Self-Perception As a Means of Personal Influence: the Foot-In-The-Door Technique

Brian Sternthal, Northwestern University
Carol A. Scott, Ohio State University
Ruby Roy Dholakia, Northwestern University
ABSTRACT - Two experiments were conducted regarding the foot-in-the-door phenomenon as a personal influence strategy. The first experiment examined the efficacy of the foot-in-the-door under different levels of incentive. The second tested the adequacy of self-perception theory as an explanation for the foot-in-the-door effect.
[ to cite ]:
Brian Sternthal, Carol A. Scott, and Ruby Roy Dholakia (1976) ,"Self-Perception As a Means of Personal Influence: the Foot-In-The-Door Technique", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 387-393.

Advances in Consumer Research Volume 3, 1976      Pages 387-393

SELF-PERCEPTION AS A MEANS OF PERSONAL INFLUENCE: THE FOOT-IN-THE-DOOR TECHNIQUE

Brian Sternthal, Northwestern University

Carol A. Scott, Ohio State University

Ruby Roy Dholakia, Northwestern University

ABSTRACT -

Two experiments were conducted regarding the foot-in-the-door phenomenon as a personal influence strategy. The first experiment examined the efficacy of the foot-in-the-door under different levels of incentive. The second tested the adequacy of self-perception theory as an explanation for the foot-in-the-door effect.

The preponderance of the research pertaining to interpersonal influence has focused on the efficacy of various mass communication strategies. In part, the relative neglect of individual influence contexts reflects the lack of appropriate experimental paradigms to study the phenomenon. In part too, the neglect is attributable to the belief that strategies demonstrated to be effective in mass communication settings will also be appropriate in personal contact situations. However, in light of the unique attributes of personal contact, such as the availability of immediate feedback from message recipients, it may well be that certain influence strategies which fail in mass communication contexts will be successful in individual influence situations.

This paper examines the effectiveness of one individual influence strategy, the "foot-in-the-door" technique. The foot-in-the-door, as it is commonly employed in personal selling, entails asking individuals to make a relatively small purchase (or trial) in the hope that compliance with the request will increase the likelihood of compliance with subsequent larger purchase solicitations. In contrast to mass persuasion strategies, which are predicated on the assumption that behavior can be modified by influencing its cognitive precursors, the foot-in-the-door technique involves direct modification of behavior. More specifically, it is based on the self-perception proposition which states t-hat individuals come to know their attitudes by inferring them from their own behavior and the circumstances in which that behavior occurs. In turn, the attitudinal dispositions acquired in this manner guide subsequent action (Bem, 1972).

Several field experimental studies have demonstrated the success of the foot-in-the-door technique. In their seminal investigation, Freedman and Fraser (1966) observed that individuals who agreed to comply with a small request evinced a greater willingness to comply with a subsequent larger request than people asked to comply with only the larger request (i.e., cold called). Moreover, this foot-in-the-door effect emerged whether or not the small request involved the same issue and task as the large request.

Subsequent investigations have generally confirmed the foot-in-the-door finding, and have specified the conditions under which it was likely to be observed. Pliner, Hart, Kohl, and Saabi (1974) reported that compliance with a large request, donating money to charity, was greater when it was preceded by a small or moderate request than by no request. However, the dollar amount given by those contributing did not vary according to whether or not experimental participants had complied with a previous small request. Baron (1973) replicated the foot-in-the-door finding, but only when males served as experimenters and the initial request was very small. Cialdini (in review), however, failed to replicate the foot-in-the-door phenomenon, probably because small and large requests were made in close temporal proximity.

These foot-in-the-door findings are generally congenial to self-perception theory (Bem, 1972). Indeed, Freedman and Fraser's interpretation of their own data was in attribution terms:

...Once he has agreed to a request, his attitude may change. He may become in his own eyes, the kind of person who does this sort of thing, who agrees to requests made by strangers, and who takes action on things he believes in, who cooperates with good causes (Freedman and Fraser, 1966, p. 201).

In effect, it is hypothesized that the foot-in-the-door is effective because people use their own behavior as a cue regarding their attitudinal dispositions. Since external pressure for the initial behavior is assumed to be minimal, people infer a positive attitude from their compliant behavior, which in turn guides subsequent action.

In a test of this explanation, Snyder and Cunningham (1975) found that, in accord with the self-perception prediction, people who complied with an initial small request (answering 8 questions in a telephone survey) were more willing to comply with a subsequent large request (answering 30 questions) than subjects who were initially asked to comply with a very large request (answering 50 questions), or subjects in the no initial request condition. Similarly, Baron (1973) found that non-compliance with an initial large request led to reduced compliance with a subsequent smaller request, whereas compliance with an initial small request led to greater compliance with the subsequent request. No differences were observed between the initial large request group and no first request group.

Thus, individuals' actions are causally related to their prior behavior. Further, it has been suggested that individuals incorporate the circumstances in which behavior occurs into the attribution rules they employ in performing causal analysis of their own behavior. Specifically, for behavior to result in belief inferences it must be self-attributed, or perceived to be elicited by the person's disposition toward the stimulus. Circumstantial forces can attenuate the probability of self-attribution of behavior and the resulting belief inferences and behavior. If the behavior is perceived to be elicited by plausible external causal factors present in the situation (e.g., coupon, incentive, high credibility source), the individual should discount internal motivations as the cause of his behavior and no belief inferences should be made (discounting principle, Kelley, 1971). In a field experiment, Uranowitz (1975) tested the self-perception explanation under two different conditions of external justification. Consistent with the self-perception prediction, women asked to watch the experimenter's packages when justification was low (experimenter had to retrieve a dollar) exhibited greater willingness to help a second experimenter than when justification for watching his packages was high (retrieve his wallet) or when there was no initial contact.

In such, the literature demonstrates that the foot-in-the-door technique is an effective means of influence relative to cold calling. Furthermore, these data are congenial to self-perception theory; the performance of a behavior increases the likelihood of more substantial behaviors of the same genre, providing that individuals do not perceive circumstantial factors to have caused initial compliance. However, none of the studies reviewed have examined the foot-in-the-door technique in a commercial setting where incentives are likely to be necessary to induce consumers to engage in even small behaviors. Nor has there been a direct test of the self-perception hypothesis that past and future behaviors are mediated by attitudinal dispositions.

The present research involves two experiments. In Experiment I, the effectiveness of the foot-in-the-door technique is examined in a typical marketing setting, and the extent to which self-perception theory can explain these findings is assessed. Experiment II provides a more rigorous test of the self-perception explanation.

EXPERIMENT I: FOOT-IN-THE-DOOR IN A COMMERCIAL CONTEXT

Overview

In Experiment I, participants were given one of four incentives to enhance compliance with a small request. Two weeks later all participants were recontacted and asked to comply with a large request. A control group was asked to comply only with the large request. In accord with self-perception theory, it was hypothesized that individuals contacted for an initial small request (trial) would be more likely to comply with a subsequent large request than those contacted only for the large request.

It was also expected that participants asked to comply with an initial small request without benefit of an incentive would be more likely to comply with a second large request than those participants offered an incentive for initial trial. Although not performed in the context of the foot-in-the-door paradigm, studies pertaining to incentive strategies support this hypothesis. Doob and his coworkers (1969) introduced new brands of inexpensive, frequently-purchased consumer products in experimental stores at a discounted price and in control stores at the regular (competitive) price. After a nine-day period, discounts were removed and the effects on product sales observed. As expected, experimental stores outsold control stores during the discount period. When the discount was eliminated, however, this advantage was lost. In fact, by the end of a four-week observation period the control stores surpassed experimental stores in sales. Similarly, Aaker (1969, 1972), used regression analysis of panel data to investigate the effects of an initial purchase under deal conditions on subsequent brand acceptance. His findings suggest that purchase on deal may be negatively associated with acceptance, although the results across product categories were less than univocal.

Evidence for the hypothesis that behavior predicated on incentives undermines the performance of subsequent behavior also obtains from studies employing an over-justification paradigm (e.g., Calder and Staw, 1975; Deci, 1971; Kurglanski, et al., 1975; Lepper and Greene, 1975). Overjustification is operationalized as behavior which is both intrinsically attractive and externally rewarded. The consistent finding of these studies is that the provision of a reward or incentive for performing an enjoyable behavior results in a decrease in intrinsic interest in subsequently engaging in that behavior.

Method

A weekly community newspaper serving the northwest suburbs of Chicago was selected for study. The paper had just been introduced to the market and was suitable for the experiment, since it could be tried on a limited basis and since the cost of trial was sufficiently low to insure a reasonable rate of compliance with an initial request.

Four hundred and twenty subjects were selected from a list of potential subscribers provided by the publisher and were randomly assigned to one of five groups: four experimental groups and one control group. Subjects in the experimental treatment conditions were called by the newspaper's regular telephone solicitation staff and asked to accept a two-week trial subscription to the paper at either the regular price (504), at half price (254), for free, or for free plus receipt of a premium gift (a coupon worth 504 at a well-known fast food restaurant). Two weeks later, all participants (all experimental participants, regardless of whether they accepted the trial offer, and control group participants) were called and asked to subscribe to the newspaper on a regular basis. Subjects were required to select at least a six-months subscription, which cost $4.

The independent variable was the amount of incentive offered for trial subscription acceptance (none, 254, 504, 504 plus a premium gift). The percent of subjects in each treatment who agreed to subscribe on a regular basis served as the dependent variable. The control group, contacted only for the regular subscription request, was included to provide a measure of the response that would be obtained with no trial offer of any kind.

Results

According to self-perception theory, experimental participants who were given a two-week trial offer should evince greater rates of subscription than those merely asked to subscribe. This hypothesis was not supported by the data. When all trial offer conditions are combined, and compared with the no trial group, the differences in subscription behavior were nonsignificant (X2 = 1.26, d.f.=l, p > .15) [All significance levels reported for 2x2 contingency tables are based on one-tailed tests, since the comparisons were performed to test directional hypotheses.] Thus, trial per se does not enhance the likelihood of repeat purchase behavior.

When responses are categorized by the amount of incentive offered for trial, the effects of trial are somewhat different. As Table 1 indicates, the fifty percent discount treatment produced a significantly greater rate of subscription than the no-trial control treatment (X2 = 3.0, d.f. = 1, p > 05). [Despite the fact that paired comparisons were planned rather than conducted in a post hoc fashion, it might be contended that inflated error rates due to the nonorthogonality of the contrasts caused intertreatment differences to be significant. To guard against this possibility an analog of Scheffe's procedure for multiple comparisons was used (Maraciullo, 1966). However, since this procedure did not change the interpretation of the findings it is not reported here.] The lowest rate of subscription occurred for the free trial group. No significant differences were found between the regular price and fifty percent discount treatments (X2 =.73, d.f. = 1, p >.20). Contrary to expectations, the free plus premium gift condition produced almost as great a subscription rate as the fifty percent discount treatment, although it was not significantly greater than the no-trial control treatment (X2 = 1.06, d.f. = 1, p > .15).

TABLE 1

SUBSCRIPTION BEHAVIOR OF EACH TRIAL OFFER AND CONTROL GROUPS (TOTAL SAMPLE)

The above analyses were conducted using all participants, irrespective of their trial acceptance behavior. This was necessary to avoid the self-selection bias that would have been present had only those who complied with the trial request been included in the analysis. From a theoretical perspective, however, there is good reason to examine the responses of only the trail com-pliers. Specifically, Lepper, et al. (1975) state that overjustification effects may be evident only when there is some initial interest in the activity. Where no initial interest exists, the smallest amount of incentive necessary to elicit behavior should be used. In addition, since discounts are usually offered to encourage product trial by those persons who would not otherwise do so, it would be useful to know how successful this strategy is in influencing repeat purchases.

The subscription behavior of the trial compliers in each incentive condition is shown in Table 2. Here, the superiority of the fifty percent discount treatment is clearly demonstrated; the rate of compliance obtained in this condition is significantly greater than that obtained from any larger incentive treatment, though not significantly greater than the regular price condition rate (X2 = .56, d.f. = 1, p > .25). Again, the lowest subscription rate occurred under the free trial condition. Further, only the regular and fifty percent discount groups produced significantly greater subscription rates than the no-trial control group (X2 = 9.14 and 18.96, respectively, d.f.= 1, p<.01).

TABLE 2

% TRIAL ACCEPTORS COMPLYING WITH SUBSCRIPTION REQUEST

Discussion

The results of this study have several practical implications. First, the finding that the fifty percent discount was the only trial offer that was significantly more effective than the cold calling approach suggests that some small incentive is necessary to gain the greatest benefits from a foot-in-the-door strategy. Second, although the fifty percent trial and the premium trial (i.e., free trial plus 504 gift certificate) were about equally effective in generating subscriptions, the latter offer was four times as expensive ($1 versus 254) and thus is less preferred. Third, the data suggest that the most efficient procedure in implementing the foot-in-the-door technique entails confining call-backs to those people who accepted the trial offer; merely recontacting people had no effect on subscription behavior. Indeed, those persons who rejected the trial offer were no more likely to subscribe to the newspaper than the control group which was not made a trail offer (Table 3). It would probably be more efficient to use a promotional mix that combined advertising and a small incentive to arouse product interest and trial acceptance.

TABLE 3

SUBSCRIPTION BEHAVIOR OF TRIAL OFFER NON-ACCEPTORS

The results of this study provide only equivocal support for self-perception theory. As expected, the free trial offer yielded less compliance with the subscription request than did smaller incentives. Furthermore, the failure to observe significant differences in compliance behavior between the regular and half-price treatment groups, though not predicted by self-perception theory, does not constitute damaging evidence. Unlike the potent discounting cues used in previous research (e.g., drugs or a sizable reward), the incentive used in the present research may have been of a sufficiently small magnitude so that it served as a positive reinforcer rather than a discounting one.

More problematic from a self-perception standpoint is the finding that the premium offer was as successful in inducing subscriptions as the fifty percent discount. Several factors may account for this finding. The small difference between the two groups when all subjects were included is due primarily to the large number of non-compilers with the fifty percent discount trial request relative to the premium trial offer (70.7% versus 23.8%). When only those who complied with the trial request are examined, the fifty percent discount resulted in more subscriptions than did the premium offer. These data imply that despite similar rates of compliance with the subscription request in the fifty percent discount and premium treatments, the reason for the observed compliance rate differed between treatments. Apparently, for many of the participants in the fifty percent discount treatment, the incentive was not sufficient to induce trial. Moreover, these non-compilers are likely to have attributed to their behavior to a negative disposition toward the newspaper and refused to subscribe when the opportunity was presented. For those accepting the trial offer despite the marginal inducement, compliance was likely to be attributed to their liking of the newspaper, thus enhancing their probability of subscription. On the other hand, the majority of individuals assigned to the premium condition complied with the trial request but were probably uncertain whether their compliance was attributable to a favorable disposition toward the newspaper or to the attractive incentive. In turn, this uncertainty regarding the appropriate cause of their behavior reduced the subscription rate relative to acceptors of the fifty percent discount. If this interpretation is correct, then the failure to observe differences between the premium and fifty percent discount condition is interpretable in self-perception terms.

Although alternative explanations for the similarity in subscription induced by a fifty percent discount and premium offer are available, they appear less tenable than the interpretation offered above. For example, it may be argued that the premium, which was a fifty percent gift certificate at a popular fast food franchise, induced a sufficiently high degree of positive affect toward the newspaper that it obviated any discounting that may have taken place. While this response may have occurred, for some participants it does not explain the aggregate finding. If it did, then one would predict higher rates of compliance among persons accepting the premium offer than among those accepting the fifty percent discount. In fact, just the opposite was found (Table 2).

EXPERIMENT II: TEST OF THE SELF-PERCEPTION MECHANISM

Overview

In Experiment I, qualified support for the self-perception explanation of the foot-in-the-door phenomenon was obtained by manipulating the hypothesized antecedents of self-attribution and observing the effects on behavior. Experiment II sought to clarify further the efficacy of self-perception theory by examining the effects of attributions on attitudes predicted by the theory. More specifically, subjects were asked to read voluntarily a persuasive appeal attributed either to a high or low credibility source and to sign a petition supporting the communication issue. A series of scaled questions were administered to determine subjects' attitudes toward the issue either before or after they had made a decision regarding whether or not they would sign the petition. In contrast to Experiment I, where incentives served as circumstantial cues, in Experiment II source credibility constituted the factor that mediated subjects' causal attributions. Moreover, Experiment II provided a test of the self-perception explanation of the foot-in-the-door in a situation where a communicator presents information about the attitudinal object in order to gain compliance--a situation that is common in purchase and consumption contexts.

According to self-perception theory, the interaction between source credibility and position of the behavioral request should be significant. Given compliance with the request made by a low credibility source, subjects' attitudes toward the issue should be more positive when determined after they have complied with a request than before compliance has occurred. Observing that they have complied despite the presence of a low credibility source should augment the attribution of behavior to internal causes. If their behavior is not available to subjects as a cue regarding their attitudes, then the situation entails a persuasion paradigm where it has been repeatedly observed that low credibility sources induce relatively little attitude change. On the other hand, if the appeal is attributed to a highly credible source, subjects will be uncertain about the underlying causes of their compliance. It may be attributable to either some personal feeling about the issue (i.e., an internal cause) or to the fact that the message was presented by a highly credible source. As a result, subjects should exhibit a less positive attitude toward the issue in this situation than when the attitudinal posttest precedes behavior.

A similar line of reasoning can be used to predict the attitudes of subjects who did not comply with the behavioral request. Non-compliant subjects exposed to a low credibility communicator are likely to be uncertain whether their failure to comply was due to the low credibility of the source or the fact that they have a negative attitude towards the issue. Conversely, if source credibility is high, non-compliance is more likely to be attributed to subjects' internal dispositions; the requested behavior was not performed despite the fact that source credibility was high.

Method

Subjects. The 110 participants in the study were recruited from three sections of an introductory course in management. During one of the regular class hours, the students were approached to participate in a study that was being conducted on federal legislation. Participation was made voluntary and the sponsors of the study were described as an external group who had sought the assistance of the course instructor to administer the study. All the students agreed to participate. [An informed consent procedure (Berscheid, Baron, Dernier, and Libman, 1974) administered during the debriefing revealed that less than five percent of the subjects had any objections to the experimental talk.]

Procedure. Experimental subjects were given booklets containing the independent and dependent variables by their class instructor. They were asked to read a one-page message which stated the goals of the Consumer Protection Agency Bill (S.707, then pending before the Senate) and which presented arguments for supporting its passage. The credibility of the source was manipulated at the beginning of the communication. For half the subjects randomly assigned to the high credibility condition, the source was described as a Harvard-trained lawyer who was a recognized expert regarding consumer issues. His support for this Bill was made distinctive by describing him as a person who generally did not favor government controls. In contrast, for those subjects assigned to the low credibility condition, the source was portrayed as an individual with no special expertise but who was interested in consumer protection because it represented a job opportunity as a consumer lobbyist. Further, he was described as a person who held socialistic views and favored all government controls. A manipulation check indicated this induction was successful.

The second independent variable involved manipulating the position of the dependent measures. The dependent variables included subjects' attitude toward the Bill, measured by six items on a seven-point Likert-type scale, their evaluation of the source, assessed on the basis of responses to nine semantic differential items, and their behavioral compliance, determined by whether or not they signed a petition to be sent to their Senators urging their representatives to vote for the Bill. For half the subjects, the verbal scales were administered after they had the opportunity to comply with or reject the request to support the Bill. The remaining subjects completed the verbal scales first and then responded to the behavioral request.

Results

The two experimental variables did not affect the extent of compliance with the request to sign the petition favoring the Consumer Protection Agency Bill. Of the subjects randomly assigned to the high credibility source condition, 58.18 percent complied, while 43.64 percent of thee assigned to the low credibility condition signed the petition (Z=.19, p>.40). Moreover, whether the request to sign the petition preceded or followed the administration of the attitude measure did not significantly affect compliance. Forty-one percent of the subjects complied when the petition request followed the attitude measures and 61.5 percent complied when it was administered immediately after the communication (Z=.19, p > .40).

To determine the effects of the experimental variables on subjects' attitudes toward the Consumer Protection Agency Bill, an analysis of variance was performed (Table 4). Consistent with the findings relative to behavioral compliance with the request to sign the petition, source credibility and position of the behavioral request did not have a significant effect on attitudes. The compliance main effect, however, was significant (F=52.798, d.f.=1/102, p<.01); those subjects who had self-selected themselves to sign the petition evinced more favorable attitudes toward the Consumer Protection Agency Bill than those who refused to comply with the petition request.

The source X position of behavioral request interaction was also significant (F=5.915, d.f.=1/102, p<.05). This effect is attributable to the fact that the highly credible source induced a more positive attitude toward the Bill when the attitude measure was administered before the behavioral request, whereas the low credibility source was more persuasive if behavioral request preceded the administration of the attitudinal post test (Figure 1). Furthermore, separate analysis of variance for compilers and non-compilers yielded a significant source X position interaction for those who complied with the request to sign the petition (Table 5, F=4.456, d.f.=1/52, p<.05). For non-compilers, the joint effect of source credibility and position was similar to that for compilers, but it did not reach conventional levels of statistical significance (F=2.628, d.f.=1/50, p=.10).

TABLE 4

ANALYSIS OF VARIANCE ON MEAN ATTITUDES TOWARD CONSUMER PROTECTION BILL

FIGURE 1

MEAN ATTITUDE AS A FUNCTION OF POSITION OF THE BEHAVIORAL REQUEST

TABLE 5

ANALYSIS OF VARIANCE OF ATTITUDE FOR COMPILERS AND NON-COMPILERS

Discussion

Demonstrations of the foot-in-the-door phenomenon have usually been interpreted in terms of self-perception theory (Bem, 1972). According to this explanation the foot-in-the-door is effective because: "compliance with a small request causes the subject to infer that he has a positive attitude toward cooperating with good causes; in turn, this positive attitude leads to compliance with the larger request" (Pliner, et al., 1974, p. 18). Experiment II provides further support for the self-perception theory explanation of the foot-in-the-door phenomenon. When behavior served as a cue and an external justification or pressure for that behavior was not present, as is the case in most demonstrations of the foot-in-the-door effect, subjects inferred stronger internal dispositions than when behavior could readily be attributed to an external cause. More specifically, for those people who complied with the high credibility source's request to sign the petition favoring the Consumer Protection Agency Bill, the source's credibility served as a discounting cue. Since compliance could be attributed to both internal dispositions and an external cause (i.e., source credibility), individuals in this treatment exhibited relatively unfavorable attitudes toward the Bill. Conversely, when the communicator was of low credibility, the source cue augmented the attribution of compliant behavior to internal or personal causes. Hence, subjects evinced a relatively favorable attitude toward the Bill.

Support for the self-perception theory would have been more compelling if the source credibility X position interaction for those who did not comply with the behavioral request had reached conventional levels of statistical significance. Nevertheless, treatment differences in attitudinal response were in the predicted direction for individuals exposed to a high credibility communicator. Their attitudes were less favorable when determined after non-compliance than before (t=1.32, d.f.=21, p=.10). A smaller change was observed among individuals receiving the message from a less credible source (t=.75, d.f.-29, p >.40).

The failure to observe a significant source x position interaction among non-compilers could be attributable to the nature of the experimental task. Individuals who complied to the behavioral request may have had fewer alternative explanations for their behavior; either their own attitudes were favorable (when source credibility was low), or the perceived external justification was high (when source credibility was high). Non-compilers, on the other hand, may have had a wider range of reasons for their behavior than did those who signed the petition. For example, they may have had a favorable attitude toward the Bill, but not toward signing the petition; or they may not have wanted to get involved; or they may have had unfavorable attitudes toward the Bill. This may have caused the greater variability in attitudinal responses observed among non-compilers than among compilers (t=3.227, d.f.=53/55, p <.01).

Finally, contrary to the self-perception prediction, the differences in attitude induced by the low credibility source before and after the behavior request were not significant either for compilers (t=.46, d.f.=22, p>.40). or for non-compilers (t=.75, d.f.=29, p >.40). This may be attributable to the fact that the low credibility source was not perceived in a very negative light. However, since the present investigation entailed gaining compliance with a behavioral request, the source could be disparaged only to a very limited extent. If a spokesman is seeking compliance, it is unlikely that he would stress biographical data that is particularly damaging to his cause. As a result of this inability to severely disparage the source, the low credibility source may have provided subjects with some external justification for behavior, accounting for the finding that attitudes toward the Bill were not significantly more favorable when behavior preceded rather than followed attitude measurement in the low credibility condition.

CONCLUSIONS

From a practical perspective, the experiments reported in this paper suggest that the foot-in-the-door strategy is a viable personal influence strategy providing that the "appropriate" level of incentive is employed to motivate trial. More specifically, some incentive, whether it is operationalized in terms of discounts, source credibility, or some other facilitator, is necessary to induce trial in commercial contexts. However, very substantial incentives may cause individuals to attribute trial to the incentive rather than a positive disposition toward the attitude object. If the incentives subsequently are retracted, the likelihood of compliant behavior diminishes, since the reason for the initial trial is no longer present. Of course, this finding is likely to be maintained only if the product in question is not clearly superior to competitive offerings.

From a theoretical perspective, investigations reported in the paper support the self-perception proposition that behavior, and the circumstances in which it occurs, are important determinants of the attitudes people acquire and that these attitudes, in turn, guide subsequent action. On this basis, it may be hypothesized that people will exhibit attitude-behavior consistency when they have had substantial experience with the attitude object that can be attributed to personal dispositions. Operationally, individuals' confidence in their attitude can serve as a measure of the extent to which behavior has been attributed to a positive disposition toward the object. Indeed, Sample and Warland (1973) observed that factors other than subjects' attitudes did not account for the variance in their behavior when they were highly confident in these judgments.

Finally, the present research provides a basis for constructing a taxonomy of consumption situations. It suggests that situational factors, such as incentives and communicator credibility, can be classified as facilitators of current behavior but inhibitors of subsequent performance, since they undermine the attribution of behavior to internal causes. Other situational variables, such as high price and low credibility, may have the opposite effect. This classification, and self-perception theory on which it is based, question the utility of Belk's (1974) notion of situation which he defines as the "factors particular to a time and place of observation which do not follow from a knowledge of personal ... and stimulus ... attributes, and which have a demonstrable and systematic effect on current behavior" (1974, p. 157). In contrast to Belk's focus on current behavior, the present research indicates the importance of examining the impact of situations on current as well as future behavioral responses.

REFERENCES

D. A. Aaker, "The Long-Term Value of Temporary Price Reductions," unpublished doctoral dissertation, Stanford University, 1969.

D. A. Asker, "A Measure of Brand Acceptance," Journal of Marketing Research, 9(1972), 160-167.

E. Aronson, J. A. Turner and J. M. Carlsmith, "Communicator Credibility and Communicator Discrepancy as Determinants of Opinion Change," Journal of Abnormal and Social Psychology, 67 (1963), 31-37.

R. A. Baron, "The 'Foot-in-the-door' Phenomenon: Mediating Effects of Size of First Request and Sex of Requester," Bulletin of Psychonomic Society, 2 (1973), 113-114.

R. Belk, "An Exploratory Assessment of Situational Effects in Buyer Behavior," Journal of Marketing Research, 11 (1974), 156-163.

D. J. Bem, "Self-perception Theory," in L. Berkowitz (ed.), Advances in Experimental Social Psychology (New York: Academic Press, 1972).

E. Berscheid, R. S. Baron, M. Dernier and M. Libman, "Anticipating Informed Consent," American Psychologist, 7 (1973), 913-925.

B. J. Calder and B. M. Staw, "Self-perception of Intrinsic and Extrinsic Motivation," Journal of Personality and Social Psychology, 31 (1975), 599-605.

R. B. Cialdini, "A Test of Two Techniques for Inducing Verbal, Behavioral, and Further Compliance with a Request to Give Blood," in review.

E. L. Deci, "The Effects of Externally Mediated Rewards on Intrinsic Motivation," Journal of Personality and Social Psychology, 18 (1971), 105-115.

A. N. Doob, J. M. Carlsmith, J. L. Freedman, T. K. Landauer and S. Tom, "Effect of Initial Selling Price on Subsequent Sales," Journal of Personality and Social Psychology, 11 (1969), 345-350.

J. L. Freedman and S.C. Fraser, "Compliance without Pressure: The Foot-in-the-door Technique," Journal of Personality and Social Psychology, 4 (1966), 195-202.

H. H. Kelley, "Attribution in Social Interaction," in E. E. Jones, et al., Attribution: Perceiving the Causes of Behavior (Morristown, N.J.: General Learning Press, 1971), 1-26.

A. W. Kruglanski, A. Riter, A. Amitai, B. Margolin, L. Shabtai and D. Zaksh, "Can Money Enhance Intrinsic Motivation: A Test of the Content-Consequence Hypothesis," Journal of Personality and Social Psychology, 31 (1975), 744-750.

M. R. Lepper, and D. Greene, "Turning Play into Work: Effects of Adult Surveillance and Extrinsic Rewards on Children's Intrinsic Motivation," Journal of Personality and Social Psychology, 31 (1975), 479-486.

L. Marascuilo, "Large Sample Multiple Comparisons," Psychological Bulletin, 65 (1966), 280-290.

P. Pliner, H. Hart, J. Kohl, and D. Saabi, "Compliance without Pressure: Some Further Data on the Foot-in-the-door Technique," Journal of Experimental Social Psychology, 10 (1974), 17-22.

J. Sample and R. Warland, "Attitude and the Prediction of Behavior," Social Forces, 51 (1973), 292-304.

M. Snyder and M. R. Cunningham, "To Comply or Not Comply: Testing the Self-Perception Explanation of the 'Foot-in-the-door' Phenomenon," Journal of Personality and Social Psychology, 31 (1975), 64-67.

S.W. Uranowitz, "Helping and Self-Attributions: A Field Experiment," Journal of Personality and Social Psychology, 31 (1975), 852-854.

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