Available online at www.sciencedirect.com
Journal of CONSUMER PSYCHOLOGY
Journal of Consumer Psychology 20 (2010) 369 – 380
The role of idiosyncratic attribute evaluation in mass customization Sanjay Puligadda a,⁎, Rajdeep Grewal b , Arvind Rangaswamy b , Frank R. Kardes c a
b
Department of Marketing, Miami University of Ohio, Oxford, OH, 45056, USA Department of Marketing, Smeal College of Business, Pennsylvania State University, University Park, PA, 16802, USA c Department of Marketing, College of Business Administration, University of Cincinnati, Cincinnati, OH, 45221, USA Received 24 July 2008; revised 7 April 2010; accepted 9 April 2010 Available online 10 May 2010
Abstract The growing use of mass customization necessitates an understanding of consumers' evaluations of mass customization platforms. We hypothesize that consumers' objective and subjective knowledge of the customized product moderate the influence of idiosyncratically evaluated (i.e., personalizable) attributes on satisfaction with a customization platform. Consistent with our theoretical framework, results from three experiments show that offering greater variety in idiosyncratically evaluated attribute options increases consumers' satisfaction to a greater extent for: (1) novices than experts (2) consumers with more subjective knowledge, and (3) miscalibrated consumers whose subjective knowledge does not match their objective knowledge, than calibrated consumers whose subjective and objective knowledge match. © 2010 Society for Consumer Psychology. Published by Elsevier Inc. All rights reserved. Keywords: Attribute classification; Mass customization; Consumer knowledge
In today's environment, it is becoming important to market products and services to small niche segments, extending even to customization to suit individual consumers (Hart, 1995), as the contexts of computers (www.dell.com), motorcycles (www.vtx. honda.com), cars (www.scion.com), specialty chemicals (www. chemstation.com), candy (www.mymms.com), and postage stamps (www.stamps.com) clearly demonstrate. Yet, despite extensive study of the production aspects of mass customization (e.g., Jiao, Ma, & Tseng, 2003; Tu, Vondermbse, & RaguNathan, 2001), only recently have scholars begun to examine mass customization from a consumer perspective (Murthi & Sarkar, 2003; Simonson, 2005; Wind & Rangaswamy, 2001). A mass customization platform (hereafter referred to as MCP) is distinctive not because it offers the “best” option of each attribute but because it enables consumers to select the options they prefer (e.g., pick a color or design of a shoe from options of colors and designs provided). Therefore, a useful place to start an investigation of an MCP is its ability to offer consumers different
⁎ Corresponding author. Fax: +1 513 529 1290. E-mail address:
[email protected] (S. Puligadda).
options among various attributes, from which consumers select their preferred option, and thereby, customize the product according to their own preferences. In this research we investigate the effect of offering varying number of options of attributes (e.g., five versus ten color options) on consumer satisfaction with an MCP. We consider two potential influences: the extent and type of knowledge of the product that consumers possess, and the manner in which consumers evaluate the attributes. Consumer knowledge is either objective, which indicates how much consumers actually know, or subjective, which indicates how much they think they know (e.g., Alba & Hutchinson, 2000). These two types of knowledge have important differential effects on consumers' satisfaction with MCPs. In addition to the notion of standardized and personalized products (Duray & Milligan, 1999), we classify product attributes according to how idiosyncratic or “personalizable” they are for consumers. That is, we classify product attributes in our research, not on the basis of their specific characteristics but on how consumers evaluate them. We distinguish between sharedpreference and idiosyncratic-preference attributes, which we consider to be similar to the universal and variable qualities of individuals (Sherman, Chassin, Presson, & Agostinelli, 1984).
1057-7408/$ - see front matter © 2010 Society for Consumer Psychology. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.jcps.2010.04.003
370
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
Shared-preference attributes (hereafter referred to as SPAs), such as battery life in cell phones, have a widely accepted or shared evaluation scheme, whereas the evaluation schemes for idiosyncratic-preference attributes (hereafter referred to as IPAs) such as the exterior color of cell phones, are idiosyncratic. We propose that due to the unique evaluation scheme for IPAs, the satisfaction derived from increasing the number of options of such attributes is conditioned on the knowledge of the consumer. We posit that objective and subjective forms of knowledge result in distinct effects (Brucks, 1985); specifically, increasing the number of options of IPAs would have a greater influence on satisfaction with the MCP among consumers as objective knowledge decreases, or subjective knowledge increases. We further argue that these results will hold only for miscalibrated consumers (Alba & Hutchinson, 2000), who are high on one knowledge construct and low on the other knowledge construct. We report results from three experiments to support these theoretical assertions. Our research augments the existing literature by revealing (1) how various distinctive features of an MCP influence consumer satisfaction with the MCP and (2) how consumers' objective and subjective knowledge of the product category moderate the effects of MCP features on satisfaction with the MCP. We identify important influences of objective and subjective knowledge, which in turn offer managerial insights concerning segmenting consumers (e.g., experts versus novices) and designing MCPs that will optimize the variety of offerings to enhance consumer satisfaction. We also adapt some constructs from social psychology and apply them to the product domain by classifying attributes as SPA or IPA, and demonstrate the applicability of this classification in a few product categories.
Satisfaction with a mass customization platform Consumer satisfaction increases when the set of options provided are sufficiently varied, because greater variety increases the likelihood that the consumer will find what he or she wants (Huffman & Kahn, 1998; Kahn, 1998). However, too much variety can seem “monumental and frustrating” (Kahn, 1998, p.48) and cause confusion and information overload (Huffman & Kahn, 1998; Lee & Lee, 2004). Thus, it is important that the variety be “just right” to maximize consumer satisfaction with the options of attributes available in an MCP. In addition to variety, personal relevance is critical (Coulter, Price, & Feick, 2003; Petty, Cacioppo, & Schumann, 1983). If the available product options match the consumer's personal goals and values, the consumer feels highly involved, which in turn activates a motivational state that “energizes” search and shopping behaviors and cognitive actions, such as attention and comprehension (Celsi & Olson, 1988). An energized motivational state is also likely to make the consumer feel more satisfied with the options provided than a less motivated consumer would. Therefore, we conceptualize satisfaction with the customization platform as the contentment experienced by consumers on the basis of how well the options offered by a customization platform satisfy the “just right” criterion of variety (e.g., Zhang & Fitzsimons, 1999) and how personally relevant those options are. Consistent with these assertions, we measured satisfaction by asking participants to respond to three items (“The set of available options gives me sufficient variety,” “With the available options, there were enough products that I could consider buying,” and “The range of options offered is appropriate for me,”) anchored by 1 = “completely disagree” and 7 = “completely agree.” We added a fourth item to this scale (“I was satisfied with the options offered for each attribute.”) in Experiment 3.
Theoretical background Shared-preference and idiosyncratic-preference attributes Mass customization Mass customization refers to the “ability to quickly design, produce, and deliver products that meet specific customer needs at close to mass-production prices” (Tu et al., 2001, p. 203), with the objective of offering superior consumer value (Pine & Gilmore, 2000). This broad definition encompasses several types of mass customization, including assembly customization, the focus of this study. Assembly customization offers options of attributes of a product that consumers may configure to create their own co-designed or co-created product. The key benefit of assembly customization is that for select product attributes, consumers may choose the option that is most appealing and satisfying to them (MacDuffie, Sethuraman, & Fisher, 1996). Yet despite their promise, many MCPs fail to deliver substantial benefits, either to firms or to consumers (Huffman & Kahn, 1998; Wind & Rangaswamy, 2001). We reason that a consumer's satisfaction with an assembly customization may depend on the number of options of product attributes offered by the customization platform (even if the range of options for other attributes remains constant).
According to Sherman et al. (1984), universally evaluated qualities are those for which “all judges (regardless of their own position on the quality) will agree on which levels of the quality are good to have and which levels are bad” (p. 1245). For example, being brave (not being brave) is considered good (bad) universally regardless of whether the judge is brave or not. In contrast, variably evaluated qualities are those for which “different judges disagree about which end of the quality is good and which is bad. In addition, attitudes toward these different levels of the quality depend on the judge's own position on that quality” (Sherman et al., 1984, p.1245). For example, supporters of capital punishment evaluate it as good, whereas non-supporters evaluate it as bad. Thus, as per Sherman et al. (1984) classification, universal qualities are those for which preferences are shared while variable qualities are those for which preferences are idiosyncratic. Although preferences can refer to a specific local state (e.g., choosing A over B) and can be argued to be largely constructed, the current use of the term preference refers to a global, stable state or disposition that is inherent to individuals (Simonson, 2008).
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
Our attribute classification based on consumer evaluations is essential in the context of mass customization because, unlike traditional purchase transactions, mass customization provides an opportunity for a consumer to select from a number of options of each attribute (e.g., number of color options). To understand the role of changing the number of options of attributes on satisfaction, it is essential to use a classification of attributes based on how consumers evaluate them, and not the inherent nature of the attributes themselves. Consequently, we classify attributes as shared-preference (SPAs) and idiosyncratic-preference attributes (IPAs). SPAs are evaluated by consumers using a shared evaluative scheme due to a common belief system; that is, there are universally accepted conceptions about the ideal value of the various attribute options. For example, for an attribute such as the coverage area of a cell phone, the common evaluative judgment across consumers is that a wider coverage is better. For an array of options ranging from local to countrywide, the shared evaluative judgment, ceteris paribus, identifies coverage across the country as the most desirable option. Similarly, regardless of the number of options available for the speed of a central processing unit—whether 1 or 5 GHz or 1, 2, 3, 4, or 5 GHz—the most preferred option should continue to be the highest processing speed, in line with the universal evaluative judgment that higher the speed, the better the product. In contrast, the assessment of an IPA depends on idiosyncratic standards, which reflect each person's position with respect to the object (Sherman et al., 1984). For example, for the color of an automobile, the option that a consumer prefers is based on his or her personal color preference, not a shared, universal preference. Because the ideal option is idiosyncratic, these attributes are “personalizable” or reflective of each consumer's preferences, so they offer consumers the ability to render a unique self-portrayal. We propose that the difference between these two types of attributes influences consumer satisfaction with an MCP. Because of the idiosyncratic nature of IPAs, satisfaction with an MCP that results from changing the number of options of the attributes should vary with the extent of a consumer's knowledge (objective and subjective) about the product. However, we theorize that the effects of objective knowledge will be opposite to those of subjective knowledge. Attribute-based processing plays an important role in consumer decision making (e.g., Carmon & Simonson, 1998; Horsky, Nelson, & Posavac, 2004), and several different attribute classification schemes have been proposed—including attribute evaluability (Hsee, 1996), alignability (Zhang, Kardes, & Cronley, 2002; Zhang & Markman, 2001), and search versus experience attributes (Jain, Buchanan, & Maheswaran, 2000; Jain & Posavac, 2001). However, none of these taxonomies address the extent of idiosyncratic attribute evaluation, which is crucial for understanding mass customization. Furthermore, none of these taxonomies have been used to generate predictions about the antecedents or consequences of idiosyncratic versus shared attribute evaluation. Consumer knowledge Because knowledge influences the formation of preferences, a natural question to explore with respect to the role of idiosyncratic-
371
preference attributes is the extent to which consumer knowledge influences idiosyncratic preferences. We consider the two dimensions of consumer knowledge, namely, objective and subjective knowledge (for a review of their relationship, see Carlson, Vincent, Hardesty, & Bearden, 2009). Research has shown that consumers' knowledge is pivotal in explaining their attitudes and behaviors (e.g., Alba & Hutchinson, 2000; Bettman & Park, 1980; Sujan, 1985). Further, objective (accurate stored information) and subjective (self-assessed) knowledge are two separate constructs (Park & Lessig, 1981) that differ in content and measurement, the manner in which they influence search and information processing (Raju, Lonial, & Mangold, 1995), their unique antecedents, and their intercorrelations (e.g., Alba & Hutchinson, 2000). It has long been recognized that consumer knowledge cannot be treated as a single construct; subjective and objective knowledge may have different effects on consumer information processing and separate constructs and measures need to be used to tap into the two dimensions (Park, Mothersbaugh, & Feick, 1994). Although a recent meta analysis has found a correlation of 0.37 between objective and subjective knowledge (Carlson et al., 2009), consistent with Park et al. (1994) we expect objective and subjective knowledge to have different, in fact, opposite effects on how IPAs influence consumers' satisfaction with an MCP (as discussed below). Objective knowledge Objective knowledge is the result of the process of cognition, or is the process of cognition itself, and represents the accumulation of an individual's cognitive activities (Coffey, 1958). Objective knowledge influences the way in which individuals search for and process information (e.g., Alba & Hutchinson, 1987; Park & Lessig, 1981). Such knowledge is often used in decision making by experienced consumers to compensate for meager cognitive resources (Yoon, Cole, & Lee, 2009). We refer to consumers with high objective knowledge as experts and those with low objective knowledge as novices. Experts possess more attribute-related thoughts and a richer store of prior knowledge than do novices (Sujan, 1985). These rich knowledge stores manifest in finely differentiated and hierarchically organized cognitive representations, as well as well-developed consumption rules with firmly entrenched beliefs and expectations about product performance (Brucks, 1985). In contrast, novices are less likely than experts to understand how an attribute relates to its functional category and, by definition are less experienced with the product (Alba & Hutchinson, 1987). Further, though novices certainly have their own personal tastes and preferences, they may not have stored in their memory a specific preference for the options of the personalizable attribute. For example, a novice cell phone shopper may not know which design (an IPA) he or she prefers, because he or she lacks any design format stored in memory. Thus, provided that such consumers are reasonably confident in their ability to discern their preferences, they are likely to evidence greater satisfaction as variety in IPA options increases. Consequently, novices should welcome increase in the number of available options of such “personalizing” attributes.
372
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
In contrast, experts, when compared with novices, due to their higher experience with the product, have access to their specific idiosyncratic preference for a specific option of the personalizable attribute. In other words, experts' higher experience should affect their predilection for large number of options of personalizable attributes, not because their experience reduces their desire for those attributes, but because they already have established preferences for the attributes. Therefore, experts should be less desirous than novices to gain access to variety in personalizable attribute options. Consequently, providing increasing variety of IPA options is superfluous for experts. As long as an option is provided that matches their preference (even if only few are offered in total), experts will be relatively satisfied, and additional options serve no purpose for them. Thus, we propose: Hypothesis 1. Variety in IPA options increases satisfaction with the MCP to a greater extent among novice consumers than among expert consumers. Subjective knowledge Subjective, or self-assessed, knowledge is the consumer's perception of what he or she knows, not necessarily a measure of the actual content of memory (Brucks, 1985), which makes it a measure of confidence (Alba & Hutchinson, 2000). In contrast to objective knowledge, subjective knowledge reflects the extent to which the consumer is confident of his or her capability to discern, evaluate, and discriminate incoming information. Because consumers are motivated to behave in a manner consistent with their subjective knowledge, such knowledge appears to influence information search and processing. However, subjective knowledge does not influence the ability to make product inferences, develop attribute-related thoughts, or improve processing efficiency, nor does it necessarily correlate with experience with the product (Alba & Hutchinson, 2000). In contrast with objective knowledge, subjective knowledge may not provide direct assistance in information search but it does influence the information search strategy by increasing consumers' reliance on their own evaluation skills (Brucks, 1985). For example, in a computerized grocery store study setting, Alba and Hutchinson (2000) find that participants who had subjective knowledge about fat exhibited a stronger need for consistency when they could not search in the low-fat aisles than when their search remained unrestricted. That is, even if subjective knowledge does not indicate actual product knowledge, it influences consumer motivation by making consumers feel they are knowledgeable, and this confidence encourages consumers to rely on their own evaluation skills. The above implies that subjective knowledge should increase consumers' predilection for personalizing attribute options. Indeed, in the absence of sufficient confidence in one's evaluations, more variety in idiosyncratic attributes options is likely to only induce confusion. Thus, whereas unsure (low subjective knowledge) consumers are unlikely to feel capable of discerning their preferred option no matter how many options are provided, confident (high subjective knowledge) consumers should feel that they can indeed ascertain their preferred option if more varieties are offered. Consequently, as the number of
options of IPAs increases, consumers with high subjective knowledge should experience greater increases in satisfaction with an MCP than do those with low subjective knowledge. Thus, we propose: Hypothesis 2. Variety in IPA options increases satisfaction with the MCP to a greater extent among consumers with high subjective knowledge than among those with low subjective knowledge. Knowledge calibration Consumers might be confident about their knowledge, but at the same time have low or high actual knowledge, because what they think they know may not be consistent with what they actually know. Thus, when we take the two knowledge constructs (objective and subjective knowledge) at high and low levels, we end up with four categories of consumers based on their knowledge: (1) low objective with high subjective knowledge (confident novices), (2) high objective with low subjective knowledge (unsure experts), (3) low objective with low subjective knowledge (unsure novices), and (4) high objective with high subjective knowledge (confident experts). Consumers whose objective knowledge does not match their subjective knowledge are referred to as miscalibrated consumers (Alba & Hutchinson, 2000). Although there is considerable literature on antecedents of miscalibration (e.g., Alba & Hutchinson, 2000; Griffin & Tversky, 1992), relatively scant attention has been paid to the consequences of miscalibration. As per the previous discussion, variety in IPA options influences satisfaction either if objective knowledge is low or subjective knowledge is high. It follows that consumers who are high in subjective knowledge and low in objective knowledge (confident novices) will be more satisfied by greater variety in IPA options than those who are low in subjective knowledge and high in objective knowledge (unsure experts) because the former have levels of both types of knowledge aligned to produce the effect, while the latter have levels of both types of knowledge that have no effect. In contrast, for consumers who possess one type of knowledge-enabling variety to have an effect, but concomitantly possess a second type of knowledge that is predicted to result in no effect of variety (calibrated consumers who are either high in objective and subjective knowledge or low in both), it is not possible to anticipate a priori which type of knowledge will dominate. However, comparing among calibrated consumers, because one type of knowledge has an effect and the other has not, there should not be any difference in the satisfaction experienced by the two types of calibrated consumers. We propose: Hypothesis 3. Knowledge calibration influences the relationship between variety in IPA options and satisfaction with the MCP, such that a) Variety in IPA options increases satisfaction with the MCP to a greater extent among consumers with high subjective and low objective knowledge than among those with low subjective and high objective knowledge.
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
b) Variety in variable attribute options does not differentially influence satisfaction with MCP among consumers with high subjective and high objective knowledge versus consumers with low subjective and low objective knowledge.
Methodology Products used For this research, we needed products that the participants could customize by selecting a specific option of certain attributes. We also needed products that the population (university students) would be familiar with. Therefore we selected cell phones, jeans, cars, and sports shoes. To determine the appropriate set of product attributes, we conducted pretests that identified a set of SPAs and IPAs for each product. All the experiments involved paper-and-pencil tasks. Experiments 1 and 2 focused solely on the effects of objective knowledge and subjective knowledge, respectively, whereas Experiment 3 enabled us to investigate both objective and subjective knowledge together. Experiment 1 Our goal for Experiment 1 was to test the first hypothesis; the experiment consisted of a 2 (number of options of SPAs: low, high) × 2 (number of options of IPAs: low, high) × 2 (levels of objective knowledge: low, high) between-participants design. Whereas the first two factors were manipulated, we measured the third factor because it is difficult to manipulate complex expectations or knowledge structures that develop over time and it is “ecologically more realistic” to measure than manipulate objective knowledge (Sujan, 1985, p. 36). The 118 undergraduate students at a Northeastern U.S. university who participated in this experiment received course credit and customized a hypothetical product (cell phone), after which they answered some survey questions. Participants took approximately 15–20 min to complete the task. Pretest 1 To identify cell phones' SPAs and IPAs, we conducted a pretest with an independent sample of students (N = 31) from a Northeastern university, who were given a detailed description of the classification and the nature of SPAs and IPAs (instrument available from the authors). They then indicated the extent to which a list of ten attributes of cell phones are shared-preference attributes by responding to three items (“______ is a universal attribute (SPA) for everyone/almost all people/a majority of people,”) anchored by 1 = “completely disagree” and 6 = “completely agree.” The same three items were repeated replacing universal with variable (“_______is a variable attribute (IPA)….”) to help participants indicate the extent of idiosyncraticity. (In the earlier stages of the research, we had used the universal and variable (Sherman et al., 1984) taxonomy and subsequently renamed them as SPA and IPA
373
respectively; hence, participants of all the experiments received instructions using the universal/variable labels). We compiled the initial list of attributes from an exhaustive examination of leading cell phone manufacturers' web sites and retail outlets. To create an overall attribute score, we first computed the sum of the scores for the three items pertaining to SPAs, and then subtracted from it the sum of the scores on the three items pertaining to IPAs. A mean difference greater than (less than) 0 implies that the attribute is shared-preference (idiosyncratic preferences). A one-sample, two-tailed t-test revealed statistically significant differences for seven of the ten attributes. From these results, we identify Design, Color, and Ring tones as the important IPAs, with Battery life, Coverage area, and Reception quality as the important SPAs. Pretest 2 We conducted a second pretest to assess any differences in the perceived importance of the six attributes. An independent sample of 81 participants from the same sample pool as the experiment participated in the pretest and indicated on a ten-point scale the importance of the six attributes (anchored at 1 = not at all important and 10 = very important). The mean comparisons of the sum of the importance of the three SPAs and the three IPAs revealed a higher importance rating for SPAs (MSPA = 8.61 N MIPA = 5.12; t = 14.52, p b 0.001). However, as we will clarify later, (in Measures section of Experiment 2) the results do not provide any support for attribute importance playing a significant role in the relationships we sought to investigate. Pretest 3 To determine the specific questions to include in the objective knowledge scale, we conducted a pretest with an independent sample (N = 34) of participants. Participants in this pretest responded to seven objective knowledge questions that assessed their overall knowledge about cell phones. Their responses were coded as 1 (0) if they answered the question correctly (incorrectly). For one item, all but one participant provided the correct answers, whereas two items generated a correct answer from only one respondent. Because the ability to discriminate between respondents who score high and low on the items is a key criterion for objective knowledge items (Park et al., 1994), we excluded the items on which participants performed uniformly well or uniformly poorly and retained four items that maximized discriminability (items available on request). Experts scored significantly higher on these items than did novices (Mexperts = 2.44 [range: 2–4] N Mnovices = 0.77 [range: 0–1]; t(116) = 16.13, p b 0.001). Procedure Upon their arrival at the behavioral lab, participants received their customization task (the customization platform was embedded in a booklet), and then completed a measurement task designed to assess their satisfaction with the MCP (hereafter, satisfaction). To mask the real purpose of the experiment, the participants were informed that their task was to customize a cell phone by selecting an option for each of the six attributes. The booklet provided instructions on the first page
374
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
that clearly explained the customization task. On the next two pages facing each other (the study instruments are available from the authors), we presented the three SPAs and the three IPAs, along with the available options of these attributes. Across experimental conditions, the range of options offered by an attribute remained constant. That is, the best and the worst options were the same, but the intermediate options varied in number across conditions. For the attribute Design, we included pictures (in black and white) of the various options while we offered verbal descriptions of the options for the other attributes knowing that because a cell phone is a highly familiar product, presenting one attribute visually would not interfere with the participants' semantic processing of the verbal descriptions of the other attributes (Wyer, Hung, & Jiang, 2008). For SPAs and IPAs, participants had three or nine options to select from in the low and high conditions, respectively. To increase the realism of the task, we associated a price with each option of each attribute, and the participants had a “budget” they could not exceed, which forced a constrained choice among options. The budget required the participants to make trade-offs among the attributes but still enabled them to choose from a wide variety of possible configurations. The budget did not influence satisfaction as evidenced by an ANOVA in which the amount spent by the participants as the dependent measure showed statistically insignificant main and interaction effects for the design variables (mean = 156.5, SD = 19.9). After customizing their cell phones, the participants reported their satisfaction using the three-item scale. A factor analysis of the scores on these three items yielded a one-factor model that explains 76% of the variation in the scale items (Cronbach's alpha of 0.84). Finally, participants responded to the objective knowledge questionnaire, which consisted of the four items from the pretest. Results We analyzed the data using regression analysis, with objective knowledge as a continuous variable (mean-centered), and the number of options of SPAs and IPAs as two-level categorical variables. As we show in Fig. 1, we find a statistically significant interaction between objective knowledge and the number of options of IPAs (F = 11.51, p b 0.001). We used the regression equation to estimate the predicted levels of satisfaction (ŷ) at one standard deviation above and below the mean value of objective knowledge. In support of Hypothesis 1, when objective knowledge is low (one standard deviation [SD] below the mean), increasing the number of options of IPAs increases satisfaction (Δŷ = 1.24, t = 3.30, p b 0.002), but with high objective knowledge (one SD above the mean), the number of options of IPAs has no effect on satisfaction (t = −0.30, p N 0.40). The interaction between objective knowledge and number of options of SPAs options is not statistically significant (F = 0.45, p N 0.50). Discussion The results of Experiment 1 suggest a meaningful difference in the manner in which objective knowledge (experts versus novices) moderates consumers' preference for the number of options of IPAs. In contrast, we find no such differences for the
Fig. 1. Effect of objective knowledge on satisfaction with MCP. This figure shows the effect of objective knowledge on the relationship between number of options of IPAs and satisfaction with MCP from Experiment 1.
effect of SPAs on satisfaction. Our results may reflect the differential importance of the two sets of attributes (SPAs and IPAs) rather than their differential idiosyncraticity. However, because our pretest showed that SPAs tend to be rated as more important than IPAs, we should have found an interaction between the levels of objective knowledge and number of options of SPAs. We do not find support for such an interaction, diminishing the possibility of attribute importance causing the results. We investigated the potential explanatory role of differential importance of the attributes further in Experiment 2, where we also tested Hypothesis 2, using different sets of products (cars and jeans) to establish the robustness of our results. Because it was established in Experiment 1 that our results hold only for IPAs, we only manipulated the number of options of the IPAs in Experiments 2 and 3. Experiment 2 An alternative explanation for the preceding results could be that our attribute classification also maps exactly onto a functional–peripheral classification. That is, the SPAs we have tested may also have been functional attributes, and the IPAs may have been peripheral, thereby undermining our explanation based on the personalizabilty of the attributes. In Experiment 2, we attempted to refute this alternative explanation by replicating the results with both functional and peripheral attributes. To this end, we conducted another pretest to identify attributes that are shared-preference or SP-functional, SP-peripheral, idiosyncratic preference or IP-functional, and IP-peripheral. Pretest 1 A panel of researchers (senior faculty at a large Northeastern U.S. university, N = 5) were invited to participate in a group discussion and received a briefing about the objective of the discussion. They generated a taxonomy of the attributes of fashion products (apparel) and cars by classifying them into four categories: SP-functional, SP-peripheral, IP-functional, and IPperipheral. The panel received detailed written and oral
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
explanations of the four classifications, along with a set of “seed” attributes of jeans and cars that they classified. These panelists also were asked to suggest other attributes of jeans and cars that could be classified similarly. A subsequent group discussion with regard to the taxonomy showed that the classification of some attributes was unanimous, though for others, there were some differences. The group discussion enabled the participants to reach consensus in assigning the conflicting attributes to a category. Pretest 2 This pretest, conducted among undergraduate students at a large Northeastern university, served to confirm the results of the first pretest. The participants identified sets of jeans and car attributes as IP-functional or IP-peripheral. To reduce participant fatigue, we asked two independent samples of students (N = 25 and N = 32) to rate 9 (6 car and 3 jeans) attributes or 9 (5 car and 4 jeans) attributes as SP or IP and functional or peripheral. To discern any significant differences in the levels of confidence in these ratings, we also measured the extent to which the participants understood the attribute descriptions and their confidence in their ratings. At the start of the pretest, we explained to the participants what we meant by SP, IP, functional, and peripheral attributes in a manner similar to the first pretest of Experiment 1 (instrument available from authors). The participants then rated each attribute as SP or IP on a continuum, in response to the following question: “How would you classify the ______ of a car/jeans?” anchored by 1 = “Universal (SPA)” and 9 = “Variable (IPA).” Participants subsequently responded to the same question but anchored by 1 = “Functional” and 9 = “Peripheral” for each attribute. The participants indicated how confident they were about each evaluation according to the item, “How confident are you in your ability to evaluate the ________ of a car/jeans?” anchored by 1 = “Not confident at all” and 9 = “Very confident.” Finally, they indicated the clarity of the attribute description by responding to the question: “To what extent do you think the ________ is clear to you?” anchored by 1 = “Not clear at all” and 9 = “Very clear.” We then computed the mean ratings on the SP–IP, functional– peripheral, confidence, and clarity scales. Because we were looking for IP-peripheral and IP-functional attributes, we needed attributes whose IP score was high and functional–peripheral score was high and low respectively. Within attributes whose IP score was at least N4, we selected the two attributes with the highest (lowest) score on the functional–peripheral scale as IPperipheral (IP-functional) attributes. Based on these criteria, among jeans attributes, Color (IPscore = 7.96; functional– peripheralscore = 8.08) and Style (IPscore = 6.53; functional– peripheralscore = 5.72) emerged as IP-peripheral whereas Fit (IPscore = 7.13; functional–peripheralscore = 3.41) and Stitch quality (IPscore = 4.4; functional–peripheralscore = 3.92) emerged as IPfunctional attributes respectively. Among car attributes, Interior color (IPscore = 7.08; functional–peripheralscore = 8.25) and Exterior style (IPscore = 6.84; functional–peripheralscore = 5.5) emerged as IP-peripheral whereas Overall design (IPscore = 6.28; functional–peripheralscore = 3.78) and Gearshift knob style (IPscore = 5;
375
functional–peripheralscore = 4.8) emerged as IP-functional attributes respectively. The degree of confidence in the evaluations and the clarity of the attributes was generally high for the selected attributes (Mconfidence = 7.08 [range: 5.66–8.81]; Mclarity = 7.23, [range: 6.18–8.58]). Measures We measured participants' subjective knowledge using a variant of the scale, used by Park et al. (1994) that asked participants to respond to three items (“I think I know a lot about cell phones in general,” “Compared to my friends, I think I know a lot about cell phones,” and “Compared to experts, I think I know a lot about cell phones.”) anchored by 1 = “completely disagree” and 7 = “completely agree.” A factor analysis of the three-item scale yielded a single factor with an eigenvalue N1 that explained 78% of the variation in the scores (Cronbach's alpha of 0.88). We measured satisfaction using the scale we developed, as described previously. A factor analysis of the three-item satisfaction scale yielded a single factor with an eigenvalue N 1 that explained 78% of the variation. We also measured the importance of the four IPAs that we tested for both jeans and cars on a seven-point scale (1 = “not at all important” and 7 = “very important”). Among jeans IPAs, Fit was considered more important than Style (Mdiff = 0.16, p b 0.05), Color (Mdiff = 0.48, p b 0.001) and Stitch quality (Mdiff = 2.21, p b 0.001); Style was considered more important than Color (M diff = 0.32, p b 0.001) and Stitch quality (Mdiff = 2.06, p b 0.001); and Color was considered more important than Stitch quality (Mdiff = 1.7, p b 0.001). Similarly, among car IPAs, Overall design was considered more important than Gear shift knob style (Mdiff = 3.24, p b 0.001), Interior style (M diff = 1.04, p b 0.001), and Interior color (M diff = 1.16, p b 0.001); Gear shift knob style was more important than Interior style (M diff = 2.2, p b 0.001) and Interior color (Mdiff = 2.09, p b 0.001).Thus we see that for both jeans and cars, the four IPAs together influenced the relationship between objective knowledge and satisfaction despite considerable variance in the importance among the four IPAs. In contrast, in Pretest 2 of Experiment 1, SPAs in general emerged as more important than IPAs, (Mdiff = 3.5 on a 10-point importance scale, p b 0.001) showing that SPAs had no influence on the relationship between objective knowledge and satisfaction despite being more important than IPAs. This evidence conclusively refutes the role of importance of attributes as an alternative explanation for our results. Procedure 134 students from a large Northeastern university took part in this experiment for extra credit. Following a procedure similar to that of the previous experiment, we asked them to customize their own pair of jeans and a car. Unlike the first experiment though, Experiment 2 focused only on the effects of IPAs, so we provided participants only with varying number of options of the IPAs in a 2 (number of options of IPAs: low, high) × 2 (subjective knowledge) between-subjects design. Although each participant was exposed to both jeans and car attributes, the set of attributes differed for the two products. We
376
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
counterbalanced the order of exposure to jeans and cars, and participants did not have an imaginary budget; therefore, we investigated the effects of the focal attributes in isolation from price and avoided any possibility of trade-offs between attributes. We provided participants with several options (three in the low and six in the high conditions) of the selected attributes (instrument available used on request) and asked them to select one option for each attribute. Next, they completed the satisfaction and subjective knowledge scales. Results We do not find a three-way interaction of the number of options of variable attributes, subjective knowledge, and order of presentation for jeans (F = 0.01, p N 0.90) or cars (F = 0.51, p N 0.46). We used regression analysis to analyze the data, with subjective knowledge as a continuous variable (mean-centered) and the number of options of IPAs as a two-level categorical variable. As we show in Fig. 2, for jeans, we find a statistically significant interaction between subjective knowledge and the number of options of IPAs (F = 5.80, p b 0.02). Again, we used the regression equation to estimate predicted values of satisfaction (ŷ) at low and high levels of subjective knowledge. In support of Hypothesis 2, when subjective knowledge is low (one SD below the mean), satisfaction does not vary significantly between the number of IPA options (t = 0.44, p N 0.20); however, with high subjective knowledge (one SD above the mean), satisfaction increases significantly with more options of IPAs (Δŷ = 1.55, t = 4.76, p b 0.01). Similarly, as we see in panel B of Fig. 2, there is a statistically significant interaction between subjective knowledge and the number of options of IPAs for the car category (F = 6.60, p b 0.01). For cars, satisfaction with MCP increases at higher number of options of IPAs both at one SD above the mean (Δŷ = 2.44, t = 6.69, p b 0.001), and at one SD below the mean, albeit to a lesser extent (Δŷ = 0.93, t = 2.56, p b 0.02). However, the differential effects as seen for jeans are replicated for cars at two SDs above and below the mean. Specifically, satisfaction with MCP, while invariant at two SDs below the mean (t = 1.18, p N 0.23), increases with higher number of options of IPAs at two SDs above the mean (Δŷ = 2.81, t = 5.77, p b 0.001). This difference in the effects between jeans and cars appears to result from a differential confidence or subjective knowledge (Alba & Hutchinson, 2000) about cars as compared to jeans among our participants; these undergraduate students not only had lesser subjective knowledge about cars than about jeans (Mcars = 3.22 b Mjeans = 3.92, t(131) = 3.72, p b 0.001) but also exhibited greater variance in their subjective knowledge about cars (σ2 cars = 2.20 N σ2 jeans = 1.84). Thus, there were fewer participants above and below one SD of the mean for cars (N = 40) than for jeans (N = 60). Discussion We find support for Hypothesis 2 for both jeans and cars. Experiments 1 and 2 separately test for and provide evidence in support of Hypothesis 1 and Hypothesis2. However, we did not investigate the interactions of calibrated versus miscalibrated consumers (Hypothesis3) in these experiments. Specifically, do
Fig. 2. Effect of subjective knowledge on satisfaction with MCP. This figure shows the effect of subjective knowledge on the relationship between number of options of IPAs and satisfaction with MCP from Experiment 2.
consumers with high levels of both objective and subjective knowledge or low levels of both differ from those who are high in one type of knowledge but not in another? To investigate the possible role of knowledge calibration, we conducted Experiment 3 and measured both objective and subjective knowledge, using a new product category (sports shoes) to test the robustness of our findings across a range of products. Experiment 3 Pretest To categorize sports shoes' attributes as shared-preference and idiosyncratic-preference attributes, we conducted a pretest among 60 undergraduate students at a large Northeastern university. At the start of the pretest, we explained to the participants what we meant by SP, IP, functional, and peripheral attributes similar to the previous experiments (instrument available from authors). The participants then rated each attribute, using the question “How would you classify the ______ of sports shoes?” on continua anchored by 1 = “Universal (SPA)” and 9 = “Variable (IPA)” and 1 = “Functional” and 9 = “Peripheral.” As in Experiment 2, the participants then indicated their confidence in their evaluations and the clarity of the attribute description. We computed the mean ratings on the SP–IP, functional– peripheral, confidence, and clarity scales, along with pair wise differences among the set of attributes (Table 1). We selected attributes that scored significantly high on the SP–IP and
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
functional–peripheral scales. Because the degree of confidence in the evaluations and the clarity of the attributes were generally high, we did not drop any of the attributes that we selected based on the SP–IP and functional–peripheral scales. The pair wise differences along with the confidence and clarity ratings suggest Color and Design as IP-peripheral attributes, with Activity type and Size as the IP-functional attributes. Measures In the pretest, we had asked participants to respond to seven items that tested their knowledge of sports shoes, which we obtained from web sites of leading sports shoe brands that offered customization (items available from the authors). We retained four items that maximized discriminability. Experts scored significantly higher on these items than did novices (Mexperts = 3.25 [range: 3–4] N Mnovices = 1.6 [range: 0–2]; t(162) = 18.78, p b 0.001). Procedure 165 students at a large Northeastern university received extra credit, as in previous experiments, for participating in this experiment. Unlike the previous experiments though, we measured participants' objective and subjective knowledge of sports shoes, using a 2 (number of options of IPAs: low, high) × 2(objective knowledge) × 2 (subjective knowledge) between-subjects design. Following a procedure similar to the previous studies, we asked them to customize their own pair of sports shoes; they were not constrained by any imaginary budget. We provided options (three in the low and six in the high conditions) of the selected attributes (instrument available on request) and asked the participants to select one option for each attribute, and then complete the satisfaction scale. We added an item to the satisfaction scale (“I was satisfied with the options offered for each attribute”) for this experiment. The four
377
items achieved a reliability of 0.90. Finally, participants completed the objective and subjective knowledge scales as in previous experiments. All participants answered the subjective knowledge questions before the objective knowledge questions. Results We used regression analysis to analyze the data, with objective and subjective knowledge as continuous variables (mean-centered) and the number of options of IPAs as a twolevel categorical variable. As expected, increasing the number of options of IPAs increases satisfaction (β = 0.81, t = 3.58, p b 0.001). Further, as seen in Fig. 3, we find a significant interaction between the number of options of IPAs and subjective knowledge (t = 2.69, p b 0.01). Furthermore, the interaction between number of options and objective knowledge also approaches significance (t = − 1.69, p b 0.10). The regression equation predicting the values of satisfaction with MCP (ŷ) at low and high levels of objective knowledge does not reveal any effects of increasing the number of options of IPAs when objective knowledge is high (one SD above mean) (t = 1.33, p N 0.18), but when objective knowledge is low (one SD below mean), increasing the number of options of IPAs increases satisfaction (Δŷ = 1.2, t = 3.69, p b 0.001), in support of Hypothesis 1. Moreover, the regression equation predicting the values of satisfaction with MCP (ŷ) at low and high levels of subjective knowledge does not reveal any effects of changing the number of options of IPAs when subjective knowledge is low (one SD below mean) (t = 0.58, p N 0.5), but when it is high (one SD above mean), increasing the number of options of IPAs increases satisfaction (Δŷ = 1.44, t = 4.37, p b 0.001), in support of Hypothesis 2. High (low) objective knowledge might cause high (low) confidence or subjective knowledge, and each consumer possesses both types of knowledge which means we need to
Table 1 Pretest—Experiment 3: to identify IP-Peripheral and IP-Functional attributes. Attribute
SP–IP score 1 = Shared-preference; 9 = Idiosyncratic preference
Functional–peripheral score 1 = Functional; 9 = Peripheral
Confidence score 1 = Not confident; 9 = Very confident
Clarity score 1 = Not clear; 9 = Very clear
Color Design Size Activity type Outsole Collar material Width Tip and heel material Weight Midsole
8.07 7.77 6.77 6.48 6.15 6.05 5.53 4.97 4.7 4.3
8.67 7.67 2.17 1.97 2.35 5.38 2.67 3.6 2.4 1.95
8.37 7.93 7.87 7.75 5.5 5.2 6.02 5.17 6.15 4.58
8.48 8.07 8.45 7.87 6.72 5.33 7.28 5.62 6.78 5.6
Note: Attribute Color and Design Size Activity type Outsole
Based on pairwise differences, using the Universal–Variable Scale, attribute significantly different (p b 0.001) from: All other attributes All other attributes except: outsole, collar material, activity type All other attributes except: outsole, collar material, size All other attributes except: collar material, width, activity type, and size
Based on pairwise differences, using the Functional–Peripheral Scale, attribute significantly different (p b 0.001) from: All other attributes All other attributes except: midsole All other attributes except: midsole, activity type, weight All other attributes except: width, activity type, weight, and size
378
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
undermines the effects produced by high subjective knowledge. In contrast, when consumers are miscalibrated, the two types of knowledge are completely aligned and the effects of low objective and high subjective knowledge get consolidated, leading to the results we achieved. General discussion
Fig. 3. Effect of objective and subjective knowledge on satisfaction with MCP. This figure shows the effect of objective and subjective knowledge on the relationship between number of options of IPAs and satisfaction with MCP from Experiment 3.
tease out the effects of the combinations of high and low levels of each of the two knowledge constructs. According to our theorizing leading to the hypotheses, the predicted effects will not be seen for participants with high levels of both subjective and objective knowledge (confident experts) or low levels of both (unsure novices). However, the miscalibrated participants who are either high on subjective but low on objective knowledge (confident novices) or low on subjective but high on objective knowledge (unsure experts) will show the proposed effects. Thus, the evidence for or against Hypothesis 3 is provided by comparisons across the four possible combinations of levels of objective and subjective knowledge. Therefore, we examined the differences between the two types of miscalibrated and the two types of calibrated participants at low and high number of options of IPAs. As expected, confident novices are significantly more satisfied (despite their lack of expertise) than unsure experts when there is a large number of options of IPAs (t = − 2.01, p b 0.05). In contrast, we find no difference between confident novices and unsure experts when the number of options of IPAs is low (t = 1.58, p N 0.10). Confident experts are not significantly more satisfied than unsure novices, regardless of whether the number of options of IPAs is high (t = 0.78, p N 0.40) or low (t = − 0.35, p N 0.70), which shows that our hypothesized effects do not hold for calibrated participants. Thus, we find strong support for Hypothesis 3. The results therefore are consistent with Alba and Hutchinson's (2000) finding that high objective knowledge
Mass customization is growing in importance because it could provide additional value for consumers who have hitherto purchased standardized products. In our research, we have sought to examine how the characteristics of an MCP (i.e., the type and number of attribute options it offers) influence consumers' satisfaction with the platform. Adapting existing literature on individual qualities (Sherman et al., 1984), we have proposed that during a mass customization process, consumers differentially process information about SPAs versus IPAs. Our results show that for novice consumers, other things being equal, increasing the number of IPA options increases satisfaction with the MCP. However, the number of SPA options does not generate differential levels of satisfaction for expert consumers, when compared to novice consumers. The results from Experiment 1 suggest that novice consumers might be evaluating the IPAs more carefully than do expert consumers. Similarly, consumers with high subjective knowledge appear to be more satisfied with an MCP when it contains more IPA options than are consumers with low subjective knowledge. Again, subjective knowledge does not moderate the relationship between SPAs and satisfaction. Finally, our results hold only for miscalibrated consumers, such that confident novices are more satisfied (despite their lack of expertise) than are unsure experts when the number of options of IPA is high, but there is no difference between confident novices and unsure experts when the number of options of the IPAs is low. Our research provides important insights for theory and practice. Theoretically, our research provides a better understanding of the processes that underlie the effects of varying the attribute options available in an MCP. We have theorized that because an MCP involves consumers configuring products with multiple attributes, their satisfaction with the process would depend on a) whether those attributes reflect universal or idiosyncratic preferences, and b) on consumer knowledge. Our research suggests that information about IPAs is more relevant than information about SPAs for consumers who have lower levels of objective knowledge and greater confidence in their knowledge. We also confirm a difference predicted by Park et al. (1994) in how objective and subjective knowledge influence the relationship between number of options of attributes and satisfaction. While we show that subjective knowledge has effects opposite to that of objective knowledge, our results also replicate and validate the calibration literature by revealing that the effects hold only for miscalibrated consumers (either high in objective and low in subjective knowledge or low in objective and high in subjective knowledge) and not for calibrated consumers (i.e., consumers high in objective and subjective knowledge or low in objective and subjective knowledge). Our research also shows that recognizing the moderating role of knowledge-related constructs
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
enables researchers to gain deeper insights into cognitions about SPAs and IPAs. Extant research already has classified attributes mainly on the basis of their intrinsic characteristics. Our classification reflects how consumers evaluate attributes and thus should be very useful for research on mass customization, as well as consumer research in general. We expect that it would be worthwhile for consumer researchers to link the shared versus idiosyncratic preference attribute dichotomy to other aspects of consumer research, such as consumer emotions. For practitioners, our consumer-focused attribute classification may help them develop market-driving strategies that incorporate consumer behavior (Carpenter, Glazer, & Nakamoto, 1997). By noting the difference in the effects of objective and subjective knowledge on satisfaction with the MCP, marketing managers designing MCPs can segment their markets (e.g., by including screening questions at the beginning of the customization process) and thereby increase satisfaction with the MCP among novice and confident consumers by offering them more options of IPAs. Our findings also could be useful in different contexts in which variety plays an important role, such as assortments of goods in a grocery store (e.g., Arnold, Oum, & Tigert, 1983). Firms can directly influence subjective knowledge (e.g., Moorman, Diehl, Brinberg, & Kidwell, 2004), such as through strong advertising messages that reiterate the consumers' sense of being knowledgeable about the product. Or, firms could invite consumers to participate in relatively trivial contests or tasks and provide positive feedback or reward good performance, which would result in increased in subjective knowledge. Managers should consider simultaneously designing MCPs that offer greater variety in IPA options and developing appropriate programs that will increase consumers' subjective knowledge to enhance their satisfaction with those MCPs. Our research represents a first step in exploring mass customization using theories and insights from consumer behavior literature. Our classification of attributes is based on how consumers evaluate them but is by no means the only useful way to classify attributes for mass customization. Additional research might explore other attribute classifications and address other cognitive and situational factors (e.g., gift giving and mood) and individual traits (e.g., tolerance of ambiguity and regulatory focus) that seem critical for understanding consumer behavior in a mass customization context. Consumers using an MCP are co-creators of the product, which suggests a shared attribution between the provider and the consumer for the success or failure of a product. Attribution theory therefore may be a relevant framework for developing an appropriate schema for evaluating MCPs (e.g., Tsiros & Mittal, 2001). In this context, the investigation of affect is especially interesting because the attribution could influence appraisal, (e.g., an appraisal of blame leading to anger in the case of an unfavorable purchase outcome) (Madrigal, 2008) which could further impact the “attachment” towards the customized product or brand (Fedorikhin, Park, & Thomson, 2008). We examined only one aspect of customization, that is, satisfaction with the MCP, though several other aspects of the
379
process (e.g., complexity and satisfaction with the product) could have important implications. Also, it is possible that other individual differences besides knowledge may influence consumers' predilection for personalizing attributes. For example, because narcissists are more likely to undervalue utilitarian value in favor of symbolic or self-presentation value (Sedikides, Gregg, Cisek, & Hart, 2007), it would be interesting to investigate whether narcissists exhibit a greater preference for idiosyncratic-preference or shared-preference attributes. While the former are more likely to represent the individual uniquely, the latter may well be more likely to be indicators of status. Finally, it would be useful for further research to explore the external validity of our results, perhaps with the cooperation of an online retailer that offers mass customization. This study design would enable researchers to explore the effects of MCPs on additional dependent variables, including the amount of effort the consumer exerts during the customization process (e.g., number of clicks/web pages opened), whether the consumer purchases the product, and the total revenue realized by the retailer at varying number of customization options. Acknowledgments The authors acknowledge feedback from Professors Hans Baumgartner and Meg Meloy. References Alba, J. W., & Hutchinson, J. W. (1987). Dimensions of consumer expertise. Journal of Consumer Research, 13(4), 411−454. Alba, J. W., & Hutchinson, J. W. (2000). Knowledge calibration: What consumers know and what they think they know. Journal of Consumer Research, 27(2), 123−156. Arnold, S. J., Oum, T. H., & Tigert, D. J. (1983). Determinant attributes in retail patronage: Seasonal, temporal, regional, and international comparisons. Journal of Marketing Research, 20(2), 149−157. Bettman, J. R., & Park, C. W. (1980). Effects of prior knowledge and experience and phase of the choice process on consumer decision processes: A protocol analysis. Journal of Consumer Research, 7(3), 234−248. Brucks, M. (1985). The effects of product class knowledge on information search behavior. Journal of Consumer Research, 12(1), 1−16. Carlson, J. P., Vincent, L. H., Hardesty, D. M., & Bearden, W. O. (2009). Objective and subjective knowledge relationships: A quantitative analysis of consumer research findings. Journal of Consumer Research, 35(5), 864−876. Carmon, Z., & Simonson, I. (1998). Price-quality trade-offs in choice versus matching: New insights into the prominence effect. Journal of Consumer Psychology, 7(4), 323. Carpenter, G. S., Glazer, R., & Nakamoto, K. (1997). Readings on marketdriving strategies: Towards a new theory of competitive advantage. New York, NY: Addison Wesley Longman. Celsi, R. L., & Olson, J. C. (1988). The role of involvement in attention and comprehension processes. Journal of Consumer Research, 15(2), 210−224. Coffey, P. (1958). Epistemology, or the theory of knowledge: An introduction to general metaphysics. Gloucester: P.Smith. Coulter, R. A., Price, L. L., & Feick, L. (2003). Rethinking the origins of involvement and brand commitment: Insights from postsocialist central Europe. Journal of Consumer Research, 30(2), 151−169. Duray, R., & Milligan, G. W. (1999). Improving customer satisfaction through mass customization. Quality Progress, 32(8), 60−66. Fedorikhin, A., Park, C. W., & Thomson, M. (2008). Beyond fit and attitude: The effect of emotional attachment on consumer responses to brand extensions. Journal of Consumer Psychology, 18(4), 281−291.
380
S. Puligadda et al. / Journal of Consumer Psychology 20 (2010) 369–380
Griffin, D., & Tversky, A. (1992). The weighing of evidence and the determinants of confidence. Cognitive Psychology, 24(3), 411−435. Hart, C. W. L. (1995). Mass customization: Conceptual underpinnings, opportunities and limits. International Journal of Service Industry Management, 6(2), 36−45. Horsky, D., Nelson, P., & Posavac, S. S. (2004). Stating preference for the ethereal but choosing the concrete: How the tangibility of attributes affects attribute weighting in value elicitation and choice. Journal of Consumer Psychology, 14(1/2), 132−140. Hsee, C. K. (1996). The evaluability hypothesis: An explanation for preference reversals between joint and separate evaluations of alternatives. Organizational Behavior & Human Decision Processes, 67(3), 247−257. Huffman, C., & Kahn, B. E. (1998). Variety for sale: Mass customization or mass confusion? Journal of Retailing, 74(4), 491−513. Jain, S. P., Buchanan, B., & Maheswaran, D. (2000). Comparative versus noncomparative advertising: The moderating impact of prepurchase attribute verifiability. Journal of Consumer Psychology, 9(4), 201−211. Jain, S. P., & Posavac, S. S. (2001). Prepurchase attribute verifiability, source credibility, and persuasion. Journal of Consumer Psychology, 11(3), 169−180. Jiao, J., Ma, Q., & Tseng, M. M. (2003). Towards high value-added products and services: Mass customization and beyond. Technovation, 23(10), 809−821. Kahn, B. E. (1998). Dynamic relationships with customers: High-variety strategies. Academy of Marketing Science Journal, 26(1), 45−53. Lee, B. -K., & Lee, W. -N. (2004). The effect of information overload on consumer choice quality in an on-line environment. Psychology & Marketing, 21(3), 159−183. MacDuffie, J. P., Sethuraman, K., & Fisher, M. L. (1996). Product variety and manufacturing performance: Evidence from the international automotive assembly plant study. Management Science, 42(3), 350−369. Madrigal, R. (2008). Hot vs. cold cognitions and consumers' reactions to sporting event outcomes. Journal of Consumer Psychology, 18(4), 304−319. Moorman, C., Diehl, K., Brinberg, D., & Kidwell, B. (2004). Subjective knowledge, search locations, and consumer choice. Journal of Consumer Research, 31(3), 673−680. Murthi, B. P. S., & Sarkar, S. (2003). The role of the management sciences in research on personalization. Management Science, 49(10), 1344−1362. Park, C. W., & Lessig, V. P. (1981). Familiarity and its impact on consumer biases and heuristics. Journal of Consumer Research, 8, 223−230. Park, C. W., Mothersbaugh, D. L., & Feick, L. (1994). Consumer knowledge assessment. Journal of Consumer Research, 21(1), 71−82. Petty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and peripheral routes to advertising effectiveness: The moderating role of involvement. Journal of Consumer Research, 10(2), 135−146.
Pine, B. J., & Gilmore, J. H. (2000). Satisfaction, sacrifice, surprise: Three small steps create one giant leap into the experience economy. Strategy & Leadership, 28(1), 18−23. Raju, P. S., Lonial, S. C., & Mangold, W. G. (1995). Differential effects of subjective knowledge, objective knowledge, and usage experience on decision making: An exploratory investigation. Journal of Consumer Psychology, 4(2), 153. Sedikides, C., Gregg, A. P., Cisek, S., & Hart, C. M. (2007). The I that buys: Narcissists as consumers. Journal of Consumer Psychology, 17(4), 254−257. Sherman, S. J., Chassin, L., Presson, C. C., & Agostinelli, G. (1984). The role of the evaluation and similarity principles in the false consensus effect. Journal of Personality and Social Psychology, 47(6), 1244−1262. Simonson, I. (2005). Determinants of customers' responses to customized offers: Conceptual framework and research propositions. Journal of Marketing, 69(1), 32−45. Simonson, I. (2008). Will I like a medium pillow? Another look at constructed and inherent preferences. Journal of Consumer Psychology, 18(3), 155−169. Sujan, M. (1985). Consumer knowledge: Effects on evaluation strategies mediating consumer judgments. Journal of Consumer Research, 12(1), 31−46. Tsiros, M., & Mittal, V. (2001). The role of attributions in evaluating exchange relationships. Advances in Consumer Research, 28(1), 447-447. Tu, Q., Vondermbse, M. A., & Ragu-Nathan, T. S. (2001). The impact of timebased manufacturing practices on mass customization and value to customer. Journal of Operations Management, 19(2), 201−217. Wind, J., & Rangaswamy, A. (2001). Customerization: The next revolution in mass customization. Journal of Interactive Marketing, 15(1), 13−32. Wyer, R. S., Hung, I. W., & Jiang, Y. (2008). Visual and verbal processing strategies in comprehension and judgment. Journal of Consumer Psychology, 18(4), 244−257. Yoon, C., Cole, C. A., & Lee, M. P. (2009). Consumer decision making and aging: Current knowledge and future directions. Journal of Consumer Psychology, 19(1), 2−16. Zhang, S., & Fitzsimons, G. J. (1999). Choice-process satisfaction: The influence of attribute alignability and option limitation. Organizational Behavior and Human Decision Processes, 77(3), 192−214. Zhang, S., & Markman, A. B. (2001). Processing product unique features: Alignability and involvement in preference construction. Journal of Consumer Psychology, 11(1), 13−27. Zhang, S., Kardes, F. R., & Cronley, M. L. (2002). Comparative advertising: Effects of structural alignability on target brand evaluations. Journal of Consumer Psychology, 12(4), 303−311.
