Lusk and Judd 1988; Wilson, Kraft, and Dunn 1989). As a result they should be more susceptible to the sort of availability effect we have described. That is, if a person's beliefs are entirely consistent, any belief that comes to mind will imply the same attitude. If their beliefs are inconsistent, however, it is more likely that the subset that come to mind will imply a new attitude.

We should note that this argument is the mirror image of Tesser's (1978) concerning the effects of thought on attitudes. He has demonstrated that mere thought will polarize attitudes when people have a set of consistent beliefs about the attitude object (see also Millar and Tesser 1986). In contrast, we have shown that asking people to explain their attitude causes attitude change in people with relatively inconsistent beliefs about the attitude object (Wilson, Kraft, and Dunn 1989).

To this point we have focused on the implications of our work for the definition of attitudes. We believe that people do not have a single attitude stored in a mental file that is the same over time. Like Tesser, Shavitt, and others, we believe that attitude reports involve a more constructive process, and depend in part on the thoughts and beliefs that are salient to people at any given time. What are the implications of this view? Is it an esoteric debate of interest only to attitude theorists, or does it have any practical implications? In the remainder of this paper we hope to show that there are some important implications of our work in the area of consumer behavior.

IMPLICATIONS FOR ATTITUDE BEHAVIOR CONSISTENCY

One implication of viewing attitudes as a constructive process is that people can be expected to behave consistently with their attitudes only under certain conditions, namely when the same attitude is elicited when people report their attitudes as when they behave toward the attitude object. For example, if people say they prefer a certain brand of soft drink, does this necessarily mean that they are more likely to purchase that brand than others? Only if the attitude constructed at the time they reported their preference is the same attitude constructed when they are standing in the soft drink aisle of the supermarket. As seen in the papers by Shavitt and Tesser, there are a number of circumstances under which the attitude that is salient to people when they report their feelings is different from the attitude that is salient when they behave, resulting in low attitude-behavior consistency.

One such circumstance is the kind of self-reflection we have investigated, where people are asked to explain their attitude before reporting it. We have shown in a number of studies that people who reflect about their attitudes in this way exhibit significantly lower attitude-behavior consistency than people who do not (see Wilson, Dunn, Kraft, and Lisle 1989 for a review). One reason for this finding is as follows: As argued earlier, when people analyze reasons a new attitude often becomes salient, leading to attitude change. This attitude change, however, does not appear to be particularly long lasting. Indeed, it would be rather surprising if a permanent change in people's attitudes could be brought about simply by asking them to explain why they felt the way they did. Instead, over time people's original attitude seems to "snap back."

If so, then the attitude people report after analyzing reasons is likely to be different than their behavior toward the attitude object, as long as the behavior occurs after people's original attitude has snapped back. Consistent with this hypothesis, we have found that if people behave toward the attitude object immediately after analyzing reasons, then their behavior is consistent with their new attitude. If enough time passes, however, people's behavior seems to be consistent with their original attitude and inconsistent with the attitude they reported after analyzing reasons (see Wilson, Dunn, Kraft, and Lisle 1989). These findings are of some importance to survey researchers who hope to measure people's attitudes toward a product, with the hope that these attitudes will predict their consumer behavior. People's attitudes may not predict behavior if they are asked to explain these attitudes before reporting them

IMPLICATIONS FOR PEOPLE'S SATISFACTION WITH CONSUMER DECISIONS

Recently we have examined the implications of our work for how people make consumer decisions. As a starting point, we assume when people are left to their own devices, they often make good personal choices. People are certainly not optimal information processors, but they often manage to assign weights to the different attributes of the alternatives that produce a satisfactory choice (satisfactory to them). People often are not fully aware of how they are weighting the information, we assume, but they often use fairly optimal weights.

If this is true, what happens when people think about why they feel the way they do? Tampering with the decision process by making people introspect, we suggest, can have deleterious consequences by altering how people weight information about the alternatives. When people analyze reasons, we have argued that they focus on those attributes of the attitude object that are available in memory and which seem like plausible causes of their evaluations. As a result, people might lose sight of which criteria of the alternatives are most important to them.

For example, consider people's preferences for a food product. When people taste the food and decide how much they like it, they might be influenced by attributes of which they are unaware (e.g., an unknown ingredient) or that are difficult to verbalize (e.g., its texture or aroma). When asked to explain their reactions, they may assign less weight to these factors because they are difficult to put into words, and as a consequence change their preferences. Assuming that their original preference was fairly optimal, this change in weights might lead to a less optimal choice.

But what do we mean by a "satisfactory" or "optimal" choice? As noted by Janis and Mann (1977), it is very difficult to evaluate how good a decision is. We have dealt with this problem by including several different measures of decision quality, including how well people's opinions of the alternatives match the opinions of experts in that domain, the correspondence between the criteria people use to make their decision and the criteria advocated by experts, and, most directly, how satisfied people say they are with their decision. We will briefly discuss some studies we have conducted to rest the hypothesis that at least under some circumstances, introspecting about reasons can lead to poorer decisions.

Comparing People's Attitudes to Expert Opinion

In one study, we compared the preferences of subjects toward a food item--strawberry jams--with the opinions of sensory panelists from Consumer Reports Magazine (Wilson, Lisle, and Schooler 1989). Subjects tasted five different brands. Half of the subjects analyzed why they liked or disliked each alternative, whereas half did not. All subjects then rated how much they liked each brand. We suggested earlier that left to their own devices, people often make reasonably good consumer decisions. Consistent with this hypothesis, the ratings made by control subjects corresponded fairly well with the experts' ratings, resulting in a mean, within-subject correlation of .55. We have also suggested that trying to explain one's attitudes can influence the salience of certain attributes of the attitude object, causing people to change their evaluations. This prediction was also borne out, in that subjects who analyzed reasons ended up with significantly different preferences for the jams than did control subjects. Finally, consistent with our hypothesis that these preferences would be in some sense "worse," the ratings made by subjects who analyzed reasons did not correspond with the experts very well, mean correlation = .11. The difference between the mean correlations in the control and reasons conditions was significant.

Choosing College Courses

The jam study examined people's preferences, without asking them to make an actual consumer decision. In another study we examined a real-life choice of some importance to college students: The decision of which courses to take (Wilson 1989). We evaluated students' choices in a variety of ways. For example, we compared the criteria they appeared to use to the opinions of faculty members as to what criteria ought to be used when choosing a course, plus we saw who was most likely to take the courses most highly rated by students: Those who introspected about the alternatives or those who did not.

A sample of introductory psychology students, who had expressed an interest in taking more psychology classes, were seen at the beginning of the week when they registered for classes for the next semester. They were given a packet of information about each of the nine sophomore level psychology classes being offered the next term. This packet contained such information as a description of the course content, student evaluations of the course, and whether or not a term paper was required. After reading through this information, some subjects were asked to describe why they might or might not take each course. To see if other kinds of reflection are equally disruptive, a second introspection condition was included: Some subjects were asked to rate each and every piece of information they received about each course according to whether it made them more or less likely to take that course. The remaining subjects were in a control condition, where they did not receive any special instructions about how to approach the course information.

The chief dependent measures were subjects' ratings of how each type of information had influenced their decision of which courses to take, their memory for the different information about the courses at the end of the session, and the courses they actually registered for and took the next semester. As predicted, introspecting about the courses seemed to change how subjects used the information about the courses (subjects who analyzed reasons and subjects who rated each piece of information responded similarly, thus for present purposes the results will be collapsed across these conditions). Subjects in the introspection conditions reported using different kinds of criteria and recalled different kinds of information about the courses than did control subjects. Further, they registered for different types of courses than control subjects.

Who is to say which group made the "right" choices? Though it is impossible to tell, there was suggestive evidence that it was control subjects and not those in the introspection conditions. First, subjects in the control condition were significantly more likely to register for and take courses that received high course evaluations (see Table 1). As another way of assessing how good subjects' choices were, we gave a list of the information our subjects received to a sample of faculty members in psychology, and asked them to rate how much students ought to weight each piece of information, in order to make the best choice. These ratings by the "experts" were then correlated with the actual subjects' ratings of how much they had used each type of information, as well as their recall of the information. Again, it was the control subjects who came out best on these measures: Both correlations were significantly higher in the control condition than in the introspection conditions (see Table 2). What appears to have happened is that subjects who introspected lost sight of which criteria (e.g., the course evaluations) were most important, and relied more on criteria that were, at least according to faculty members, not as valuable.

Post-Choice Satisfaction With Consumer Choices

The results of the first two studies suggest that analyzing reasons changes the criteria people use to evaluate alternatives. These studies are at least suggestive of the possibility that the criteria used by people who introspect are not as "good," leading to worse decisions. But will people themselves be dissatisfied with these choices? On the one hand they might not, because several studies have found that whatever alternative people choose, they will become more favorably disposed toward it in order to reduce dissonance (e.g. Brehm 1966). On the other hand there are probably limits to this kind of dissonance reduction (to illustrate this, one need only refer to the high rate of divorce in this country! ).

There are two reasons why people who introspect might regret their choices more than those who do not. First, if people who introspect end up choosing inferior alternatives, such as poorly-rated jams, then they should not be as satisfied with these choices as people who do not introspect. Second, we argued earlier that the attitude change produced by analyzing reasons is temporary, and snaps back over time. If so, people should come to regret a choice they made immediately after analyzing reasons, because their initial, pre-introspective attitude has returned.

Two studies have been performed to test these hypotheses (Wilson, Lisle, and Schooler 1989). In the first study, subjects were presented with five art posters and allowed to choose one to take home. Two of the posters were reproductions of Impressionist paintings, and were very popular with our student population, according to our pretesting. The remaining three were contemporary posters of a style we- can only call "cute," such as a photograph of a cat perched on a rope with the caption, "Gimme a Break." They were considerably less popular with our pretest subjects. Half of the subjects were asked to write down why they felt the way they did about each poster, supposedly to organize their thoughts. The other half completed a filler questionnaire of equal length. All subjects then rated the posters and chose one to take home. The experimenter left the room while the subject chose a poster from bins containing several of each type; thus, in the minds of the subjects, the experimenter would not know which one they took.

As in our other studies, analyzing reasons seemed to change the criteria subjects used to evaluate the posters. Subjects who did not analyze reasons overwhelmingly preferred the Impressionist paintings: As seen in Table 3, virtually everyone chose one of these posters to take home. Subjects who analyzed reasons, however, were significantly more likely to choose one of the contemporary posters. Having to explain their attitudes seems to have increased the salience of positive features of the contemporary posters, in that subjects listed more positive than negative attributes of these posters. What happened after people took their posters home? We telephoned them 3 weeks later to find out. As predicted, and as seen in Table 3, subjects in the reasons condition reported a lower satisfaction with their choice of poster, possibly because their initial attitude--that the Impressionist paintings were preferable--had by then returned.

TABLE 1

COURSES PREREGISTERED FOR AND ACTUALLY TAKEN BY CONDITION

TABLE 2

CORRELATIONS BETWEEN FACULTY RATINGS OF HOW MUCH THE COURSE INFORMATION SHOULD BE WEIGHTED AND (A) SUBJECTS REPORTED INFLUENCE OF THE INFORMATION AND (B) SUBJECTS' RECALL OF THE INFORMATION

According to our prediction, it should only be people who convinced themselves they liked the unpopular, contemporary posters who should come to regret this choice later. As seen in Table 3 there was a tendency in this direction, in that the difference in post-choice satisfaction between reasons and control subjects was greater among those who chose a contemporary poster than those who chose an Impressionist poster. The interaction between condition and poster choice, however, was not significant. One reason for this might be that the control condition/contemporary poster cell was represented by one person, making this value highly unreliable. We replicated the poster study with a new stimuli, felt-tip pens, and found results that were more consistent with our hypothesis. The people who were least satisfied with their choice of a pen were those in the reasons condition who convinced themselves that they liked the one that was rather unpopular, and chose to take it home. When asked later how much they liked it, their original attitude seemed to have returned, making them relatively unhappy with their choice.

CONCLUSIONS

Viewing attitudes as the result of a constructive process, dependent on the thoughts that are salient to people at any given moment, has some interesting implications for consumer behavior. We have found that asking people to explain their attitudes changes their preferences, presumably by influencing the salience of different attributes of the attitude object. Further, we have shown that this type of self-reflection can lower the quality of people's choices. We should, however, add some caveats to our findings. First, we are not suggesting that all kinds of reflection should be avoided, or even that people should never analyze the reasons for their feelings. In most of our studies people are asked to make off-the-cuff, fairly brief explanations of their attitudes, and it is possible that a more in depth kind of reflection, such as that advocated by Janis and Mann (1977), will have very different effects on the quality of people's decisions. Further, there are times when people make decisions too impulsively, and would benefit from a more measured, reflective approach to the decision. We have just begun to explore the conditions under which people should and should not introspect about the reasons for their preferences. Perhaps the best conclusion at this point is a variation of Socrates oft-quoted statement that the "unexamined life is not worth living." We suggest that at least at times, the unexamined choice is worth making.

TABLE 3

RESULTS OF THE WILSON, LISLE, AND SCHOOLER (1988) POSTER STUDY

REFERENCES

Brehm, Jack W. (1956), "Post-decision changes in the desirability of alternatives," Journal of Abnormal and Social Psychology,52 (May), 384-389.

Chaiken, Shelly and Yates, Suzanne (1985), "Affective-Cognitive Consistency and Thought-Induced Attitude Polarization," Journal of Personality and Social Psychology, 49 (December), 1470- 1481.

Janis, Irving L. and Mann, Leon (1977), Decision Making: A Psychological Analysis of Conflict, Choice, and Commitment, New York: The Free Press.

Lusk, Cynthia M. and Judd, Charles M. (1988). "Political Expertise and the Structural Mediators of Candidate Evaluations," Journal of Experimental Social Psychology, 24 (March), 105-126.

Millar, Murray G. and Tesser, Abraham (1986), "Thought-Induced Attitude Change: The Effects of Schema Structure and Commitment," Journal of Personality and Social Psychology, 51 (August), 259 -269.

Petty, Richard E. and Cacioppo, John T. (1981), Attitudes and Persuasion: Classic and Contemporary Approaches, Dubuque, IA: Wm. C. Brown Co.

Salancik, J. R. (1974), "Inference of One's Attitude from Behavior Recalled Under Linguistically Manipulated Cognitive Sets," Journal of Experimental Social Psychology, 10 (September), 415-427.

Seligman, Clive, Fazio, Russell H., Zanna, Mark P. (1980), "Effects of Salience of Extrinsic Rewards on Liking and Loving," Journal of Personality and Social Psychology,38 (M arch), 453460.

Shavitt, Sharon (1989), "Operationalizing Functional Theories of Attitude." in Attitude Structure and Function, ed. Anthony R. Pratkanis, Steven J. Breckler, and Anthony G. Greenwald, Hillsdale, NJ: Lawrence Erlbaum Associates, 311 -337.

Sherif, Muzafer and Hovland, Carl I. (1961). Social Judgment: Assimilation and Contrast Effects in Communication and Attitude Change New Haven: Yale University Press.

Tesser, Abraham (1978). "Self-Generated Attitude Change." in Advances in Experimental Social Psychology, ed. Leonard Berkowitz, New York: Academic Press, Vol. 11, 289-338.

Wilson, Timothy D. (1989). "The Effects of Self Reflection on Choosing College Courses," Unpublished data, Department of Psychology, University of Virginia, VA 22903.

Wilson, Timothy D. (in press), "Self-Persuasion via Self-Reflection," in Self-Inference Processes: The Ontario Symposium, ed. James M. Olson and Mark P. Zanna, Hillsdale, NJ, Vol 6.

Wilson, Timothy D., Dunn, Dana S., Kraft, Dolores, and Lisle, Douglas J. (1989), "Introspection, Attitude Change, and Attitude-Behavior Consistency: The Disruptive Effects of Explaining Why We Feel the Way We Do." in Advances in Experimental Social Psychology, ed. Leonard Berkowitz, Orlando, FL: Academic Press, Vol. 19, 123-205.

Wilson, Timothy D., Kraft, Dolores, and Dunn, Dana S. (1989), 'The Disruptive Effects of Explaining Attitudes- The Moderating Effect of Knowledge About the Attitude Object," Journal of Experimental Social Psychology, 25 (September), 379-400.

Wilson, Timothy D., Lisle, Douglas J., and Schooler, Jonathan (1989), "Some Undesirable Effects of Self-Reflection," unpublished manuscript, Department of Psychology, University of Virginia, VA 22903.

Wilson, Timothy D. and Pollack, Scott (1989). "Effects of Explaining Attitudes on Survey Responses," unpublished manuscript, Department of Psychology, University of Virginia, VA 22903.

Zanna, Mark P. and Rempel, John K. (1988), "Attitudes: A New Look at an Old Concept," in The Social Psychology of Knowledge, ed. Daniel Bar-Tal and Arie Kruglanski, New York: Cambridge University Press, 315-334.

----------------------------------------