Perceived Control in Consumer Choice: a Closer Look

ABSTRACT - This paper theorizes that in most choice situations within the consumer-marketplace domain, perceived control reduces to consumer confidence in the context of making the Aright@ choice, and consequently, perceived control in consumer choice may be measured as confidence in making the Aright@ choice. It further opines that utilitarian choice models should explicitly account for such perceived control in order to enhance predictive validity, and makes a case for this by comparing the Aright@ choice hit rates of a utilitarian approach with that of an attitudinal approach in light of perceived control.

Citation:

Rajan Nataraajan and Madhukar G. Angur (1998) ,"Perceived Control in Consumer Choice: a Closer Look", in E - European Advances in Consumer Research Volume 3, eds. Basil G. Englis and Anna Olofsson, Provo, UT : Association for Consumer Research, Pages: 288-292.

European Advances in Consumer Research Volume 3, 1998      Pages 288-292

PERCEIVED CONTROL IN CONSUMER CHOICE: A CLOSER LOOK

Rajan Nataraajan, Auburn University, U.S.A.

Madhukar G. Angur, University of Michigan-Flint, U.S.A.

ABSTRACT -

This paper theorizes that in most choice situations within the consumer-marketplace domain, perceived control reduces to consumer confidence in the context of making the "right" choice, and consequently, perceived control in consumer choice may be measured as confidence in making the "right" choice. It further opines that utilitarian choice models should explicitly account for such perceived control in order to enhance predictive validity, and makes a case for this by comparing the "right" choice hit rates of a utilitarian approach with that of an attitudinal approach in light of perceived control.

INTRODUCTION

The area of "perceptions of control" has interested social psychologists and researchers from related disciplines for several decades, and there is a rich body of literature particularly on the aspect of perceived control pertaining to human judgment and behavior (e.g., Averill, 1973; Ajzen & Madden, 1986; Bagozzi & Warshaw, 1990; Langer, 1983; Rotter, 1966; Schifter & Ajen, 1985; Schulz & Decker, 1985; Steiner, 1970). Yet the role of perceptions of control has received very limited attention in consumer behavior settings (Hui & Bateson, 1991). In particular, the literature does not reveal an interpretation of perceived control that is specific to the consumer choice context. However, it reveals an interpretation of perceived control in a choice setting in general. Although perceived choice has been referred to as one type of "decisional control" (Averill, 1973), subsequent interpretations and resulting operationalizations have typically restricted perceived control in a choice situation to whether it is a person’s own decision to enter or stay in a situation. An example of this type of interpretation is seen in the work of Hui and Bateson (1991). Other examples include the works of Madden, Ellen and Ajzen (1992) and Mathur (1996), although they have used multiple indicators to capture this.

The purpose of this paper is two fold: 1. To present a succinct interpretation of perceived control in a choice setting in general, and then tailor it to the typical consumer choice context. 2. To give credibility to the idea that utilitarian choice models should explicitly consider such perceived control as part of the model structure in order to enhance realism, and thus increase predictive validity. Toward this end, we report the findings of an empirical study wherein the "right" choice hit rates from a utilitarian approach and an attitudinal approach were compared under two levels of perceived control.

PERCEIVED CONTROL IN A CHOICE SETTING IN GENERAL

Perceived control in a choice setting is more complex than in other settings. This is because the choice situation involves two stages, and different perceptions of individual control may occur in each stage. The first stage concerns the enactment of choice behavior per se. In this stage, perceptions of control essentially culminate in the volitional control that a person thinks he/she has over the decision to enact or not enact choice behavior itself. This is in line with the traditional interpretation of perceived control in social psychology as the concept typically pertains to one intended behavior (e.g., stopping alcohol consumption, losing weight etc.). The enactment of choice behavior is the intended behavior in the choice context.

The second stage pertains to the actual choice process, and here the rational person would naturally desire to make the "right" choice among the available alternatives in the choice set. However, this is not always easily accomplished. Sometimes the person knows a great deal about a product and brands and is sure of not making a mistake in choice; at other times the person is not so confident. Sometimes the person can very easily find enough reliable information from the store/package /manual/ elsewhere to base the selection on; at other times it may not be so easy. In some cases, after-sale service and warranty conditions are always straight forward and easy to understand; in other cases they may be confusing. In any case, it is clear that with some products/services the choice process is a cinch; with others it becomes relatively more difficult or down right frustrating. And of course, this ease/difficulty level varies from product to product, from consumer to consumer, and perhaps, from context to context (e.g., purchasing something for oneself; purchasing something for somebody else).

Thus, in this second stage, perceptions of control would essentially depend on the degree of confidence that the person has over making the "right" choice, given that the person indeed enacts choice behavior. Mathematically, if the first stage is represented by the marginal probability P_B (the probability of enacting choice behavior), and the second stage by the conditional probability P_C/B (the probability of making the "right" choice given that choice behavior is enacted),then perceived control over making the "right" choice in a choice situation is given by the joint probability,

P_R = (P_B * P_C) = (P_B) . (P_C/B).

PERCEIVED CONTROL IN THE CONSUMER CHOICE CONTEXT

However, in choice situations within the consumer-marketplace domain, the above interpretation may become simpler. While there may be exceptions (e.g., a child consumer under the supervision of a parent may think that he/she has complete volitional control when in fact he/she may have no actual control over the decision whether or not to enact choice behavior in the marketplace), in the majority of choice situations within the consumer-marketplace domain, the rational adult consumer typically thinks and in fact does have complete volitional control over the decision whether or not to enact choice behavior. In other words, the first stage is a certainty, or P_B = 1. Thus, P_R simply reduces to P_C/B.

Further, a rational consumer would always like to choose the "right one" among the available brands in the market. Success in choosing the "right one" will result in the consumer experiencing positive consequences including enhanced confidence in one’s decision making ability. Failure to choose the "right one," on the other hand, will likely yield negative consequences pertaining to buyer remorse or dissonance. In effect, perceived control in a consumer choice setting typically reduces to his/her level of confidence in making the "right" choice.

In other words, such perceived control is purely on human judgment and not on behavior, and follows the concept of self-efficacy of beliefs advanced by Bandura (1977; 1982). Therefore, such perceived control may be measured as "confidence" in judgment. The reasonableness of this interpretation is further substantiated by a quote from Ajzen and Madden (1986, p. 457): "Bandura and his associates (e.g., Bandura, Adams, & Beyer, 1977; Bandura, Adams, Hardy, & Howells, 1980) have provided evidence showing that people’s behavior is strongly influenced by their confidence in their ability to perform it (i.e., by perceived behavioral control)." Note that Ajzen and Madden (1986) interpret "confidence" as perceived control.

MODELING CONSUMER CHOICE

Since perceived control in a consumer choice setting typically reduces to his/her level of confidence in making the "right" choice, we can assess an individual’s perceived control by measuring such confidence. This reduction makes the measurement of perceived control simpler and straight forward. While researchers have been trying to finetune models to enhance their predictive abilities, low predictive validity continues to be a paramount concern in consumer behavior. Considering that the choice process is ubiquitous in the consumer-marketplace domain, and that the prediction of consumer choice is a marketer’s ultimate game, this concern is legitimate.

While the vast majority of choice models stem from a utilitarian base, there are a few attitudinal models that may be used for predicting consumer choice (e.g., the theory of reasoned action; Ajzen & Fishbein, 1980). However, there is no evidence that these attitudinal models have been widely used for consumer choice prediction. The lack of popularity of attitudinal models for choice prediction could possibly be attributed to the perception that they are cumbersome and not specifically tailored to the choice situation. Yet these models appear to have a greater element of realism embedded in them because they include psychological variables and are essentially contextdependent. On the other hand, the utilitarian models which comprise the -vast- majority class of choice models, typically suffer from a severe paucity of psychological variables. They are also essentially context independent. For instance, the basic conjoint model is fundamentally context free (Rao, 1977). However, within a certain research, contexts may be created for respondents. Green, Helsen, and Chandler (1988) have surmised that an explicit consideration of psychological variables in conjoint analysis would inject in such techniques a significant dose of realism, which, in all likelihood, would help in enhancing predictive validity. In any case, although hybrid conjoint models have the self-explicated component which is obtained in a manner that bears a close similarity to data obtained for a belief/attitude type model, to our best knowledge, conjoint researchers are yet to incorporate something even akin to perceived control in their modeling let alone incorporating the simplified interpretation of it as furnished in the previous section.

Some of the relatively recent attitudinal models include perceived control as a specific variable. Nevertheless, two points must be mentioned. First, these models are purely psychological models that have a broader purpose than predicting consumer choice behavior. While, in theory, they may be used for predicting consumer choice behavior, they are largely not used for such specific purpose. So, in the realm of predicting consumer choice, it is the mathematical psychology based utilitarian techniques that reign supreme. Second, as with the utilitarian techniques, none of these attitudinal models interpret perceived control in the specific realm of consumer choice, the way it has been done in the previous section.

Yet it is plausible that an explicit recognition of perceived control as a concrete, measurable aspect in utilitarian based choice modeling may have the potential to enhance the explanatory as well as the predictive powers of such utilitarian techniques. To aid in giving credence to this thinking, an empirical study was conducted wherein the "right" choice hit rates of traditional conjoint analysis (possibly the most well known utilitarian approach) and the theory of reasoned action (possibly the most well known attitudinal approach) were assessed under two levels of perceived control.

THE STUDY

The interested reader may refer to the works of Green and Rao (1971), Green and Srinivasan (1978), and Rao (1977) for a quick review of traditional conjoint analysis, and to the works of Ajzen and Fishbein (1980) and Sheppard, Hartwick and Warshaw (1988) for a comprehensive review of the theory of reasoned action. For a theoretical comparison of conjoint analysis and the theory of reasoned action, one may refer to Nataraajan and Warshaw (1991).

The hypothesis underlying the application of conjoint analysis to consumer choice situations is that a consumer enters a choice situation with a predefined algebraic judgment policy or utility function which defines how the observed attributes of products will be integrated to form overall evaluations of utility. The consumer is expected to independently evaluate each available alternative, and then choose the option with the highest total utility. The hypothesis underlying the application of the theory of reasoned action to consumer choice situations would be that a consumer has a certain level of BI (behavioral intention, the predictor of behavior) for each of the available alternatives, and would end up selecting that alternative for which he/she has the highest level of BI.

Hypotheses

Conjoint analysis is primarily object-based (as opposed to person-based), context free (Rao, 1977), sterile, without a normative component, and does not assume that consumers actually process cognitively according to the maximization principle (subjectto constraints) but only that consumers choose as if they did. The theory of reasoned action, on the other hand, is primarily person based, extremely contextual, and has a normative component (the subjective norm).

In light of the above differences, and given that perceived control in consumer choice is, in effect, confidence in judgment, it would be reasonable to posit that when consumer confidence in making the "right" choice is low, a primarily person-based model would be a relatively more accurate predictor of the "right" choice than a primarily object based model. First, one would expect the normative component (the subjective norm) in the theory of reasoned action to instill some predictive accuracy especially when consumer confidence in making the "right" choice is low. The consumer may rely on the opinions of important others, and thus, his/her responses to the normative component may reflect what he/she will choose. Second, the validity of the part-worths resulting from conjoint analysis obviously depends on the accuracy of the profile ratings furnished by the subject. Needless to say, if consumer confidence in making the "right" choice is low (probably due to lack of sufficient product knowledge), then he/she is not likely to rate object-based profiles in a manner reflecting his/her actual choice.

In view of the above, it seems reasonable to speculate that when confidence in making the "right" choice is low, the theory of reasoned action would be a relatively better predictor of consumer choice than traditional conjoint analysis. When such confidence is high however, it is speculated that there would be no difference in the predictive abilities of the two models. In other words, a high level of confidence in making the "right" choice would propel the subject toward giving equally "right" input to the models. Based on the above, it is hypothesized that:

H_1a: Under conditions of low perceived control, an attitudinal approach is significantly more effective than a utilitarian approach in predicting consumer choice.

H_1b: Under conditions of high perceived control, there is no significant difference between a utilitarian approach and an attitudinal approach in the effectiveness of predicting consumer choice.

EXHIBIT

35mm CAMERA ATTRIBUTES AND LEVELS

METHOD

The predictive abilities of the two models as indicated by their "right" choice hit rates were examined under two levels (#high’ and #low’) of perceived control in a hypothetical purchase setting involving six comparable (within a certain price range) real 35mm cameras. A total of 136 undergraduate business students enrolled at a North Eastern university formed the test sample out of which 70 (51%) were males and 66 (49%) were females.

Perceived control was measured as the degree of confidence in choosing the "right" brand using a single, carefully worded scale that asked the respondent to indicate the likelihood of his/her choosing the best (for him/her and for his/her situation) among six brands of 35mm cameras without assistance from anyone else on a 7 point likely-unlikely scale. This scale was preceded by the following explanation of "degree of confidence."

Every time a consumer goes to the market to purchase a product, he/she is forced to make a choice among the various brands available in the market. Sometimes the consumer knows a great deal about a product and brands and is sure that he/she will not make a mistake in his/her choice; at other times he/she i not so confident. Sometimes the consumer can very easily find enough reliable information from the store/package /manual/ elsewhere to base his/her selection on; at other times it may not be so easy. In some cases, after-sale service and warranty conditions are always straight forward and easy to understand; in other cases they may be confusing. In any case, it is clear that with some products/ services the choice process is a cinch; with others it becomes relatively more difficult or down right frustrating. And of course, this ease/difficulty level varies from product to product, from consumer to consumer, and perhaps, from context to context (e.g., purchasing something for oneself; purchasing something for somebody else).

However, despite the fact that it is not always an easily accomplished task, a rational consumer would always like to choose the BEST (for him/her, and for his/her situation) among the available brands in the market. Note that if the consumer succeeds in choosing the BEST brand, then he/she will experience one or more positive consequences (e.g., getting a good deal, happiness, satisfaction, enhanced confidence in one’s decision making ability). On the other hand, if the consumer fails to choose the best, then he/she is likely to face one or more negative consequences (e.g., dissonance, dissatisfaction, getting a bad deal).

Pretests had indicated that this type of instrument was actually better than either a single scale with a short explanation without examples or a multi-scale approach preceded by short or long explanations pertaining to perceived control.

The scores of the subjects were rank ordered and split into upper and lower halves which became the #high’ (n=68) and #low’ (n=68) "perceived control" groups respectively. However, to avoid any kind of testing effect, groups were not overtly formed; the subjects were unaware that they were being classified.

Every subject completed both the conjoint analysis and the theory of reasoned action questionnaires. However, the order of assignment of these questionnaires was randomized to avoid order bias. The camera attributes and levels used in the conjoint design are shown in the Exhibit. An orthogonal main effects design through CONJOINT DESIGNER yielded 16 profiles. Before rating the 16 profiles, subjects were asked to imagine that they were going to purchase a 35mm camera for their own use in the next month (the hypothetical purchase setting). To facilitate their rating task, they were asked to adopt a #sort’ and #rate’ process. The questionnaire exhibited a likelihood of purchase (coded 1=not at all likely to purchase, to 9=extremely likely to purchase) scale for each profile.

For the holdout set, six real 35mm cameras were picked from camera catalogs after ensuring that they could be completely described in terms of the appropriate levels of the attributes used for profile generation in conjoint analysis. These were Nikon, Kodak, Vivitar, Canon, Olympus and Pentax.

For data via the theory of reasoned action, subjects were asked to go through the descriptions of the six real brands, and then fill out a questionnaire under the same hypothetical purchase setting as in conjoint analysis. The items were AB_y (attitude toward behavior -purchasing a particular brand- for each subject and for each brand; y=brand) and SN_y (subjective norm -influencing the purchase of a particular brand- for each subject and for each brand). These were patterned after the standard formats furnished by Ajzen and Fishbein (1980) and measured directly. Since prediction and not explanation was the main interest, the belief and evaluation stage of the theory of reasoned action (the ¦b_ie_i and ¦NB_jMC_j terms) was deemed quite unnecessary, and measurement took place only in the subsequent stages. This is in line with what Ajzen and Fishbein (1980) have suggested, and others (e.g., Warshaw, 1980) have followed.

TABLE

"RIGHT" CHOICE HIT RATES

AB_y was computed as a summated score of four semantic diffrential scales [For summation, the unidimensionality of the four scales was confirmed through factor analysis.] (coded from +3 to -3 or reversed depending on the polarity) with the following bipolar adjectives: 1. Pleasurable-Painful 2. Wise-Foolish 3. Unpleasant- Pleasant 4. Punishing-Rewarding. An example is shown below:

My choosing the Canon would be

punishing ___ ___ ___ ___ ___ ___ ___ rewarding

extremely quite slightly neither slightly quite extremely

SN_y was measured using a scale (coded from +3 to -3) of the type shown below:

Most people who are important to me (friends, family, et al.) would think I should choose the Canon.

likely ___ ___ ___ ___ ___ ___ ___ unlikely

extremely quite slightly neither slightly quite extremely

One month after the prediction phase ended, every one of the 136 subjects was asked to indicate his/her choice assuming that he/she was going to choose and purchase one of the six holdout brands for his/her own use. [Unlike many other predictive validity studies (where there is virtually no time gap between the prediction phase and the actual choice behavior phase), we wanted to minimize if not obliterate memory effects. One month seemed like a safe bet to accomplish that.]

For conjoint predictions, part worths for all attribute-levels for every subject were computed using profile ratings through CONJOINT ANALYZER, [The assumption that a linear model fitted every subject was verified.] and then a total utility score for each holdout brand for every subject was computed. Thus, 6 total utility scores were computed for each subject. In each case, the brand with the highest total utility score was deemed the "right" choice of a subject as predicted by conjoint analysis.

For predictions via the theory of reasoned action, the dependent variable, BI_y, (behavioral intention -to purchase a particular brand- for each subject; y=brand) was computed as BI_y= w_1y. AB_y+ w_2y.SN_y, where w_1y and w_2y were the brand specific beta weights. [As suggested by Ajzen and Fishbein (1980), these weights were estimated from a separate and smaller sample of subject; it may be noted that weights were estimated separately for the 'high' and 'low' groups.] Thus, for each subject, 6 BI_y scores were computed. In each case, the brand with the highest score was deemed the "right" choice of a subject as predicted by the theory of reasoned action.

The predicted "right" choices were compared to the actual "right" choices, and hit rates were computed. McNemar’s test for related samples was used to test the significance of the difference in hit rates of the two models.

RESULTS AND DISCUSSION

The Table reports the results for both models under the two levels of perceived control, and they support the stated hypotheses fully. While there is no significant difference in the hit rates of the theory of reasoned action and conjoint analysis when perceived control is high, the difference is statistically significant in favor of the theory of reasoned action when perceived control is low.

Additionally, while the difference between hit rates for the two groups is significant for the theory of reasoned action, it is not significant for conjoint analysis. The theory of reasoned action is clearly more sensitive to perceived control in consumer choice even if such "control" variable is not an integral part of the model. Traditional conjoint analysis, on the other hand, seems fairly insensitive to perceived control, which underscores the mechanistic, object oriented and context free approach inherent in the conjoint measurement model.

The results point to the theory of reasoned action as a more realistic predictor of consumer choice. Yet it is clearly cumbersome in the sense that prediction is not possible without exposing the holdout set to the respondent in the prediction phase itself. Regardless of the positive attributes heaped on such attitudinal approaches (see Ajzen and Fishbein, 1980), this is a major weakness which, in all likelihood, heavily contributes to the perception that attitudinal approaches are cumbersome for predicting consumer choice. The beauty of conjoint analysis lies in the fact that holdout is never exposed to the respondent during the prediction phase, and prediction of eventual consumer choice among real brands is done purely with artificially generated profiles. This is undoubtedly a major strength and lends tremondous support for the greater usage and popularity of such utilitarian approaches in the choice context. Nevertheless, the findings point to one important implication despite the fact that the study did not involve a real life situation with real purchase behaviors. Utilitarian approaches for use in predicting consumer choice would become more sensitive and therefore more realistic through the explicit consideration of "perceived control" during the model development phase itself. For instance, although traditional conjoint analysis may not offer much scope in this regard, the self-explicated component in hybrid models may be refined to capture perceptions of control.

Finally, the qualification "right" needs clarification. It is to be noted that "right choice hit rates" were, in effect, the traditional first choice hit rates. "Right" seemed a more appropriate qualification than "first" given the way perceived control was operationalized in this study. However, it is acknowledged that although every consumer would always like to pick the option that would be "right" for him/her, this does not mean that he/she did in fact pick the absolute right choice. It can only be concluded that he/she picked what seemed right (will bring happiness, no cognitive dissonance etc.) at that time. As an analogy, many men/women may now feel that their choices for spouses were not right but they seemed "right" at that moment. Some of them realize this after five minutes, others after 50 years! In any case, the study involved a hypothetical purchase situation and so the question of real experience with the product did not arise.

Also, "right" does not mean that the choice has to be economically rational or even be rational in the eyes of others. The choice has to be rational only to the person effecting that choice.

REFERENCES

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall.

Ajzen, I., & Madden, T. J. (1986). Prediction of goal-directed behavior: Attitudes, intentions, and perceived behavioral control. Journal of Experimental Social Psychology, 22, 453-474.

Averill, J. R. (1973). Personal control over aversive stimuli and its relationship to stress. Psychological Bulletin, 80, 286-303.

Bagozzi, R. P., & Warshaw, P. R. (1990). Trying to Consume. Journal of Consumer Research, 17, 127-140.

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191-215.

Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122-147.

Bandura, A., Adams, N. E., & Beyer, J. (1977). Cognitive processes mediating behavioral change. Journal of Personality and Social Psychology, 35, 125-139.

Bandura, A., Adams, N. E., Hardy, A. B., & Howells, G. N. (1980). Tests of the generality of self-efficacy theory. Cognitive Therapy and Research, 4, 39-66.

Green, P. E, Helsen, K., & Shandler, B. (1988). Conjoint internal validity under alternative profile presentation. Journal of Consumer Research, 15, 392-397

Green, P. E., & Rao, V. R. (1971). Conjoint measurement for quantifying judgmental data. Journal of Marketing Research, 8, 355-363.

Green, P. E., & Srinivasan, V. (1978). Conjoint analysis in consumer research: Issues and outlook. Journal of Consumer Research, 5, 103-123.

Hui, M. K., & Bateson, J. E. G. (1991). Perceived control and the effects of crowding and consumer choice on the service experience. Journal of Consumer Research, 18, 174-184.

Langer, E. J. (1983). The psychology of control. Beverly Hills, CA: Sage.

Madden, T. J., Ellen, P. S., & Ajzen, I. (1992). A comparison of the theory of planned behavior and the theory of reasoned action. Personality and Social Psychology Bulletin, 18, 3-9.

Mathur, A. (1996). An examination of trying as a moderator and control as a moderator of intention-behavior relationship. Working paper. Hofstra University, NY.

Nataraajan, R., & Warshaw, P. R. (1991). Are the models compatible for empirical comparison? An illustration with an intentions model, an expectations model, and traditional conjoint analysis. In Advances in Consumer Research (Vol. XIX, pp. 472-481), the 1991 Proceedings of the Association for Consumer Research Conference, Chicago, IL.

Rao, V. R. (1977). Conjoint measurement in marketing analysis, in Multivariate Methods for Market and Survey Research. Jagdish N. Sheth, ed., American Marketing Association, Chicago. 257-286.

Rotter, J. B. (1966). Generalized expectancies for internal veruss external control of reinforcement. Psychological Monographs, 80 (1, Whole No. 609).

Schifter, D. E., & Ajzen, I. (1985). Intention, perceived control, and weight loss: An application of the theory of planned behavior. Journal of Personality and Social Psychology, 49, 843-851.

Schulz, R., & Decker, S. (1985). Long-term adjustment to physical disability: The role of social support, perceived control, and self blame. Journal of Personality and Social Psychology, 48, 1162-1172.

Sheppard, B. H., Hartwick, J., & Warshaw, P. R. (1988). The theory of reasoned action: A meta-analysis of past research with recommendations for modifications and future research. Journal of Consumer Research, 15, 325-343.

Steiner, I. D. (1970). Perceived freedom. In L. Berkowitz (Ed.), Advances in Experimental Socil Psychology (Vol. 5, pp. 187-248). New York: Academic Press.

Warshaw, P. R. (1980). A new model for predicting behavioral intention: An alternative to Fishbein.Journal of Marketing Research, XVII (May), 153-172.

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