The Perceived Behavioral Control Construct: Test of a Twobdimensional Conceptualization For the Consumption Domain

ABSTRACT - While a voluminous body of research has shown that a positive attitude is a major determinant of purchase intentions, it has neglected the effect of resources on purchase intentions. Accordingly, we propose a twoBdimensional construct that taps two major factors that constrain purchase, viz., financial and temporal resources. This construct is a modification of a uniB dimensional construct proposed by Ajzen in the social psychology literature. The results indicate that a measure of a person's resources improves variance explained in purchase intentions over and above that provided by attitude alone. This was true for both the uniB and the proposed twoB dimensional measures. Our assumption that the twoBdimensional construct would perform better than the unidimensional construct was not supported. However, we found that the twoB dimensional measure provides unique information that can potentially be very useful for consumer research.

Citation:

Arti Sahni (1995) ,"The Perceived Behavioral Control Construct: Test of a Twobdimensional Conceptualization For the Consumption Domain", in E - European Advances in Consumer Research Volume 2, eds. Flemming Hansen, Provo, UT : Association for Consumer Research, Pages: 393-397.

European Advances in Consumer Research Volume 2, 1995      Pages 393-397

THE PERCEIVED BEHAVIORAL CONTROL CONSTRUCT: TEST OF A TWOBDIMENSIONAL CONCEPTUALIZATION FOR THE CONSUMPTION DOMAIN

Arti Sahni, University of San Diego

ABSTRACT -

While a voluminous body of research has shown that a positive attitude is a major determinant of purchase intentions, it has neglected the effect of resources on purchase intentions. Accordingly, we propose a twoBdimensional construct that taps two major factors that constrain purchase, viz., financial and temporal resources. This construct is a modification of a uniB dimensional construct proposed by Ajzen in the social psychology literature. The results indicate that a measure of a person's resources improves variance explained in purchase intentions over and above that provided by attitude alone. This was true for both the uniB and the proposed twoB dimensional measures. Our assumption that the twoBdimensional construct would perform better than the unidimensional construct was not supported. However, we found that the twoB dimensional measure provides unique information that can potentially be very useful for consumer research.

Almost everyday, we are confronted with products we would like to purchase. However, we are unable to purchase them because we are constrained by our resources. While purchase prediction is a prime concern of consumer researchers, there has been an alarming neglect of the effect of these constraints on purchase behavior. More effort has been directed towards the process by which attitude or preference affects purchase. A case in point is a very popular theory from the social psychology literature termed the Theory of Reasoned Action (henceforth TRA; Ajzen and Fishbein 1980; Fishbein and Ajzen 1975). This theory has been the subject of consumer research for the last two and a half decades (Sheppard, Hartwick and Warshaw 1988), and continues to be to date (e.g., Bagozzi 1992; Netemeyer et al. 1993). According to this theory (see Figure 1), the best predictor of behavior is intention to perform the behavior. Intention, in turn, is best predicted by a person's attitude towards performing the behavior, and expectations of significant others', i.e., subjective norms. While attitude and subjective norms can no doubt be expected to play an important role in determining behavior, other factors like a person's access to "resources" should play an equally, if not a more important role.

In sharp contrast to psychology, economic theory posits that individuals attempt to maximize their utility, but they achieve this by staying within their budgets. However, due to our heavy reliance on psychology, we seem to have made the same error as the psychologists. The error is that of predicting behavior based solely on preference ratings, and neglecting the costs involved in carrying out the behavior (Meyer 1982).

The objective of this paper is to incorporate resource constraints into a traditional model of purchase prediction like the TRA. The TRA has been extended to address the impact of "behavioral control" problems on behavioral achievement. We believe that this extended theory, the Theory of Planned Behavior, can be used to model resource constraints in purchase prediction. We will perform a partial test of the Theory of Planned Behavior (henceforth TPB; Ajzen 1985; Ajzen 1991) in a consumption context. In addition, we will modify the TPB to better address the needs of the consumption domain. Specifically, we will reconceptualize and operationalize the focal construct of this extended theory, perceived behavioral control (pbc), as a two dimensional construct. Since it is specially crafted for the consumption domain, we believe that this twoBdimensional construct should provide information that the uniBdimensional measure cannot. We will compare the extant pbc measure with our modified twoBdimensional pbc measure and discuss the value of the latter for consumer research.

Several researchers have noticed the lack of attention given to resource constraints in consumer research in general (van Raaij 1981) and TRA in particular (Ajzen 1985, 1991; Kuhl 1985; Liska 1984; Sarver 1983; Triandis 1977; Verhallen and Pieters 1984). The TPB, which is a direct extension of the TRA, can help us address this shortcoming of consumer research. The TPB features an additional predictor of behavior termed perceived behavioral control. Pbc is defined as, "a person's belief as to how easy or difficult 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, p.457). According to this theory, pbc can influence behavior in two ways (see Figure 2). According to version 1, pbc is an independent predictor of intention; the rationale being that if a person perceives control over the behavior, s/he is likely to form strong intentions to perform it. According to version 2, pbc can influence behavior directly. The rationale being that behavioral attainment depends not only on motivation, but also on actual control over the behavior.

We believe that the pbc construct can be used to model the effect of resource constraints in predicting purchase. However, since it has its origins in social psychology, it may not fully reflect the realities of the consumption domain. It is currently operationalized as a uniBdimensional measure consisting of a standard scale comprised of three to four 7Bpoint items (e.g., Ajzen and Driver 1992; Madden et al. 1992). This uniBdimensional measure is expected to result in a global assessment of a person's behavioral control over a particular behavior. We believe that due to multiple, and often competing constraints operating in the consumption context, such a global operationalization may be inadequate for consumer research.

We suggest that the pbc construct be reconceptualized for it to have value in the consumption domain. There are two major factors that constrain purchase which we consider especially important, viz., financial and temporal resources. Accordingly, we propose a twoB dimensional conceptualization of pbc that taps into these two constraints. In proposing this modified construct, we are also motivated by the recurrent criticism by scholars in our field who say that consumer research has borrowed carelessly from other disciplines, and has made little effort to modify constructs for its domain and develop its own theory (e.g., Chakravarti 1992; Olson 1981). (The operationalization of this measure is discussed in the methodology section.)

In this study we will test the TPB in two typical consumption settings. The TPB will be examined for predicting purchase intention for both the uniBdimensional and the proposed twoB dimensional pbc measures. Based on the review above, we expect the following. First, based on previous studies, we expect that a measure of pbc should offer more variance in purchase intention compared to attitude alone. The rationale for this expectation is that while liking (attitude) of the product is important, a measure of whether a person has access to resources should provide additional information in explaining purchase intention. We believe this is also a more realistic representation of the purchase process. Second, the twoBdimensional measure of pbc should perform better than the uniBdimensional pbc measure. The rationale for this expectation is that the extant uniBdimensional pbc measure may not be sensitive enough to tap constraints that operate in the consumption domain. On the other hand, the twoBdimensional pbc measure is crafted to address two major constraints that operate in the consumption domain, viz., financial and temporal. Therefore, it can be expected to provide valuable information that the uniB dimensional measure cannot.

FIGURE 1

THE THEORY OF REASONED ACTION

FIGURE 2

THE THEORY OF PLANNED BEHAVIOR

METHODOLOGY

A survey was administered to 286 undergraduate students enrolled in a business school. Purchase intentions as a function of attitude, subjective norms, perceived behavioral control (measured as uniBdimensional and twoBdimensional) were examined for two consumption activities: buying a compact disc, and going out for a movie (henceforth, referred to as CD and Movie). These contexts were chosen after pilot tests revealed that students engaged in these activities with some regularity. The questionnaire consisted of the following measures for each activity:

Attitude

Attitude towards both the behaviors was measured on a 7Bpoint scale consisting of seven items: pleasant/unpleasant, boring/interesting, good/bad, unfavorable/favorable, enjoyable/unenjoyable, useful/useless, harmful/harmless. These items were summed to serve as a measure of attitude. The coefficient alphas for the CD and Movie activities were .89 and .88 respectively.

Subjective Norms

Subjective norms for both the behaviors were measured in a global fashion on a 7Bpoint scale. Normative beliefs and motivation to comply were assessed nonB contiguously with respect to "most people who are important to me". Normative belief was multiplied by motivation to comply to serve as a measure of subjective norms.

Perceived Behavioral Control

The uniBdimensional pbc measure consisted of four items (Ajzen and Driver 1992; Madden et al. 1992): "If I wanted to, I could easily engage in this activity" (extremely likely/extremely unlikely); "For me to engage in this activity is" (easy/difficult); "the number of events outside my control which could prevent me from" (numerous/very few); and, "I believe I have the resources to perform this activity" (true/false). These four items were summed to form a composite measure of the uniBdimensional pbc. The coefficient alphas for the CD and Movie activities were .97 and .96 respectively.

The twoBdimensional measure was comprised of a financial and a temporal dimension. The financial resource dimension is comprised of three 7Bpoint items: "If I want to, I can easily afford to.." (definitely true/definitely false); "For me to spend.." (easy/difficult); "My personal income permits me to easily.." (strongly agree/strongly disagree). These three items were summed to serve as the financial resource dimension of the twoBdimensional pbc measure. The coefficient alphas for the CD and Movie activities were .92 and .88 respectively.

The temporal dimension consisted of the following three items measured on a 7Bpoint scale: "If I want to, I can easily find the time to.." (definitely true/definitely false); "My schedule permits me to easily find the time to.." (strongly agree/strongly disagree); "For me to find the time to buy.." (easy/difficult). These three items were summed to serve as the temporal resource dimension of the twoBdimensional pbc measure. The coefficient alphas for the CD and Movie activities were .96 and .93 respectively.

TABLE 1

PREDICTING INTENTIONS FROM ATTITUDE, SUBJECTIVE NORMS, AND PERCEIVED BEHAVIORAL CONTROL: THE CD CASE

TABLE 2

PREDICTING INTENTIONS FROM ATTITUDE, SUBJECTIVE NORMS, AND PERCEIVED BEHAVIORAL CONTROL: THE MOVIE CASE

Purchase Intention

The intention measure consisted of three 7Bpoint items: "I intend to.." (definitely do/definitely do not); "I will try to.." (definitely will/definitely will not); "I will make an effort to.." (definitely will/definitely will not). The coefficient alphas for the CD and Movie activities were .97 and .96 respectively.

RESULTS AND DISCUSSION

The data were analyzed by means of multiple regression analysis. To test our expectations, the following three models were estimated:

(1) TRA: intention was regressed on attitude (ATT) and subjective norms (SN) to serve as a base model for comparison purposes.

(2) TPB_ud: intention was regressed on ATT, SN, and the uniBdimensional pbc (PBC_ud) to test the TPB_ud.

(3) TPB_td: intention was regressed on ATT, SN, and the twoBdimensional measure of pbc (PBC_td) to test the TPB_td. PBC_td consisted of a financial resource dimension (PBC_f), and a temporal resource dimension (PBC_t).

The results of the CD and Movie contexts are presented in Tables 1 and 2 respectively. It can be seen from Table 1 that ATT (beta=.424, p<.01) and SN (beta=.247, p<.01) together explain 28% variation in intention. When PBC_ud (beta=.404, p<.01) is added, the variance explained in intention increases to 41.7%. The variance explained by TPB_ud is significantly greater than that explained by TRA (F=65.65 (1, 282); p<.01). (The FBtest for comparing two regression models can be found in Cohen and Cohen, 1975).

Table 1 also shows the performance of the twoBdimensional measure in predicting intentions for the CD context. The PBC_f (beta=.412, p<.01) dimension was a significant predictor of intentions. PBC_td explained 41.9% of variation in intention. The TPB_td explains significantly more variation in intention than the TRA (F=33.74 (2, 281); p<.01). Thus, our first expectation that a measure of pbc should increase variation explained in intention over and above attitude is supported by both the uniB and twoBdimensional pbc measures for the CD context.

Table 2 presents results for the Movie context. It can be seen that ATT (beta=.510, p<.01) and SN (beta=.126, p<.01) together explain 28.5% variation in intention. When PBC_ud (beta=.423, p<.01) is added, the variance explained in intention increases to 41.3%. In addition, the variance explained by TPB_ud is significantly greater than that explained by TRA (F=61.51 (1, 282); p<.01). Table 2 also presents results for the twoBdimensional measure. It can be seen that both PBC_f (beta=.247, p<.01) and PBC_t (beta=.197, p<.01) were significant predictors of intentions. The TPB_td explained 40.5% of the variation in intention. In addition, the variance explained by TPB_td is significantly greater than that explained by TRA (F=28.30 (2, 281); p<.01). Thus, our expectation that a measure of pbc should explain variance in intention over and above that explained by attitude is supported by both the uniB and twoBdimensional measures for the Movie context also. These results for both the contexts show that besides liking a product (as embodied in attitude), measuring a person's access to resources can lead to better estimates of purchase intention. These results are consistent with the perspective of economic theory that recognizes the importance of utility and constraints in modeling consumer choice. Similarly, these results support the viewpoint of several researchers that call for examining the purchase process by examining economic and psychological variables simultaneously (Hogarth and Redder 1987; van Raaij 1981).

It is interesting to note from the results for the TPB_td model in Tables 1 and 2 that the twoBdimensional measure does not perform uniformly in the two contexts. While only the financial resource dimension is significant in the CD context, both the financial and temporal resource dimensions are significant in the Movie context. This shows that while constraints do operate in both contexts, they differ in their nature. While both the uniBdimensional and twoBdimensional measures explain significantly more variation in intentions, the latter provides us with additional information that the former cannot. The twoBdimensional measure informs us that financial resources play a significant role in determining purchase intention for purchasing a CD for our sample. On the other hand, both financial and temporal resources significantly impact purchase intention for going out to a movie. In fact, even economics, which has traditionally been concerned with budget constraints as impacting consumption, has recognized that individuals are constrained by both money and time. More recent papers in economics model time constraints in addition to budget constraints (e.g., Becker 1965, 1985; Tanzi and Zee 1986). Thus, our twoB dimensional conceptualization, and the concommitant results it experienced, are also in sink with this newer economic viewpoint.

Having estimated TPB_ud and TPB_td, we can now compare the performance of the uniB dimensional and twoBdimensional pbc measures in terms of variance explained in purchase intention. We do not require a formal test to comapre these models since a simple eyeballing of the R² values in Tables 1 and 2 indicates that the TPB_td did not perform any better than the TPB_ud. Specifically, the twoBdimensional measure explained only 0.2% more than the uniBdimensional measure in the CD context. In the case of the Movie context shown in Table 2, the twoB dimensional measure explained 0.8% less variation in intentions than the uniBdimensional measure. In terms of variance explained, the twoBdimensional measure performed no better than the uniB dimensional measure. Thus, our second expectation that the twoBdimensional measure would outperform the uniBdimensional measure was not supported.

To summarize, our results show that a measure of a person's resources can explain variation in purchase intention over and above that offered by attitude. We found this to be the case for both the uniBdimensional and the proposed twoBdimensional measures. This result replicates previous research in social psychology that shows that the TPB explains more variation in intentions than the TRA (see Ajzen 1991). In addition, this also shows that the TPB has great potential for the consumption domain. Second, a comparison of the two pbc measures show that the twoBdimensional measure is no better than the uniBdimensional measure in explaining variation in purchase intention. Third, the twoBdimensional measure, however, better reveals the nature of the constraints that operate in a given context. It thus provides us with information on the process of intention formation, something that the uniBdimensional pbc is not in a position to provide. The practical value of this third result is discussed next.

IMPLICATIONS FOR PRACTICE

A vast majority of market research is carried out in the area of predicting purchase intentions from attitudinal type variables (e.g., The Burke Institute; BASES). The results of this study indicate that the proposed twoBdimensional construct may not offer much more than the uniBdimensional measure in terms of predictive ability. However, this study found that the process of purchase intention can be better understood by explicitly measuring a person's perception of his/her access to financial and temporal resources. The proposed twoBdimensional measure can offer value to market researchers by helping to identify segments in a sample. It is plausible that segments based on whether they value time or money can be isolated when such a measure is used. In turn, this information could be used for developing different promotion strategies for the different segments.

LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH

We have proposed a conceptualization of the pbc that addresses two constraints we believe are most prevalent in the consumption context. However, there are no doubt other constraints that make purchase difficult. For example, Bagozzi and Warshaw (1990) suggest that scarce supply of certain goods can make purchase problematic. Future research should examine the value of incorporating other constraints, and examine how they combine with the twoB dimensional measure proposed in this study.

We tested the performance of the uniB and twoBdimensional pbc with respect only to purchase intention. We found that the twoBdimensional measure did not explain more variation in intention than the uniBdimensional measure. It, however, provided more information as to the nature of the constraints that operate in each scenario. Future research should compare the performance of these measures with respect to actual purchase.

REFERENCES

Ajzen, Icek (1991), "The Theory of Planned Behavior: Some Unresolved Issues," Organizational Behavior and Human Decision Processes, 50, 179B211.

Ajzen, Icek (1985), "From Intentions to Actions: A Theory of Planned Behavior," in ActionB Control: From Cognition to Behavior (pp. 11B39), eds. J. Kuhl and J. Beckman, Heidelberg: Springer.

Ajzen, Icek and B. L. Driver (1992), "Contingent Value Measurement: on the Nature and Meaning of Willingness to Pay," Journal of Consumer Psychology, 1 (4), 297B316.

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

Ajzen, Icek and Thomas J. Madden (1986), "Prediction of GoalBDirected Behavior: Attitudes, Intentions, and Perceived Behavioral Control," Journal of Experimental Social Psychology, 22, 453B474.

BASES: Simulating Market Demand for New Services, A Division of BBI Marketing Services, Inc., 1994 brochure, Covington, KY.

Bagozzi, Richard P. and Paul R. Warshaw (1990), "Trying to Consume," Journal of Consumer Research, 17, 127B140.

Bagozzi, Richard P. (1992),, "The SelfBRegulation of Attitudes, Intentions, and Behavior," Social Psychology Quarterly, 55, 178B204.

Becker, Gary S. (1965), "A Theory of Allocation of Time," Economic Journal, 75 (September), 493B517.

Becker, Gary S. (1985), "Human Capital, Effort, and the Sexual Diversion of Labor," Journal of Labor Economics, 3 (January), S33BS58.

The Burke Institute: Center for Applied Marketing Research, A Division of BBI Marketing Services, Inc., 1994 brochure, Covington, KY.

Chakravarti, Dipankar (1992), "Appraising Consumer Research: There's More to Vision than Meets the Eye," Journal of the Academy of Marketing Science, 20, 361B366.

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

Fishben, Martin (1980), "A Theory of Reasoned Action: Some Applications and Implications," in Nebraska Symposium on Motivation, eds. H. Howe and M. Page, Lincoln, NB: University of Nebraska Press, 65B116.

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

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. Mather, New York: Academic Press.

Liska, A. E. (1984), "A Critical Examination of the Causal Structure of the Fishbein/Ajzen AttitudeBBehavior Model," Social Psychology Quarterly, 47, 61B74.

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 (1), 3B9.

Meyer, Willi (1982), "The Research Programme of Economics and the Relevance of Psychology," British Journal of Social Psychology, 21, 81B91.

Netemeyer, Richard G., J. Craig Andrews and Srinivas Durvasula (1993), "A Comparison of Three Behavioral Intention Models: The Case of Valentine's Day giftBGiving," in Advances in Consumer Research (Vol. 20), eds. Leigh McAlister and Michael Rothschild, Vancouver, BC, Canada.

Olson, Jerry (1981), "ACR Presidential AddressBToward a Science of Consumer Behavior," in Advances in Consumer Research, ed. Andrew A. Mitchell, Provo, UT: Association for Consumer Research, 1B9.

Sheppard, Blair A., Jon Hartwick, and Paul R. Warshaw (1988), "The Theory of Reasoned Action: A MetaBAnalysis of Past Research with Recommendations for Modifications and Future Research," Journal of Consumer Research, 15 (December), 325B343.

Tanzi, Vito and Howell H. Zee (1993), "Time Constraints in Consumption and Savings Behavior," the Journal of Public Economics, 50, 2, 253B257.

Triandis, H. C. (1977), Interpersonal Behavior, Monterey, CA: Brooks/Cole.

van Raaij, W. Fred (1981), "Economic Psychology," Journal of Economic Psychology, 1, 1B24.

Verhallen, Theo M. and Rik G. M. Pieters (1984), "Attitude Theory and Behavioral Costs," Journal of Economic Psychology, 5, 223B249.

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