Incorporating Perceptions of Financial Control in Purchase Prediction: an Empirical Examination of the Theory of Planned Behavior

Arti Sahni, University of Cincinnati
ABSTRACT - While consumer researchers have recognized that financial control over purchase obviously has an impact on purchase, more attention has been devoted to attitude and subjective norms as antecedents of purchase to the neglect of financial capability. The Theory of Reasoned Action has been modified by Ajzen to predict goals, i.e., behaviors that are not completely under a person's control. This modified theory has been termed the Theory of Planned Behavior. It provides a natural framework for explicit consideration of the facilitating and constraining effects of financial capability. This paper adapts and operationalizes the central construct of this modified theory, 'perceived behavioral control', as 'perceived financial control', and offers an initial test of the Theory of Planned Behavior in a consumer behavior context. Hypotheses consider the differential role played by perceived financial control in predicting purchase of an inexpensive and a relatively-expensive product. The results support the Theory of Planned Behavior and highlight the importance of explicitly including financial capability when predicting purchase.
[ to cite ]:
Arti Sahni (1994) ,"Incorporating Perceptions of Financial Control in Purchase Prediction: an Empirical Examination of the Theory of Planned Behavior", in NA - Advances in Consumer Research Volume 21, eds. Chris T. Allen and Deborah Roedder John, Provo, UT : Association for Consumer Research, Pages: 442-448.

Advances in Consumer Research Volume 21, 1994      Pages 442-448

INCORPORATING PERCEPTIONS OF FINANCIAL CONTROL IN PURCHASE PREDICTION: AN EMPIRICAL EXAMINATION OF THE THEORY OF PLANNED BEHAVIOR

Arti Sahni, University of Cincinnati

ABSTRACT -

While consumer researchers have recognized that financial control over purchase obviously has an impact on purchase, more attention has been devoted to attitude and subjective norms as antecedents of purchase to the neglect of financial capability. The Theory of Reasoned Action has been modified by Ajzen to predict goals, i.e., behaviors that are not completely under a person's control. This modified theory has been termed the Theory of Planned Behavior. It provides a natural framework for explicit consideration of the facilitating and constraining effects of financial capability. This paper adapts and operationalizes the central construct of this modified theory, 'perceived behavioral control', as 'perceived financial control', and offers an initial test of the Theory of Planned Behavior in a consumer behavior context. Hypotheses consider the differential role played by perceived financial control in predicting purchase of an inexpensive and a relatively-expensive product. The results support the Theory of Planned Behavior and highlight the importance of explicitly including financial capability when predicting purchase.

Researchers in diverse fields have been interested in predicting human behavior. A theory that has provided useful direction and generated considerable research is the Theory of Reasoned Action (Ajzen and Fishbein 1980; Fishbein and Ajzen 1975). The Theory of Reasoned Action has received substantial empirical support in consumer and social psychology (Sheppard, Hartwick, and Warshaw 1988).

Recently, however, researchers have recognized the limits of the Theory of Reasoned Action. For example, Bagozzi and Warshaw (1990) draw a distinction between goals and behaviors and point out that the Theory of Reasoned Action (henceforth TRA), as its name implies, is concerned with the prediction of 'behaviors' and not goals. They say, "the key assumptions underlying theories of reasoned behaviors are (1) that action is preceded by a deliberative process culminating in a conscious decision to act and (2) that, if the individual tries to act, no impediments are likely to stand in the way, such as ability limitations, lack of money, environmental contingencies, and unconscious habits. Behaviors subject to such impediments (e.g., purchasing a new house, hiring a top lawyer, stopping smoking) may be considered goals" (Bagozzi and Warshaw 1990, page 127). Thus, a major assumption of the TRA is that the behavior being predicted is under the person's complete volitional control.

While the merit of the TRA lies in its parsimonious representation of complex human decision making, the same can also be its drawback. With the increasing interest in goals versus outcomes, this parsimonious nature proves to be a constraining factor. Specifically, the assumption that the behavior being predicted is under complete volitional control is unrealistic, since most behaviors, even mundane ones, pose problems of behavioral control (Ajzen 1985; Ajzen and Madden 1986).

Purchase behavior has been examined frequently in the TRA framework (e.g., Bonfield 1974; Fishbein and Ajzen 1980; Miniard, Obermiller, and Page 1982; Ryan 1982; Warshaw 1980). However, purchase is often only partially under a person's volitional control because of financial constraints. As Sheppard, Hartwick and Warshaw (1988) suggest, "a variety of consumer activities involve limits on the consumer's ability to perform a given intended action or to achieve a certain outcome. Examples include intentions to purchase expensive items (for which the necessary resources may not be available)" (p. 326). Indeed, predicting purchase of many expensive products can be conceived of as a problem of predicting a goal, and not a behavior.

Recently, there have been attempts by various researchers to modify the TRA to allow for prediction of goals (Ajzen 1985, 1991; Bagozzi 1990; Bagozzi and Warshaw 1990; Warshaw and Davis 1984, 1985; Warshaw, Sheppard, and Hartwick, forthcoming). One direct extension of the TRA is the Theory of Planned Behavior (henceforth TPB, Ajzen 1985, 1991), which relaxes the assumption of complete volitional control. The TPB provides a viable framework for investigating situations where consumers may have only limited financial control over their purchases.

This paper has two related purposes. The first is to conduct an empirical test of the TPB in a consumer behavior context. The second involves adapting and operationalizing the TPB to make it more relevant to the needs of consumer research. This operationalization explicitly considers the role of financial capability in predicting purchase-a factor that has been recognized but nevertheless neglected by consumer psychologists.

THE THEORY OF PLANNED BEHAVIOR

According to the TRA, whether a person performs a behavior depends immediately upon his or her intention. Intention in turn is determined by two factors-attitudes and subjective norms. The constructs specified in the TRA are motivational in nature. For example, the stronger the intention to perform a behavior, the more likely it is that the person will try harder to achieve it. But the performance of many behaviors depends not only on motivation but also on non-motivational factors like the person's ability to actually perform the behavior. Various researchers have suggested that notions of ability and behavioral control be included in models that attempt to predict human behavior (e.g., Ajzen 1985; Kuhl 1985; Liska 1984; Sarver 1983; Triandis 1977).

According to Ajzen's TPB, whenever control over a behavior is not complete, intentions will not be sufficient as the sole predictor of behavior. The TPB features a construct called 'perceived behavioral control' to capture non-motivational factors like ability, availability of resources, co-operation of others, environmental contingencies, etc., that may be required in addition to intention in predicting behavior. Perceived behavioral control is defined as "the person's belief as to how difficult or easy performance of the behavior is likely to be-and beliefs about resources and opportunities may be viewed as underlying perceived behavioral control" (Ajzen and Madden 1986, page 457). Thus, if a person believes she/he possesses the required resources and expects few impediments in reaching the goal, the person perceives greater control over the behavior.

Two versions of the TPB have been offered (see Figure 1). According to version 1, perceived behavioral control is an independent predictor of intention.

This suggests that, even if one has a positive attitude towards the behavior and important others would approve of it, one may not form strong intentions to perform the behavior if perceived control is lacking. This version assumes that the effect of perceived behavioral control on behavior is completely mediated by intention, making intention the sole direct antecedent of behavior.

FIGURE 1

THEORY OF PLANNED BEHAVIOR: VERSION 1 WITHOUT BROKEN ARROW, VERSION 2 WITH BROKEN ARROW

According to version 2, perceived behavioral control, besides having an influence on intention, can also have a direct affect on behavior. This link is proposed because, when it is difficult to perform a behavior, intention will not readily translate to behavior. In such cases, a measure of how easy or difficult a person thinks it would be to engage in the behavior, should provide useful information on the ability of a person to engage in the behavior. Ajzen (1991) offers two reasons for this direct affect. First, if two people have equally strong intentions to engage in a behavior, the one that has less doubts about his ability to achieve it, is more likely to achieve it, perhaps because he is more likely to persevere. Second, because a measure of perceived behavioral control can serve as a substitute for a measure of actual control. In the case that a person from the very outset realistically perceives that it will be problematic for some reason or the other for him to achive the goal, the, even if he sincerely intents to achieve it, it is very likely that he will not be sucessful in his endeavor.

The TPB has received considerable empirical support. According to Ajzen (1991), 16 empirical tests of version 1 have been conducted in the past 5 years and reveal that, "the addition of perceived behavioral control to the model led to considerable improvements in the prediction of intentions; the regression coefficients of perceived behavioral control were significant in every study" (p. 189). Also, 12 empirical tests of version 2 have been conducted and show, "that the combination of intentions and perceived behavioral control permitted significant prediction of behavior in each case, and that many of the multiple correlations were of substantial magnitude" (p. 187).

FINANCIAL CONTROL AND PURCHASE

While consumer researchers have recognized that financial capability has an impact on the purchase of some expensive products, much more attention has been devoted to brand attitude as the primary antecedent of purchase. There could be two reasons for this. One reason could be that, in accordance with the TRA, price, income, and notions of affordability have been treated as external variables whose effect is expected to be mediated by beliefs about the brand or beliefs of important others. However, empirical tests of the TPB in diverse contexts have shown that perceived behavioral control exerts an influence on intention and behavior that is independent of attitudes and subjective norms (Ajzen 1991). Therefore, it might be expected that beliefs about 'financial control' would exert an influence on purchase intention and purchase that is independent of attitude and subjective norm.

A second reason comes from the fact that there is a difference between purchasing a brand from a given product category versus purchasing a given product. Consumer research is predominantly concerned with the former; that is, given a product category, what factors determine brand choice. In this situation the effect of perceived financial control is minimal: beliefs about brands should determine choice of one brand from the consideration set (Hartwick 1983; Warshaw et. al, forthcoming). However, our study is concerned with the likelihood that a person will make a purchase from a general product category. In this situation perceived financial control should be predictive since its effect may not be mediated by beliefs about the product. A measure of whether a person (perceives he) has the required financial resources at his disposal should impact intention to purchase and actual purchase.

When predicting purchase, problems of behavioral control are most likely to be problems of financial control. Therefore, the construct of perceived behavioral control translates to 'perceived financial control' in the consumption context. Thus, the TPB is a framework that allows explicit consideration of the constraining and facilitating effects of financial resources, in combination with the traditional constructs of attitude and subjective norms. As specified in the hypotheses below, financial considerations should improve prediction of both purchase intentions and purchase.

HYPOTHESES

Perceived financial control should enhance prediction of intentions for both inexpensive and expensive products. This is possible since perceived financial control has motivational implications for intentions (Ajzen and Madden 1986). The more a person believes s/he possesses the required financial resources for purchase, the more s/he is likely to form strong intentions to purchase. Thus,

H1: Perceived financial control will add explained variance in the prediction of intention over and above that provided by attitudes and subjective norms.

According to the TPB, the incremental variance added by perceived behavioral control in predicting behavior is inversely related to the degree of control over the behavior (Madden et al. 1992). Hence, when actual financial control is relatively high, perceived financial control will add little to the prediction of behavior. In other words, intentions alone will be sufficient to predict purchase because the financial resources required for purchase are under the person's control.

Whenever actual financial control in a situation is low, the effect of perceived financial control on purchase will not be entirely mediated by intention. Therefore, "to ensure accurate prediction of behavior over which individuals have only limited control, we must assess not only intention but also obtain some estimate of the extent to which the individual is capable of exercising control over the behavior in question" (Ajzen and Madden 1986, page 456). The rationale for this is that, "holding intention constant, the effort expended to bring a course of behavior to a successful conclusion is likely to increase with perceived behavioral control. For instance, even if two individuals have equally strong intentions to learn to ski, and both try to do so, the person who is confident that he can master this activity is more likely to persevere than is the person who doubts his ability" (Ajzen 1991, p. 184). Thus,

H2: Perceived financial control will add explained variance in the prediction of behavior, but this effect would only be expected for relatively-expensive products.

METHODOLOGY

Sample and Procedure

To identify product categories that would allow for a test of the hypotheses, informal one-on-one interviews were conducted with undergraduate students. The students were asked to supply events, and activities associated with those events, that they would consider inexpensive versus relatively-expensive. They also supplied dollar amounts corresponding to each activity in each event category. On the basis of these interviews, 'celebrating the completion of mid-term exams by going out to dinner, a sporting event, or a rock concert' was chosen as the inexpensive scenario; 'buying a gift/dinner for someone for Valentine's day' was chosen as the relatively-expensive scenario.

The study was described as a survey being conducted by the University to determine students' spending habits and activities. Respondents were undergraduate students enrolled in an introductory marketing class during the winter term. Data were collected at two time periods. At time 1, the students indicated their attitude, subjective norm, intention and perceived financial control for both scenarios. The total number of responses obtained was 154. At time 2-two weeks later-behavioral self-reports were solicited for the mid-term celebration and Valentine gift scenarios. Of the 154 students who had participated at time 1, 117 (76%) participated at time 2. The respondents were given class participation credit.

Questionnaire

To anchor the cost aspect of the scenarios, actual dollar ranges that were obtained from the one-on-one interviews were mentioned in the measures. For example, intention for the Valentine scenario read "I intend to buy dinner/ a present (in the range of $25-$50) as a Valentine's day gift". For the mid-term scenario it read "I intend to celebrate the completion of mid-term exams by going out to dinner, a sporting event, or a rock concert, where I might spend $10-$25".

All the constructs were measured on 7-point scales. Attitudes were measured by 7 items: pleasant/unpleasant, boring/interesting, good/bad, unfavorable/favorable, enjoyable/unenjoyable, useful/useless, harmful/harmless. The alpha coefficients for the scenarios were .91 (mid-term exam) and .93 (Valentine).

Subjective norms for each scenario were measured in a global fashion. Normative beliefs and motivations to comply were assessed non-contiguously in the questionnaire with respect to "most people who are important to me". Consistent with the TRA, normative beliefs were combined with motivations to comply to obtain a measure of subjective norms.

Intentions were measured by 3 items for each scenario: "I intend to-", "I will try to-" and "I will make an effort to-". The first item was anchored by "definitely do/definitely do not"; the second and third items by "definitely will/ definitely will not". The alpha coefficients were .96 (mid-term exam) and .97 (Valentine).

Perceived financial control was measured by 4 items: "how much financial control do you have over spending -"; "If I want to, I could easily afford-"; "For me to spend-"; and "My personal income permits me to easily spend-". For the mid-term exam scenario, each item continued as "$10-$25 on celebrating the completion of mid-term exams by going out to dinner, a sporting event, or a rock concert." For the Valentine's day scenario, the items continued as "in the range of $25-$50 on buying dinner/a present as a Valentine's day gift." The 7-point scales were anchored by "complete control/very little control," "extremely likely/extremely unlikely," "easy/difficult," "strongly agree/strongly disagree," respectively. The first three items were adapted from Ajzen and Madden (1986); the last item was created for this study to tap into the affordability and income aspect more directly. An exploratory factor analysis using varimax rotation resulted in a single factor solution for both the scenarios. The first item was found to load weakly (.490 for mid-term, .531 for Valentine) in comparison to the other items, all of which had loadings between .651 and .915. The first item was thus dropped from the scale for both scenarios. The alpha coefficients of the remaining three items were .78 (mid-term exam) and .84 (Valentine).

Behavior self-reports (taken at time 2) were coded as 1 if the subject reported performance of the behavior and 0 if s/he reported the behavior was not performed.

RESULTS

The results are presented by scenario for the hypotheses. Data were analyzed by means of hierarchical regression. For the prediction of intention, attitude and subjective norm were entered in the first step (model 1) to test the TRA. In the second step, perceived financial control was added to the predictors in model 1, resulting in the TPB (model 2). Since behavior is a dichotomous variable, the behavioral data were analyzed by means of logistic regression. Intention as the sole predictor of behavior was entered in the first step to test the TRA (model 1). In the second step, perceived financial control was added, resulting in model 2.

The Inexpensive Scenario

Predicting intentions. The results for prediction of intention are shown in Table 1. It was hypothesized that the inclusion of perceived financial control would add explained variance over and above that provided by attitude and subjective norm in predicting intention. As can be seen from model 1, attitude (b=.58, p<.01) and subjective norm (b=.11, ns) resulted in an R2 of .39. The inclusion of perceived financial control (model 2) resulted in a significant effect (b=.24, p<.01) and increased the R2 to .44. The variance explained by model 2 is significantly greater than that explained by model 1 (F=13.3 (1,149), p<.01). (The formula for testing R-square difference between two models is given in Cohen and Cohen, 1975). These results support hypothesis 1.

Predicting behavior. Out of the 117 subjects available at time 2, 37 (31.6%) reported celebrating the completion of mid-term exams by going out to dinner, a sporting event, or a rock concert. The results for prediction of behavior are shown in Table 2. In logit analysis, the significance of a set of k independent variables is determined by a likelihood-ratio (LR) test. The LR test is the counterpart of the F-test in analysis of variance or regression analysis. This involves computing a LR statistic as follows. First, the model is estimated by constraining and not constraining the impact of the set of k independent variables to zero. Then, corresponding log-likelihood (LL) values denoted as L1 and L2, the LR statistic is computed as 2(L2 - L1) and is chi-square distributed with k degrees of freedom. It was hypothesized that perceived financial control would add no variance beyond that explained by intention in predicting behavior for the inexpensive scenario. Consistent with hypothesis 2, the addition of perceived financial control (model 2) did not add variance to the prediction of behavior beyond that accounted for by intention. This can be seen from the non-significant beta coefficient for perceived financial control and a non-significant LR.

The Relatively-Expensive Scenario

Predicting intentions. The results pertaining to the prediction of intention are shown in Table 3. It was hypothesized that perceived financial control would add variance beyond that provided by attitude and subjective norm in predicting intention. As can be seen from model 1, attitude (b=.41, p<.01) and subjective norm (b=.34, p<.01) resulted in a R2 of .39. Consistent with hypothesis 1, the inclusion of perceived behavioral control (model 2) resulted in a significant coefficient (b=.27, p<.01) and increased the R2 to .45. The variance explained by model 2 is significantly greater than that explained by model 1 (F=16.3 (1,149), p<.01).

TABLE 1

PREDICTING INTENTIONS: CELEBRATING MID-TERM EXAMS

TABLE 2

PREDICTING BEHAVIOR: CELEBRATING MID-TERM EXAMS

Predicting behavior. Out of the 117 subjects available at time 2, 73 (62.4%) reported going out to dinner or having bought a present for someone, while 44 (37.9) reported not having done either. Additionally, out of the 72 that had a significant other, 62 (62.07%) engaged in the behavior, while 10 did not. The results pertaining to the prediction of behavior are shown in Table 4. It was hypothesized that perceived financial control would add variance beyond that explained by intention. The inclusion of perceived financial control (beta=.09, p<.10) added to the prediction of behavior vis-a-vis intention alone (Chi-sq.=3.4; 1 df., p<.10). Thus, the data support hypothesis 2.

DISCUSSION

Predicting Intention

It was hypothesized that perceived financial control would add explained variance to the prediction of intention over and above that furnished by the traditional constructs in the TRA. Consistent with this hypothesis, perceptions of financial control were found to contribute to the prediction of intentions for both the inexpensive and the relatively-expensive scenarios. Thus, perceived financial control serves a motivational function in forming intentions-the higher the financial control, the higher the likelihood of a person intending to perform the focal behavior.

TABLE 3

PREDICTING INTENTIONS: VALENTINE'S DAY GIFT

TABLE 4

PREDICTING BEHAVIOR: VALENTINE'S DAY GIFT

Ajzen also expects interactions between perceived behavioral control and attitude and/or subjective norms. According to Ajzen and Madden (1986), "-perceived behavioral control is a necessary, but not a sufficient condition for the formation of intention to perform a behavior. Besides believing that one could perform the behavior, one must also be inclined to do so for other reasons. These considerations imply the possibility that perceived behavioral control affects intention in interaction with attitude and subjective norm" (page 459, emphasis original). Thus, for the consumption context, this expectation was tested by adding the interaction terms of perceived financial control with attitude and subjective norms to the variables in model 2 which resulted in model 3. For the mid-term exam scenario (model 3, table 1), the perceived financial control x attitude interaction is significant (b=.91, p<.05). In addition, R2 increased from .44 (model 2) to .47 (model 3); this increase is statistically significant (F=4.2 (2,147), p<.05). This shows that the effects of attitude and perceived financial control on intention are best explained by their interaction: simply stated, intention is highest when both attitude and perceived financial control are high.

The possibility of an interaction effect was also tested for the Valentine's day scenario. Here, perceived financial control combined with attitude and subjective norm in an additive manner to predict intention. The interaction terms were not significant (model 3, table 3). When the interaction terms were added, it is interesting to note that the only significant effect was that of subjective norms. This is most likely due to the nature of the context. According to Ajzen (1991), the relative importance of attitudes, subjective norms, and perceived behavioral control in predicting intention will vary across contexts. Ajzen notes that out of 16 empirical tests of the TPB conducted to date, the effect of subjective norms was mixed across contexts, perhaps because personal considerations were more important than social pressures for the behaviors considered in those studies. In a gift-giving scenario like Valentine's day, social norms might be expected to exert greater pressure on consumers' intentions (Netemeyer et al. 1993).

Predicting Behavior

It was hypothesized that perceived financial control would have a direct and independent influence in predicting behavior for the more expensive Valentine's day scenario, but would make little or no contribution for the less costly, mid-term scenario. The results confirm these expectations.

Ajzen also posits an interaction between intention and perceived behavioral control for the prediction of behavior (Ajzen 1985; Ajzen and Madden 1986). According to Ajzen (1991), "the assumption is usually made that motivation and ability interact in their effects on behavioral achievement" (page 183). Thus, in the consumption context, financial control is a necessary condition when purchasing expensive items and reflects the ability component required for purchase; intention reflects the motivational component. To test for the interaction effect, the interaction between intention and perceived financial control was added to the variables in model 2, resulting in model 3. In the mid-term exam scenario, the intention x perceived financial control interaction was not significant. This result supports the TPB, since the consumption of an inexpensive product is under the volitional control of subjects -when this is the case, the TRA is sufficient to predict behavior.

However, perceived financial control and intention interacted to predict behavior in the Valentine's day scenario. It can be seen from model 3 (table 4), that the interaction is significant (beta=.02, p<.05), and the likelihood ratio (LR) test reveals that the fit of model 3 improves beyond model 2 (Chi-sq.=6; 1 df, p<.05). This implies that the likelihood of performance of a focal behavior will be greatest when a person possesses both ability and motivation. In the present study this translates to: the likelihood of purchase of a relatively-expensive item is greatest when financial control and intention are both strong.

According to Ajzen (1991), though interactions are intuitive and are hypothesized, it is puzzling that they rarely emerge. He reports that out of the 12 empirical tests of version 2 of the TPB, 7 hypothesized interactions, but only 1 study (Schifter and Ajzen 1985) evidenced a marginally significant interaction (p<.10) between intention and perceived behavioral control. In fact, because of no empirical support for the interaction between perceived behavioral control and intention, Doll and Ajzen (1992) are considering dropping this expectation from the theory altogether. The results of our study suggest that the interaction hypothesis should be retained in the TPB.

Limitations and Suggestions for Future Research

The present study operationalized perceived behavioral control as perceived financial control with the assumption that problems of control in consumption contexts are most likely to be problems of financial constraints. Future research should also consider factors such as lack of time, location etc. that hamper purchase. In addition, future research should measure both perceived financial control and the more global perceived behavioral control in the same study to examine whether perceived financial control performs better than perceived behavioral control in consumption contexts. Indirect support for perceived financial control comes from Netemeyer et al. (1993) who also tested the TPB in a Valentine's day gift giving context and examined gift items falling in various price ranges. They used the global perceived behavioral control construct and did not find any support for the hypothesis that perceived behavioral control has a direct effect on behavior. The authors suggest that the perceived behavioral control construct may not have captured financial constraints adequately.

While the TPB was supported by our data, more research needs to be done to explain the additive and interactive effects obtained. We are not able to offer any reasons as to why we obtained interactions while prior research has not. Future research should examine when can we expect additive versus interactive effects and vice-versa and what the additive and interactive effects imply about consumer decision processes.

REFERENCES

Ajzen, Icek (1985), "From Intentions to Actions: A Theory of Planned Behavior," in Action-control: From Cognition to Behavior, eds., J. Kuhl and J. Beckman (pp.11-39). Heidelberg: Springer.

Ajzen, Icek (1991), "The Theory of Planned Behavior," Organizational Behavior and Human Decision Processes, 50, 179-211.

Ajzen, Icek and Martin Fishbein (1980), Understanding Attitudes and Predicting Social Behavior, Englewood and Cliffs, NJ: Prentice-Hall.

Ajzen, Icek and Thomas J. Madden (1986), "Prediction of Goal-Ddrected Behavior: Attitudes, Intentions, and Perceived Behavioral Control," Journal of Experimental Social Psychology, 22, 453-474.

Bagozzi, Richard P. (working paper), "On the Neglect of Volition in Consumer Research: A Critique and Proposal," University of Michigan Graduate School of Business Administration, Ann Arbor, MI 49109.

Bagozzi, Richard P. and Paul R. Warshaw (1990), "Trying to Consume," Journal of Consumer Research, 17, 126-140.

Bonfield, E.H. (1974), "Attitude, Social Influence, Personal Norm, and Intention Interactions as related to Brand Purchase Behavior," Journal of Marketing Research, 11, 379-389.

Cohen, Jacob and Patricia Cohen (1975), Applied Multivariate Regression/ Correlation Analysis for the Behavioral Sciences, John Wiley & Sons, NY.

Doll, Jorg and Icek Ajzen (1992), "Accessibility and Stability of Predictors in the Theory of Planned Behavior," Journal of Personality and Social Psychology, 63, 754-765.

Fishbein, Martin and Icek Ajzen (1975), Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research. Reading, MA: Addison-Wesley.

Greene, W.H. (1990), Econometric Analysis, New York: Macmillan.

Hartwick, Jon (1983). Attitudes and Purchase Intentions: Headaches for Fishbein and Ajzen? Paper presented at the annual convention of the American Psychological Association, Anaheim, CA.

Kuhl, J. (1985), "Volitional Aspect of Achievement Motivation and Learned Helplessness: Toward a Comprehensive Theory of Action Control," in Progress in Experimental Personality Research (Vol. 13), ed. B.A. Maher, New York: Academic Press.

Liska, A.E. (1984), "A Critical Examination of the Causal Structure of the Fishbein/Ajzen Attitude-Behavior Model," Social Psychology Quarterly, 47, 61-74.

Madden, Thomas J., Pamela Scholder Ellen and Icek Ajzen (1992), "A Comparison of the Theory of Planned Behavior and the Theory of Reasoned Action," Personality and Social Psychology Bulletin, 18, 3-9.

Miniard, Paul W., Carl Obermiller and Thomas J. Page (1982), "Predicting Behavior with Intentions: A Comparison of Conditional versus Direct Measures," in Advances in Consumer Research (Vol. 9, pp. 461-464), ed. Andrew A. Mitchell, Ann Arbor, MI: Association for Consumer Research.

Netemeyer, Richard G., J. Craig Andrews and Srinivas Durvasula (1993), "A Comparison of Three Behavioral Intention Models: The Case of Valentine's Day Gift-Giving," unpublished manuscript, Marketing Department, Lousiana State University at Baton Rouge.

Ryan, Michael J. (1982), "Behavioral Intention Formation: A Structural Equation Analysis of Attitudinal and Social Influence Interdependency," Journal of Consumer Research, 9, 263-278.

Sarver, V.T. Jr., (1983), "Ajzen and Fishbein's "theory of reasoned action": A Critical Assessment," Journal for the Theory of Social Behavior, 13, 155-163.

Schifter, Deborah E. and Icek Ajzen (1985), "Intention, Perceived Control, and Weight Loss: An Application of the Theory of Planned Behavior," Journal of Personality and Social Psychology, 49, 843-851.

Sheppard, Blair H., Jon Hartwick and Paul R. Warshaw (1988), "The Theory of Reasoned Action: A Meta-Analysis of Past Research with Recommendations and Modifications for Future Research," Journal of Consumer Research, 15, 325-343.

Triandis, Harry C. (1977), Interpersonal behavior, Monterey, CA: Brooks/Cole.

Warshaw, Paul R. (1980), "A New Model for Predicting Behavioral Intentions: An Alternative to Fishbein," Journal of Marketing Research, 17, 153-172.

Warshaw, Paul R.and Fred D. Davis (1984), "Self-Understanding and the Accuracy of Behavioral Expectation," Personality and Social Psychology Bulletin, 10, 111-118.

Warshaw, Paul R. and Fred D. Davis (1985), "Disentangling Behavioral Intention and Behavioral Expectation," Journal of Experimental Social Psychology, 21, 213-228.

Warshaw, Paul R., Blair H. Sheppard and Jon Hartwick (forthcoming), "The Intention and Self-Prediction of Goals and Behavior," in Advances in Marketing Communication, ed., R.P. Bagozzi, Greenwich, CT:JAI Press.

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