Asymmetric Choice Patterns ACRoss Higher-Quality and Lower-Quality Brands

Subimal Chatterjee, State University of New York at Stony Brook
Timothy B. Heath, University of Pittsburgh
ABSTRACT - Existing research suggests that adding a dominated brand to a choice set benefits higher-quality brands more than lower-quality brands. For example, although adding a dominated brand to a choice set generally increases the choice share of the dominating or target brand (attraction effect; Huber, Payne, and Puto 1982), the added brand increases choice shares of higher-quality targets more than choice shares of lower-quality targets (Heath and Chatterjee 1993). We report an experiment that investigates choices and relative preferences between a higher quality (higher priced) brand and a lower quality (lower priced) brand. The results indicate that (1) subjects choosing the higher-quality brand prefer their brand more than subjects choosing the lower-quality brand prefer theirs, (2) adding a dominated brand increases the choice share of the higher-quality brand but not the lower-quality brand, and (3) adding a brand dominated by the lower-quality but not higher-quality brand can increase the higher-quality brand's choice share if that brand is obviously superior to the added brand.
[ to cite ]:
Subimal Chatterjee and Timothy B. Heath (1995) ,"Asymmetric Choice Patterns ACRoss Higher-Quality and Lower-Quality Brands", in NA - Advances in Consumer Research Volume 22, eds. Frank R. Kardes and Mita Sujan, Provo, UT : Association for Consumer Research, Pages: 328-332.

Advances in Consumer Research Volume 22, 1995      Pages 328-332

ASYMMETRIC CHOICE PATTERNS ACROSS HIGHER-QUALITY AND LOWER-QUALITY BRANDS

Subimal Chatterjee, State University of New York at Stony Brook

Timothy B. Heath, University of Pittsburgh

ABSTRACT -

Existing research suggests that adding a dominated brand to a choice set benefits higher-quality brands more than lower-quality brands. For example, although adding a dominated brand to a choice set generally increases the choice share of the dominating or target brand (attraction effect; Huber, Payne, and Puto 1982), the added brand increases choice shares of higher-quality targets more than choice shares of lower-quality targets (Heath and Chatterjee 1993). We report an experiment that investigates choices and relative preferences between a higher quality (higher priced) brand and a lower quality (lower priced) brand. The results indicate that (1) subjects choosing the higher-quality brand prefer their brand more than subjects choosing the lower-quality brand prefer theirs, (2) adding a dominated brand increases the choice share of the higher-quality brand but not the lower-quality brand, and (3) adding a brand dominated by the lower-quality but not higher-quality brand can increase the higher-quality brand's choice share if that brand is obviously superior to the added brand.

Mounting evidence suggests that market interventions benefit higher-quality brands more than lower-quality brands. In real-world markets, price discounts move consumers from lower-quality to higher-quality brands more than from higher-quality to lower-quality brands (Allenby and Rossi 1991; Blattberg and Wisnewski 1989; Kamakura and Russell 1989). In experimental choice settings, brands of moderate price and quality (compromise brands) attract more consumers from lower quality (lower priced) brands than from higher quality (higher priced) brands (Simonson and Tversky 1992). Similarly, asymmetrically dominated brands increase choice probabilities of higher-quality brands more than choice probabilities of lower-quality brands (Heath and Chatterjee 1993). Consider the scenarios below where Scenarios 2 and 3 include a brand (C) dominated by one original brand but not by the other. Adding such brands generally increases the choice share of the dominating (target) brand, an effect dubbed the attraction effect (Huber, Payne, and Puto 1982). However, recent meta-analytic evidence shows that attraction effects are common with higher-quality targets (Scenario 2), but rare with lower-quality targets (Scenario 3; see Heath and Chatterjee 1993).

TABLE

Using asymmetric dominance within experimental choice tasks, the current study extends evidence of differential preferences for higher-quality and lower-quality brands in three ways. First, it replicates evidence suggesting differential sensitivity to asymmetric dominance. Second, it tests the possibility that under certain circumstances, adding a brand dominated by the lower-quality but not higher-quality brand can increase the higher-quality brand's share. Third, it tests directly whether subjects choosing the higher-quality brand hold stronger relative preferences for their brand than subjects choosing the lower-quality brand hold for their brand. Relative preferences are operationalized as differences in attitudes toward higher-quality and lower-quality brands.

THEORY AND RESEARCH

Figure 1 illustrates the attraction effect. In Figure 1, Brands A and B are competitors and the entrants (S1, S2, and S3) are added one at a time. Neither Brand A nor Brand B dominates the other since each is superior on one attribute. However, Brand S1 is dominated by A but not by B (i.e., Brand S1 is asymmetrically dominated). Attraction effects due to such brands have been reported across numerous product classes and studies (Huber, Payne, and Puto 1982; Huber and Puto 1983; Lehmann and Pan 1994; Mishra, Umesh, and Stem 1993; Pan and Lehmann 1993; Ratneshwar, Shocker and Stewart 1987; Simonson 1989; Wedell 1991). However, attraction effects vary across different types of target brands. Heath and Chatterjee (1993) reanalyzed data from thirty experimental choice sets and found that attraction to higher-quality brands occurred in nine of fourteen cases, but attraction to lower-quality brands occurred in only one of sixteen cases. The pattern parallels findings from real-world markets where it is easier to move consumers to higher-quality brands than to lower-quality brands. In one study, price discounts of frequently purchased grocery items attracted consumers from lower-quality to higher-quality brands in eight out of twelve cases, but attracted consumers from higher-quality to lower-quality brands in only three out of twenty-seven cases (Blattberg and Wisnewski 1989).

Researchers have proposed three explanations that may account for the observed asymmetric responses to price discounts. First discounting higher-quality brands results in favorable substitution and income effects, whereas discounting lower-quality brands results in favorable substitution effects but an unfavorable income effect (Allenby and Rossi 1991). The unfavorable income effect occurs because discounting the lower-quality brand increases discretionary income that enables consumers to buy higher-quality (higher priced) brands. Second, asymmetric price competition may be due to asymmetries in a property called loss aversion, the tendency for losses to be more unpleasant than gains of equal economic value are pleasant (Kahneman and Tversky 1979; Tversky and Kahneman 1991). Consumers may be more loss-averse for quality than price. Hardie, Johnson, and Fader (1993), for example, report loss-aversion coefficients that are greater for quality than price. Third, asymmetries in price competition may arise if consumers of higher-quality brands are more committed to quality than consumers of lower-quality brands are committed to non-quality attributes such as price (Blattberg and Wisnewski 1989).

In experimental choice settings, subjects do not actually buy brands or get an opportunity to trade one brand for the other. Thus, income or substitution effects are less likely to occur in experimental markets. Differential loss aversion and/or differential commitment to quality and price, on the other hand, can account for the asymmetries in choice patterns between higher and lower quality brands in experimental tasks (see Simonson and Tversky 1992)

FIGURE 1

BRAND CONFIGURATIONS

Study Overview and Hypotheses

We tested three hypotheses investigating choices and relative preferences between a higher-quality and lower-quality brand in one experiment in which student subjects chose between two fictitious brands of beer, A and B (Figure 1). Brand A was the higher-quality, higher-priced brand (75-quality, $4.95 per six-pack). Brand B was the lower-quality, lower-priced brand (65-quality, $4.25 per six-pack). The control condition consisted of Brands A and B, whereas an out-of-stock brand was added in each of the three experimental conditions (Brands S1, S2, and S3).

Based on prior research we expected that an asymmetrically dominated brand designed to make Brand A look more attractive (S1) would increase A's choice share and reduce B's, whereas an asymmetrically dominated entrant targeting Brand B (S2) would have no such effects.

H1: Introducing an asymmetrically dominated brand targeting the higher-quality brand (Brand S1 targeting Brand A) will increase the choice share of the higher-quality brand. Introducing an asymmetrically dominated brand targeting the lower-quality brand (Brand S2 targeting Brand B) will not increase the choice share of the lower-quality brand.

We extend prior research by testing the possibility that under certain circumstances, adding a brand dominated by the lower-quality but not the higher-quality brand can increase the higher-quality brand's share. Consider Entrant S3 that is dominated by Brand B and not by Brand A (Figure 1). In addition to making Band B appear attractive through dominance, it may make Brand A appear attractive by being obviously inferior to A (Brand A offers 10 units more quality than S3 at a similar price). However, if people choosing higher-quality brands are less likely to switch than people choosing lower-quality brands, then S3 should increase Brand A's choice share by moving people off Brand B, even though A does not dominate it. [Past research has shown that people are less likely to prefer the cheapest alternative when they choose from a set of three brands (Simonson, Nowlis and Lemon 1993). This does not threaten our test since Brand S3 is kept out of stock and subjects always choose between two brands: A (highest price/quality) and B (lowest price/quality).]

H2: When asymmetrically dominated brands are obviously inferior to both lower-quality and higher-quality brands (e.g., S3), their effect will be to increase choice shares of the higher-quality brands.

Past research and Hypotheses 1 and 2 suggest that consumers of lower-quality brands are more likely to switch to higher-quality brands than consumers of higher-quality brands are likely to switch to lower-quality brands. The implication is that consumers preferring higher-quality brands prefer them more strongly over lower-quality brands, than consumers of lower-quality brands prefer their brands over higher-quality brands. Our third hypothesis is that the relative preferences of consumers of higher-quality brands are stronger than those of lower-quality brands, where relative preferences are operationalized as differences in attitudes toward higher-quality and lower-quality brands.

H3: Subjects choosing the higher-quality brand (A) will display stronger relative preferences than subjects choosing the lower-quality brand (B).

EXPERIMENT

Subjects and Procedure

Two-hundred-and-five business students from three large eastern universities served as subjects in a one-way between-subjects experiment consisting of one control group and three experimental groups. Subjects were told to imagine that they wished to buy a six-pack of beer. The control group was asked to choose from a choice set consisting of two brands: Brand A was priced at $4.95 and had a quality rating of 75 (1-100 scale). Brand B was priced at $4.25 and had a quality rating of 65. The three experimental groups chose between the same two brands, although an out-of-stock brand was included as well.

The brand configurations are shown in Figure 1. To assess differential effects on higher-quality and lower-quality brands, Brands S1 and S2 were configured to be equidistant from their respective target brands (i.e., S1's proportional deviation from A's quality matched S2's proportional deviation from B's price). Thus, S1 had A's price but 4.0% lower quality, whereas S2 had B's quality but 4.3% higher price. Brand S3 was configured such that it was dominated by Brand B and not by Brand A, and yet made Brand A look attractive. Thus, S3's quality was made equal to that of the lower-quality brand (B), and its price only slightly lower than A's (by 2%). Attraction effects to Brand A were expected with entrants S1 and S3. No attraction effects were expected to Brand B. After making their choices, subjects in all conditions rated each brand on a 1-20 like / dislike scale.

Analyses and Results

The data are analyzed in two stages. First, between-group differences in choice were assessed with Fisher's Exact Test of Independence. A separate 2 X 2 contingency table was formed for each pairwise comparison between control and experimental groups. Second, differences in relative preferences between subjects choosing Brand A and Brand B were assessed using ANOVAs.

Hypothesis 1 predicted that Brand S1 would increase Brand A's share, whereas Brand S2 would not increase Brand B's share. These effects were assessed by comparing target brands' shares between control and experimental conditions. Hypothesis 1 was supported. Brand S1 significantly increased Brand A's share from 49.02% to 80.85% (p <.001). Brand S2 did not significantly affect B's share (BControl=50.98% vs. BExperimental =42.59%, p =.855).

Hypothesis 2 predicted that Brand S3 would increase A's choice share rather than B's share, even though B, not A, dominated it. The hypothesis was supported. Brand S3 increased A's share from 49.02% to 66.04% (p =.059).

Hypothesis 3 predicted that relative preferences would be stronger among subjects choosing the higher-quality brand (A) than subjects choosing the lower-quality brand (B). Relative preference was operationalized by taking the difference between the attitude towards the chosen brand and foregone brand.

Relative preference

= LikingBrand A - LikingBrand B ,

for subjects choosing Brand A,

= LikingBrand B - LikingBrand A ,

for subjects choosing Brand B.

[To ensure uncontaminated choice processes, attitudes were measured only after choices were made. Although relative preferences may have been inflated to justify choices, such justification need not threaten our test since there is no a priori reason for expecting justification to vary across consumers of lower-quality and higher-quality brands.]

Relative preferences were tested separately in the control condition and experimental conditions. In the control condition, relative preferences were subjected to a one-way fixed-effects ANOVA: Choice (A vs. B). In the experimental conditions, relative preferences were subjected to a two-way fixed-effects ANOVA : Choice (A vs. B) and Added Brand (S1 vs. S2 vs. S3). [The control condition provides a pure test for differential preference strengths since it is not contaminated by perceptual biases arising from irrelevant alternatives.]

Hypothesis 3 was supported in both conditions. In the control condition, although there was a 50:50 split in choices, subjects choosing Brand A showed significantly stronger relative preferences than subjects choosing Brand B (MBrand B=1.42, MBrand A=4.13, F (1,46)=6.38, p <.05). In the three experimental conditions, Brand A was favored by a 2:1 margin overall (67.53% for Brand A vs. 32.47% for Brand B) and subjects choosing Brand A showed significantly stronger relative preferences than subjects choosing Brand B (MBrand B=4.62, MBrand A=6.21, F (1,144)=8.60, p <.01). However, the effect of the brand chosen was moderated by brand added (Interaction F (2,144)=10.93, p <.01; see Table 1). When Brands S1 or S3 were added to the choice set, the pattern matched the pattern in the control condition. Relative preferences were significantly stronger among subjects choosing Brand A than among subjects choosing Brand B. However, when Brand S2 was added to the choice set, relative preferences were significantly stronger among subjects choosing Brand B than among subjects choosing Brand A. Since Brand S2 was clearly dominated by Brand B, subjects may have focused on the dominance relationship to justify a higher rating for Brand B. [Although S3 was dominated by Brand B, it was obviously inferior to both Brands A and B. Hence S3 provides less reasons for justifying a higher rating and Brand B alone.]

GENERAL DISCUSSION

Summary

Recent research suggests that asymmetrically dominated brands are more effective in improving choice shares of higher-quality than lower-quality target brands (Heath and Chatterjee 1993). The current study replicated and extended such asymmetries in three ways. We found that (1) asymmetrically dominated brands increased the share of the high-quality brand, but not the lower-quality brand, (2) adding a brand dominated by the lower-quality but not the higher-quality brand can increase the choice probability of the higher-quality brand if that brand is obviously superior to the new brand, and (3) relative preferences among subjects choosing the higher-quality brand were significantly greater than among subjects choosing the lower-quality brand. Thus, we replicated and extended prior asymmetries in choice patterns, and suggest and support an explanation for the asymmetries.

TABLE 1

RELATIVE PREFERENCE STRENGTHS ACROSS CONSUMERS OF HIGHER-QUALITY (A) AND LOWER-QUALITY (B) BRANDS

The current study suffers from three shortcomings that must be addressed. First, the experimental data were limited to a simple three-brand choice set where each brand was described on price and quality, and the added brand differed from targets only on one attribute. Such limitations are only partially mitigated by the fact that asymmetries in the attraction effect have been reported in a meta-analysis that used a wide variety of products and entrants that varied on two attributes compared to the targets (Heath and Chatterjee 1993). Future research needs to address scenarios that more closely approximate multiple-brand, multiple-attribute markets of the real world.

Second, difference measures were used to operationalize relative preferences. Difference scores may lead to a variety of potential problems including low reliability, false discriminant validity, spurious correlation, and variance restriction (see Peter, Churchill, and Brown 1993). One way of getting around these problems is to operationalize relative preferences more directly such that calculation of difference scores is unnecessary (Johns 1981). For example, a direct comparison measure of relative preferences may be operationalized in a question like "Overall how much do you prefer the brand chosen over the brand that was not chosen" that is scored on a scale anchored from "about the same" to "very much more than the brand not chosen." [However, direct comparison measures also suffer from measurement problems. For example, Dhar and Simonson (1992) found that in a judgement task involving a comparison between two options, focusing attention on one alternative (in this case the chosen option) tends to enhance its attractiveness.]

Third, we did not test the potential explanation of differential loss-aversion across higher-quality and lower-quality brands. Future research might do so with measures such as "Overall how much does an x% decrease in quality make you more unhappy than an x% increase in price," that is scored on a scale anchored from "about the same" to "very much more than an increase in price."

REFERENCES

Allenby, Greg M. and Peter E. Rossi (1991), "Quality Perceptions and Asymmetric Switching Between Brands," Marketing Science, 10 (Summer), 185-204.

Blattberg, Robert C. and Kenneth J. Wisnewski (1989), "Price Induced Patterns of Competition," Marketing Science, 8 (Fall), 291-309.

Dhar, Ravi and Itamar Simonson (1992), "The Effect of Focus of Comparison on Consumer Preferences, " Journal of Marketing Research, 29 (November), 430-440.

Hardie, Bruce G. S,. Eric J. Johnson, and Peter S. Fader (1993), "Modeling Loss Aversion and Reference Dependence Effects on Brand Choice," Marketing Science, 12 (Fall), 378-394..

Heath, Timothy B. and Subimal Chatterjee (1993), "Attraction and Compromise Effects in Choice: Moderating Influences and Differential Loss Aversion to Quality and Nonquality Attributes," in Advances in Consumer Research, Vol. 20, Association for Consumer Research, 475.

Huber, Joel, John W. Payne, and Christopher Puto (1982), "Adding Asymmetrically Dominated Alternative: Violations of Regularity and the Similarity Hypothesis," Journal of Consumer Research, 9 (June), 90-98.

Huber, Joel and Christopher Puto (1983), "Market Boundaries and Product Choice: Illustrating Attraction and Substitution Effects," Journal of Consumer Research, 10 (June), 31-44.

Kahneman, Daniel and Amos Tversky (1979), "Prospect Theory: An Analysis of Decision Making Under Risk," Economterica, 47 (March), 263-291.

Lehmann, Donald R. and Yigang Pan (1994), "Context Effects, New Brand Entry, and Consideration Sets," Journal of Marketing Research, 30 (August), 331-349.

Mishra, Sanjay, U. N. Umesh, and Donald E. Stem, Jr. (1993), "Antecedents of the Attraction Effect, " Journal of Marketing Research, 30 (August), 331-349.

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.

Peter, J. Paul, Gilbert A. Churchill, Jr., and Tom J. Brown (1993), "Caution in the Use of Difference Scores in Consumer Research," Journal of Consumer Research, 19 (March), 655-662.

Johns, Gary (1981), "Difference Score Measures of Organizational Behavior Variables: A Critique, " Organizational Behavior and Human Performance, 27 (June), 443-463.

Ratneshwar, Srinivasan, Allan D. Shocker, and David 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, Nowlis, Stephen, and Katherine Lemon (1993), "The Effect of Local Consideration sets on Global Choice Between Lower price and Higher Quality, " Marketing Science, 12 (Fall), 357-377.

Tversky, Amos and Daniel Kahneman (1991), " Loss Aversion in Riskless Choice: A Reference Dependent Model," Quarterly Journal of Economics, 106 (November), 1040-1061.

Wedell, Douglas H. (1991), "Distinguishing Among Models of Contextually Induced Preference Reversals," Journal of Experimental Psychology: Learning, Memory, and Cognition, 17 (4), 767-778.

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