Advances in Consumer Research Volume 23, 1996 Pages 218-224
DECISION PROCESSES OF THE ATTRACTION EFFECT: A THEORETICAL ANALYSIS AND SOME PRELIMINARY EVIDENCE
Lianxi Zhou, Concordia University
Chankon Kim, Concordia University
Michel Laroche, Concordia University
Drawing from theories of adaptive behavior and reasons in choice, this study proposes two decision processes leading to the attraction effect. First, an attribute-based process suggests that this effect results from consumers' evaluations, comparisons, or other computations on attribute values. Second, a reason-based process implies that this effect results from consumers' reliance on a dominance and/or a compromise relationship contained in the choice task. Retrospective protocols obtained from a pilot study provide support for the reason-based process but not for the attribute-based process. In addition, theoretical and marketing implications of this study are briefly discussed.
There is much evidence showing the existence of the attraction effect (e.g., Huber, Payne, and Puto 1982; Huber and Puto 1983; Ratneshwar, Shocker, and Stewart 1987; Simonson 1989; Pan and Lehmann 1993; Pan, O'Curry, and Pitts 1995). This effect describes the phenomenon that a new alternative, when added to a choice set, increases the preference and choice probability of an existing alternative. Major findings reported in this area indicate that the attraction effect occurs when the choice set consists of an asymmetrically dominated, a relatively inferior, or a compromise alternative. There is also evidence showing that the attraction effect is influenced by the ambiguity of the information presented about the attribute values of the alternatives. It has been shown that the attraction effect is greater when there is information ambiguity in the choice task than when there is not (Ratneshwar, Shocker, and Stewart 1987; Mishra, Umesh, and Stem 1993). Ambiguity of attribute information may arise from the consumer's lack of knowledge with the product category or when the attribute is fuzzy or less meaningful (Mishra, Umesh, and Stem 1993). Accountability is another construct which has been examined for its impact on the occurrence of the attraction effect. Simonson (1989) reports that the attraction effect was stronger when consumers expected to justify their choices to others.
A number of explanations for the attraction effect have been proposed, including the use of choice rules or strategies, perceptual framing of the decision problem, need for justification, changes in attribute importance, changes in subjective brand evaluations, and consideration set memberships. Past research provides mixed results for these explanations, suggesting the need for further study in this area.
Most of the previous studies investigating the attraction effect used aggregate choices as the dependent variable. As a result, much remains to be known as to the basic processes leading to the attraction phenomenon. The use of process tracing methods such as verbal protocols (Ericsson and Simon 1980) would be potentially useful in gaining direct insights into the reasons for the occurrence of the attraction effect. Against this background, the present study attempts to improve our understanding of the decision processes or mechanisms that underlie the attraction effect. We first present a theoretical analysis of the attraction process and then report preliminary empirical evidence.
Adaptive Decision Making
Numerous empirical findings of decision research support the notion that the decision process is governed by a number of rules or strategies (Abelson and Levi 1985). Decision strategies are adaptive to a particular choice task (Payne 1982). Adaptation may occur in two contrasting modes: top-down process and bottom-up process (Payne, Bettman, and Johnson 1992). The cost/benefit framework for strategy selection implicitly assumes a top-down adaptation, whereas the constructive view of decision making reflects a bottom-up adaptation.
There is much evidence showing the adaptive use of heuristics in a more bottom-up fashion (Payne, Bettman, and Johnson 1988; 1992; Klein and Yadav 1989). Support for such adaptive decision making comes from the fact that decision makers tend to use multiple strategies in arriving at a final choice (Payne 1976; Lussier and Olshavsky 1979; Gertzen 1992). The use of hybrid strategies or phased heuristics in decision making implies that people adapt on-line during the decision. Specifically, individuals adapt to changing environments and use combinations of different strategies, often constructing a strategy as they proceed (Bettman and Zins 1977). Of particular importance is the finding that an elimination strategy, such as elimination by aspects, is often used to reduce the choice set to a manageable size and the remaining alternatives are then processed in a more compensatory manner (Johnson and Puto 1987).
The choice set typically used in the past studies has been relatively simple, consisting of two or three alternatives and two attribute dimensions per alternative. According to the view of adaptive decision making, consumers in such a simple decision situation are more likely to use attribute-based strategies, taking into account comparative characteristics of the alternatives when making a final decision. This process may account for the occurrence of the attraction phenomenon.
Recently, Wernerfelt (1995) suggested that the attraction phenomenon can be seen as an outcome of consumers' rational inferences about utilities from market offerings. In fact, some of the other previously proposed theoretical explanations for the attraction effect, such as the consumer's reliance on relative attribute comparisons (Huber and Puto 1983) and tradeoff contrast (Simonson and Tversky 1992), also imply the role of attribute comprehension and inferences in producing the attraction phenomenon. Thus, we put forward the following hypothesis:
H1: The attraction effect results from consumers' evaluations, comparisons, or any other computations on attribute values - namely, the attribute-based process.
Search for Reasons in Decision Making
A number of researchers have proposed a view that in certain situations decision makers tend to make their choices on the basis of available reasons and justifications (Simonson 1989). An example of the reason-based choice can be seen in Slovic's (1975) study. He reported that when faced with a choice between two equally valued alternatives, decision makers tend to prefer the one that is better on the more important attribute. According to the author, such an approach to problem solving is likely to occur because the chosen option can be easily justified for oneself and others as being the best decision.
The current theorizing on reason-based decision making has advanced the idea that the decision process involves the search for a dominance structure - a cognitive representation in which one alternative can be perceived as dominant over the others (Montgomery 1983, 1989). The search for a dominance structure in decision making is an appealing process because it provides a ground for justifying the final choice (Montgomery 1983). It permits decision making be based on clear reasons without a reliance on relative weights, attribute tradeoffs, or other effortful computations, thus easing the demands on the decision maker's limited information processing capacity. In this sense, the desire to search for a dominant structure is compatible with the characterization of consumers as limited information processors (Montgomery 1989). As Montgomery further argues, decision makers have the tendency to protect their choices from the competing alternatives, thus the construction of a dominance structure is a desirable goal for decision makers.
In the context of brand decisions, the dominance-search model suggests that brand preference and choice is a function of the dominance structure in the choice set. In most cases, however, a dominance structure may not exist in its pure sense. Therefore, the decision maker needs to restructure the given information in such a way that a dominance structure can be obtained. Montgomery (1989) proposed a number of operations that can be used to achieve this. For example, the decision maker may de-emphasize a given disadvantage of the promising alternative and/or bolster the disadvantages of non-promising alternatives. According to Montgomery, these operations may result in a fully developed dominance structure in which the preferred alternative is better than the other alternatives on all of the attributes under consideration, or the disadvantages of the preferred choice are completely eliminated, neutralized or counterbalanced.
Unfortunately, little empirical evidence has been reported thus far to support the depiction of decision making as a search for dominance. In studying the effect of the dominance relationship on adaptive decision making, Klein and Yadav (1989) have recently shown that increasing the number of dominated alternatives significantly improve choice accuracy and reduced choice effort. These findings clearly indicate that decision bahavior is affected by the dominance structure encountered (or constructed) in the decision process.
In addition to the dominance structure, Simonson (1989) indicated that decision behavior may also be accounted for by the compromise structure - a cognitive representation in which one alternative is seen as a compromise choice (or a middle option) in terms of its attribute values between the existing alternatives. Drawing from the notion of loss aversion (Tversky and Kahneman 1991), Simonson and Tversky (1992) argued that decision makers tend to avoid the selection of extreme alternatives. The construction of a compromise structure is likely to make decision makers feel safe or less risky about their decisions (Huber and Puto 1983; Simonson 1989). It implies that a compromise relationship found among decision alternatives may serve as another good basis for justifying a choice. As in the case of searching for a dominance structure, this decision process similarly involves less considerations of attribute values.
In short, decision making has been described as a process of searching for contextual factors that may provide reasons or justifications for the final choice. Two of these factors are the dominance and compromise relationships among the alternatives in the choice set. The view of reason-based decision making implies that little information integration is required for making a decision. Rather, the main decision task is to establish a dominance and/or a compromise relationship among the alternatives, which in turn can lead to the final choice.
The perspective of reason-based decision making provides an alternative explanation for the occurrence of the attraction effect. As mentioned earlier, decision contexts used in the past studies of the attraction effect incorporated dominance (or near-dominance) and/or compromise relationships. According to the reason-based theory of decision making, these characteristics of the decision context are likely to be used as bases for justification when consumers make their decisions. Therefore, we postulate our second hypothesis as follows:
H2: The attraction effect results from consumers' reliance on a dominance and/or a compromise relationship contained in the choice task - namely, the reason-based process.
It should be noted that Simonson (1989) attempted to examine the reason-based process in explaining the attraction effect. In one of his experiments, Simonson collected think aloud protocols from subjects. However, because the attraction effect was measured only at the aggregate level, the obtained cognitive thoughts could not be used to explain the individual subject's decision process leading to the attraction effect. This lack of an individual level analysis of the decision process renders the author's explanation for the attraction effect less conclusive.
Subjects and Stimuli
Subjects for this preliminary study were twenty university students. Each subject was paid $5 for their participation. Three product categories were used: cars, calculators, and orange juice. They were employed as product stimuli for their relevance to the subject population as well as their representation of different risk and consumer involvement levels (Mishra, Umesh, and Stem 1993).
The attributes used in this study, together with their levels, are similar to those used in previous studies: city mileage per gallon and ride quality rating for cars (Ratneshwar, Shocker, and Stewart 1987), number of functions and probability of repair in the first two years for calculators (Simonson 1989), and price and quality ratings for orange juice (Huber, Payne, and Puto 1982).
Study Design and Procedure
The study was conducted on a within-subject basis to compare the preference pattern found in the core set consisting of two non-dominating brands A and B to that found in the three-alternative set including the original two brands (labelled as X and Y, respectively, to reduce the demand effects) and a new brand (Z). We manipulated the new entrant to be relatively inferior to brand Y. Structurally, the addition of such a new brand will objectively establish not only a dominance (or near-dominance) relationship but also a compromise relationship as well for a given choice task. These characteristics enable us to test the proposed hypotheses (i.e., effects of the attribute-based and reason-based processes). Table 1 gives the details of the choice tasks used in this study.
THE CHOICE TASKS USED IN THE STUDY
As a within-subject design, this experiment requires choice tasks to be completed by each respondent. Consequently, we can estimate the attraction effect for each respondent. Individual-level estimates of the attraction effect are essential to the understanding of the underlying decision processes.
Subjects were first presented with the core choice set (X and Y of Table 1), and asked to choose the brand they would buy and provide preference ratings on a 0-100 constant sum scale (Mishra, Umesh, and Stem 1993). Next, they repeated this set of tasks for the other two product categories. After taking a ten minute break, they were given the three-alternative set consisting of the two original brands and a new brand (Z) in each product category, and then asked to respond to similar questions to those used in the core choice tasks. Finally, participants were required to complete a set of demographic questions.
Following previous practices, the options were presented in an alternative (row) by attribute (column) matrix format. The order in which the new brand and product category were presented was counterbalanced across subjects.
To provide direct evidence of the attraction effect, retrospective verbal protocols (Ericsson and Simon 1984) were collected from respondents. Subjects were asked to describe how they arrived at the decision immediately after the completion of each choice task involving the three-alternative set. They were instructed to state everything that went on in their minds while they made the choice, as in Ratneshwar, Shocker, and Stewart (1987).
Attribute-based and Reason-based Processes. Subjects' protocols were separated into individual thoughts, and then coded by two independent judges into categories for attribute-oriented thoughts and reason-oriented thoughts. Thoughts relating to the search for the dominance and/or compromise structure provided in the choice tasks were construed as reason-based, whereas thoughts relating to evaluations, comparisons, or any other computations on attribute values, which do not reflect subjects' focus on the choice set structure, as attribute-based. The number of thoughts was used as an indicator of the extensiveness of the decision process. This measure is similar to that used by Sujan (1985) in his study of category-based and piecemeal processes underlying brand evaluations. Table 2 presents the details of the coding procedure for the two-way decision process.
The criteria for determining the reason-based process were largely based on the operations proposed by Montgomery (1989) for establishing a dominance relationship. Given the purpose of this research, the operations were modified in such a way that the criteria covered dominance as well as compromise relationships objectively contained in the decision context. The criteria used in delineating the attribute-based thoughts come from the elementary information processes proposed by Johnson and Payne (1985). A count of the total number of elementary information processes (EIPs) used for decision making provides a measure of the effort associated with the use of a certain decision strategy (Payne, Bettman, and Johnson 1988; 1992). This study is particularly concerned with consumers' evaluations, comparisons, or any other types of computations on attribute values that would indicate a construction of the overall worth of a specific alternative from the pieces of attribute information provided.
A caution should be taken in interpreting the coding scheme. For example, de-emphasizing and bolstering, criteria for reason-oriented thoughts, appear to be similar to comparing, a criterion for attribute-oriented thoughts. The distinction between them lies in that de-emphasizing and bolstering primarily focus on the comparisons of brands between which there is a dominance relationship. Any thoughts related to comparisons between the dominating and the dominated brands (such as, brands Y and Z in Table 2) are deemed to be indications of searching for the dominance relationship contained in the choice task. If comparisons occur between non-dominating brands (such as, brands X and Y), the related thoughts were classified as attribute-based as they do not involve the search for the dominance structure contained in the choice task.
CODING SCHEME FOR RESPONSES
Attraction Effect. The attraction effect measure used in this study was adopted from Mishra, Umesh, and Stem (1993). It computes the difference between the observed preference share of the target brand and the estimated share of the target derived from the principle of proportionality (Luce 1959). A positive difference found upon addition of the new brand into the choice set signifies the occurrence of the attraction effect. Here, the attraction effect is defined as the net change in market share of the target brand after an adjustment is made for the expected proportional loss based on the constant ratio model. A better understanding of this definition can be seen from the following example given by Mishra, Umesh, and Stem (1993). Consider a core set share of 60 for the target brand Y and 40 for the competitor X. If the decoy Z captures a share of 20, the expected shares of brands Y and X will decline proportionately to 48 (=60-20 x .6) and 32 (=40-20 x .4), respectively. The attraction effect will exist if the observed share of brand Y (target) is greater than 48.
Distinguishing Reason-based Process from Attribute-based Process
The research hypotheses in this study focus on a two-way decision process in producing the attraction effect: (1) the search for a dominance and/or a compromise relationship contained in the choice task (reason-based processing), and (2) the inferences of the values of alternatives from the available attribute information (attribute-based processing).
Subjects' responses were separated into individual thoughts and coded by two judges. The judges were blind to the hypotheses. The interjudge agreement was 87 percent. Disagreements were resolved through discussion, so that all responses were coded. A sample of the attribute-oriented thoughts and the reason-oriented thoughts coded by the two judges is found in table 3, for the choice task involving cars.
Note that the reason-oriented thoughts differ from the attribute-oriented thoughts in the way that they explicitly reflect subjects' consideration of the dominance and/or compromise relationship provided in the choice task. The sample responses suggest that the target brand (Y) can be benefited not only from the choice set structure such as a dominance relationship but from inferences of attribute values as well.
A simple correlation analysis was performed to examine the relative impact of the two-way decision process on the attraction effect. Because the results regarding the amount and pattern of the attribute-oriented and reason-oriented thoughts obtained for the other two product categories were similar to those for cars, the analysis was done only for the choice task involving cars. We were interested in how the attraction effect is related to both the attribute-based process and the reason-based process. Noting that decision behavior is likely to consist of multiple systems that interact in various ways (Payne 1982), we expected that the two processes explaining the occurrence of the attraction effect are more complementary than competitive. A significant relationship was found between the attraction effect and the reason-based process (r=.48; p<.01). But, the relationship between the attraction effect and the attribute-based process was not significant. Consequently, the empirical results reject the hypothesis 1, and provide support for the hypothesis 2.
DISCUSSION AND CONCLUSION
The substantive point of this study is that a dominance and/or a compromise structure contained in a choice task can lead to the attraction effect through two conceptually different decision processes, namely, attribute-based processing and reason-based processing. The attribute-based processing reflects the theory of adaptive decision behavior. This decision process suggests that the attraction effect can be seen as a manifestation of consumers' using other alternatives to infer the values of a specific option in the choice set. Such a conceptualization of the rationales underlying the attraction effect is consistent with the theoretical arguments made by Wernerfelt (1995). Using a number of examples, Wernerfelt describes the attraction phenomenon as an outcome of consumers' rational inferences about utilities from market offerings.
The reason-based processing reflects the theory of the search for reasons in choice. This mechanism suggests that consumers can choose on the basis of the relationships of the alternatives in the choice set rather than the comprehension of the attribute values of alternatives. The attraction phenomenon constructed through the dominance/compromise search can be seen as a manifestation of deviations from rationality in choice. In fact, a number of researchers (e.g., Simonson and Tversky 1993; Pan, O'Curry, and Pitts 1995) have indicated that certain normative assumptions of consumer choice models such as value maximization are inadequate for understanding of the attraction effect and need to be relaxed in order to account for context effects.
While the attribute-based process and the reason-based process represent different construction of the attraction effect, they are more complementary than competitive. Under certain situations, both decision processes may produce the attraction effect simultaneously. There are also situations in which either decision process may operate. The relative impact of the two decision processes on the attraction effect are likely to be affected by a number of individual and task factors. Two noteworthy variables are accountability and information ambiguity.
The retrospective protocols obtained in the present study provide preliminary support for the reason-based process but not for the attribute-based process. The findings could be due to the ambiguity in attribute information. Given the ambiguity of choice alternatives, subjects may feel uncertain about the choices based on attribute values. As a result, they tend to rely on the available dominance structure when making choice decisions. Further research may examine this speculation by manipulating information ambiguity. In addition, the influence of other individual and task variables such as accountability on the attraction process also deserves researchers' attention.
In practice, the focus on the roles of attribute values and choice set structure in consumer decision making would have important implications for competitive strategies. If our conceptualization about context effects in choice is sound, it may suggest that brand competition is both a race to meet customer needs at attribute values and a battle to add distinctive cues such as a dominance relationship over the structure of customer's choice set (Taylor and Fiske 1978). Hence, marketers should devote resources not just to satisfy consumers better than competitors but to create value for consumers by shaping the context of preferences and thus competition (Carpenter, Glazer, and Nakamoto 1994).
Of course, there are several limitations to this preliminary investigation. First, the small size of the sample employed in this study makes our findings tentative. A large sample is certainly needed for further examination of the two-way decision process leading to the attraction effect. Second, the method of data analysis is very simple. Strong conclusions can only be made when other advanced methods are applied here. Yet in spite of these obvious weaknesses, this paper provides encouraging insights into the reasons for the observed attraction effect, and should be followed by more rigorous investigations.
Abelson, R.P., and Levi, A. (1985), "Decision Making and Decision Theory," In the Handbook of Social Psychology, eds., G. Lindzey and E. Aronson, Vol. 1, New York: Random House, 231-309.
Bettman, James R. and Michel A. Zins (1977), "Constructive Processes in Consumer Choice," Journal of Consumer Research, 4(September), 75-85.
Boush, David M. and Barbara Loken (1991), "A Process-Tracing Study of Brand Extension Evaluation," Journal of Marketing Research, Vol. XXVIII (February), 16-28.
Carpenter, G.S., R. Glazer, and K. Nakamoto (1994), "Meaningful Brands From Meaningless Differentiation: The Dependence on Irrelevant Attributes," Journal of Marketing Research, Vol. XXXI (August), 339-350.
Ericsson, K. Anders and Herbert A. Simon (1980), "Verbal Reports as Data," Psychological Review, 87 (3), 215-251.
Gertzen, H. (1992), "Component Processes of Phased Decision Strategies," Acta Psychologica, 80, 229-246.
Huber, Joel, John W. Payne, and Christopher Puto (1982), "Adding Asymmetrically Dominated Alternatives: Violations of Regularity and the Similarity Hypothesis," Journal of Consumer Research, 9, 90-98.
Huber, Joel and Christopher Puto (1983), "Market Boundaries and Product Choice: Illustrating Attraction and Substitution Effects," Journal of Consumer Research, 10, 31-44.
Johnson, Michael D. and Christopher Puto (1987), "A Review of Consumer Judgment and Choice," in Review of Marketing, ed., Michael J. Houston (Chicago, IL: American Marketing Association), 236-292.
Klein, N.M. and M.S. Yadav (1989), "Context Effects on Efforts and Accuracy in Choice: An Enquiry into Adaptive Decision Making," Journal of Consumer Research, 15(March), 411-421.
Lehmann, Donald R. and Yigang Pan (1994), "Context Effects, New Brand Entry, and Consideration Sets," Journal of Marketing Research, Vol. XXXI, August, 364-374.
Luce, R. Duncan (1959), Individual Choice Behavior, New York: John Wiley.
Lussier, D. A. and R. W. Olshavsky (1979), "Task Complexity and Contingent Processing in Brand Choice," Journal of Consumer Research,, 6 (September), 154-165.
Mishra, Sanjay, U.N. Umesh, and Donald E. Stem, Jr. (1993), "Antecedents of the Attraction Effect: An Information-Processing Approach," Journal of Marketing Research, Vol. XXX (August), 331-349.
Montgomery, Henry (1983), "Decision Rules and the Search for a Dominance Structure: Towards a Process Model of Decision Making," in Analyzing and Aiding Decision Processes, eds., P. Humphreys, O. Svenson, and A. Vari, North-Holland and Hungarian Academic Press, Amsterdam/Budapest, 343-369.
Montgomery, Henry (1989), "From Cognition to Action: the Search for Dominance in Decision Making," in Process and Structure in Human Decision Making, eds., H. Montgomery and O. Svenson, John Wiley & Sons Ltd. 23-49.
Olshavsky, R. W. (1979), "Task Complexity and Contingent Processing in Decision Making: A Replication and Extension," Organizational Behavior and Human Performance, 24, 300-316.
Pan, Yigang and Donald R. Lehmann (1993), "The Influence of New Brand Entry on Subjective Brand Judgments," Journal of Consumer Research, 20 (June), 76-86.
Pan, Yigang, S. O'Curry, and R. Pitts (1995), "The Attraction Effect and Political Choice in Two Elections," Journal of Consumer Psychology, 4(1), 85-101.
Payne, John W. (1976), "Task Complexity and Contingent Processing in Decision Making: An Information Search and Protocol Analysis," Organizational Behavior and Human Performance, 16, 366-387.
Payne, John W. (1982), "Contingent Decision Behavior," Psychological Bulletin, Vol.92, No.2, 382-402.
Payne, John W., James R. Bettman, and Eric J. Johnson (1988), "Adaptive Strategy Selection in Decision Making," Journal of Experimental Psychology: Human Learning, Memory and Cognition.
Payne, John W., James R. Bettman, and Eric J. Johnson (1992), "Behavioral Decision Research: A Constructive Processing Perspective," Annual Review of Psychology, 43
Ratneshwar, S., A.D. Shocker, and D.W. Stewart (1987), "Toward Understanding the Attraction Effect: The Implications of Product Stimulus Meaningfulness and Familiarity," Journal of Consumer Research, 13 (March), 520-533.
Simonson, Itamar (1989), "Choice Based on Reasons: The Case of Attraction and Compromise Effects," Journal of Consumer Research, 16 (September), 158-174.
Simonson, Itamar and Amos Tversky (1992), "Choice in Context: Tradeoff Contrast and Extremeness Aversion," Journal of Marketing Research, Vol. XXIX (August), 281-295.
Slovic, Paul (1975), "Choice Between Equally-Valued Alternatives," Journal of Experimental Psychology: Human Perception and Performance, 1 (3), 280-287.
Sujan, Mita (1985), "Consumer Knowledge: Effects on Evaluation Strategies Mediating Consumer Judgments," Journal of Consumer Research, 12(June), 31-46.
Taylor, S.E. and S.T. Fiske (1978), "Salience, Attention, and Attributions: Top of the Head Phenomena," in Advances in Experimental Social Psychology, Vol. 11, ed. L. Berkowitz, New York: Academic Press.
Tversky, Amos and Daniel Kahneman (1991), "Loss Aversion in Riskless Choice: A Reference Dependent Model," Quarterly Journal of Economics, 106(November), 1040-1061.
Wernerfelt, Birger (1995), "A Rational Reconstruction of the Compromise Effect: Using Market Data to Infer Utilities," Journal of Consumer Research, 21(March), 627-633.