Impact of retail environment extraordinariness on customer self-concept

Impact of retail environment extraordinariness on customer self-concept

Journal of Business Research 64 (2011) 551–557 Contents lists available at ScienceDirect Journal of Business Research Impact of retail environment ...

237KB Sizes 0 Downloads 16 Views

Journal of Business Research 64 (2011) 551–557

Contents lists available at ScienceDirect

Journal of Business Research

Impact of retail environment extraordinariness on customer self-concept Velitchka D. Kaltcheva a,⁎, Anthony Patino a,1, Jean-Charles Chebat b,2 a b

Department of Marketing and Business Law, Loyola Marymount University, 1 LMU Drive, Los Angeles, CA 90045, United States ECSC Chair Professor at HEC Montréal and Affiliate Professor at RMS, 5540 avenue Louis Colin, Montréal (Québec), Canada H2T 3A7

a r t i c l e

i n f o

Article history: Received 1 January 2009 Received in revised form 1 March 2009 Accepted 1 June 2010 Keywords: Retail environment Self-concept Accessibility

a b s t r a c t This research examines the effect of extraordinary retail environments on consumer self-concept. Two between-subjects experiments manipulate the extraordinariness of the retail environment and evaluate participants' self-concept in the environment. In both experiments, high-extraordinariness retail environments elicit a more atypical working self-concept than low-extraordinariness environments. Content analysis of participants' working self-concept in the two environments offers insight into the cognitive processes underlying the effect. The article discusses managerial and theoretical implications of the research. © 2010 Elsevier Inc. All rights reserved.

A number of retailers invest in retail environments that create extraordinary experiences for customers, vicariously transporting them to a different time and/or place. For example, the Johnny Rockets restaurant chain immerses customers in the atmosphere of the 1940s with its themed fixtures and music (Bellantonio, 2003); the restaurants and stores at Disney's EPCOT Center in Orlando, Florida, take visitors on a vicarious tour around the world (Johnson, 2002). Some retailers even create environments that have no prototypes in the real world. For example, the Mars 2112 restaurant in New York takes customers on a voyage to Mars (Elan, 1998). The retail literature suggests that any impact extraordinary retail environments may have on consumer self-concept will ultimately impact consumer shopping behavior (Chebat et al., 2006; Sirgy et al., 2000). Thus, the first objective of this research is to test whether extraordinary retail environments elicit identities that are significantly different from the consumer's typical self-concept. The second objective is to offer insight into the process of how retail environments varying in extraordinariness may elicit different self-concepts. The next section defines environment extraordinariness and the self-concept. Then, the article describes two experiments in which participants view retail environments that differ in extraordinariness. The experiments measure participants' self-concept in the environ-

ment and then compare this environment-elicited self-concept to the participant's typical self-concept. The final section of the article discusses implications of this research, its limitations, and future research directions.

⁎ Corresponding author. Department of Marketing and Business Law, Loyola Marymount University, 1 LMU Drive, Los Angeles, CA 90045, United States. Tel.: +1 310 338 5167; fax: +1 310 338 3000. E-mail addresses: [email protected] (V.D. Kaltcheva), [email protected] (A. Patino), [email protected] (J.-C. Chebat). 1 Tel.: + 1 310 338 5167; fax: + 1 310 338 3000. 2 Tel.: + 1 514 340 6846; fax: + 1 514 340 6097.

1.2. The self-concept

0148-2963/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jbusres.2010.06.011

1. Theoretical background 1.1. Environment extraordinariness Extraordinary retail environments are “marvelously different from normal … contemporary sources of contact with the miraculous and numinous” (Belk, 2000, p. 106), “new worlds that consumers interpret as different realities” (Kozinets et al., 2004, p. 669). Environment extraordinariness is purely subjective. A retail environment may appear highly extraordinary to some consumers, while other consumers may find the same environment mundane. Environment extraordinariness is not the same as environment atypicality. Retail environment typicality defines the degree to which an environment matches consumers' expectations (Babin and Babin, 2001). Thus, if the ambience of a Nordstrom store were designed like that of a Wal-Mart store, consumers would be likely to consider that Nordstrom store as atypical but hardly as a different world. Extraordinary retail environments are generally atypical; however, consumers are not likely to perceive all atypical environments as extraordinary.

The self-concept involves various self-relevant attributes (mental representations of the self): dispositional characteristics and traits; perceptions of the ongoing or past experiences, social roles, and behaviors; self-related attitudes and affect (Markus and Wurf, 1987;

552

V.D. Kaltcheva et al. / Journal of Business Research 64 (2011) 551–557

Sirgy, 1982). Typically, people have numerous self-relevant attributes, but only a subset of those attributes is salient in working memory at any one time (e.g., Markus and Kunda, 1986; Niedenthal and Beike, 1997; Rhodewalt, 1986). This subset of momentarily salient selfrelevant attributes is known as the working self-concept (Markus and Kunda, 1986) or the phenomenal self (Baumeister, 1998). Thus, while the self-concept includes all the self-relevant attributes that an individual has, the working self-concept includes only those attributes that are salient at the moment. Based on extensive social validation and/or continued use, some self-relevant attributes form and stabilize over years of personality development; these attributes are more likely to be invoked in the individual's self constructions and behavior across a variety of different contexts and over time, which makes them more typical for the individual (Higgins, 1987; Schlenker and Weigold, 1989; for a review see Baumeister, 1998; Kihlstrom and Klein, 1994). Other selfrelevant attributes are activated and become salient infrequently; such attributes are more atypical for the individual (Markus and Kunda, 1986; Nurius and Markus, 1990; for a review see Baumeister, 1998; Kihlstrom and Klein, 1994). On some occasions, predominantly typical self-relevant attributes are active in a person's working selfconcept (high typicality of the working self-concept). On other occasions, relatively few typical attributes and relatively more atypical attributes are salient in the working self-concept (high atypicality of the working self-concept). Once formed, a self-relevant attribute is stored in long-term memory from where the attribute can be retrieved into working memory (Baumeister, 1998; Kihlstrom and Klein, 1994; Markus and Wurf, 1987). Some self-relevant attributes are retrieved more easily than others; such attributes are said to be more accessible (Bargh, 1982). The accessibility of self-relevant attributes defines their activation readiness or ease of retrieval (Higgins, 1996). One determinant of the accessibility of self-relevant attributes is the frequency of their activation in the past: the more frequently a selfrelevant attribute has been activated in the past, the more accessible this attribute is likely to be (Andersen and Chen, 2002; Bargh et al., 1986; Higgins and Brendl, 1995; for a review see Higgins, 1996). Thus, typical self-relevant attributes are likely to be more accessible than atypical self-relevant attributes (Andersen and Chen, 2002; Bargh, 1982; Markus, 1977). For example, Jones et al. (1974) asked research participants to describe themselves and found that the most frequently listed self-relevant attributes were also more likely to be listed first, suggesting that the more typical self-relevant attributes were more readily accessible. 1.3. Impact of the environment on the self-concept Another source of accessibility for self-relevant attributes is cues in the environment (Kihlstrom and Klein, 1994; Markus and Kunda, 1986; Schlenker and Weigold, 1989). The environment can elicit selfrelevant attributes through generating experiences (Fazio et al., 1981; Markus and Kunda, 1986; Morse and Gergen, 1970) and/or prompting social roles (Forehand and Deshpande, 2001; Forehand et al., 2002; Hong et al., 2000). Because of the experiences people have in an environment and/or the social roles they perform in that environment, people may come to view themselves in a certain way; in other words, certain self-relevant attributes are likely to become more salient in the working self-concept. Often, previously formed self-relevant attributes are elicited; however, novel experiences and social roles may result in the formation of self-relevant attributes that are completely new for the individual (Markus and Wurf, 1987). In an experiment, Nurius and Markus (1990, Experiment 2) showed research participants false feedback (purportedly from other participants) that was either complimentary (social acceptance condition) or critical (social rejection condition). Then, participants described their self-concept at the time. Compared to the partici-

pants in the social acceptance condition, those in the social rejection condition listed fewer positive and more negative self-relevant attributes, and rated the negative attributes as more likely to characterize them in the future. In a study conducted in a retailing context, Chebat et al. (2006) found that upscale shopping mall environments elicit upscale self-relevant attributes. Since people are likely to view themselves as more upscale than they actually are, consumers feel higher self-congruity with upscale shopping malls and evaluate the stores located in such malls more positively (Chebat et al., 2006). Extraordinary retail environments are likely to create experiences consumers do not regularly have and/or to initiate social roles consumers do not normally engage in. Thus, through generating unusual experiences and/or roles, extraordinary retail environments are likely to momentarily activate more atypical self-relevant attributes in the working self-concept. For example, The Showcase of Nations at Disney's EPCOT Center in Orlando, Florida, features characteristic architectural styles, restaurants, and merchandise from different countries giving visitors an opportunity to experience life in the showcased countries. Therefore: Hypothesis 1. Retail environment extraordinariness increases the atypicality of the working self-concept. The next two hypotheses offer insight into the process of how retail environments varying in extraordinariness may activate selfrelevant attributes. The literature on accessibility suggests that the typical self-relevant attributes elicited by low extraordinariness environments are likely to have a very high level of accessibility, whereas the atypical self-relevant attributes elicited by high extraordinariness environments are not likely to have especially high levels of accessibility. Two sources impacting the accessibility of selfrelevant attributes are the frequency of past activation (a chronic source of accessibility) and the environment (a temporary source of accessibility). The two sources have an additive effect on accessibility, each increasing accessibility independently of the other (Bargh et al., 1986; Higgins and Brendl, 1995; Johar et al., 2003). Typical self-relevant attributes have a high level of chronic accessibility. Additionally, low extraordinariness environments are likely to activate typical self-relevant attributes, thus acting as a temporary source of accessibility for those attributes. Because chronic and temporary sources of accessibility have an additive effect on the accessibility of self-relevant attributes, the typical attributes elicited by a low extraordinariness environment are likely to be very highly accessible. Highly accessible self-relevant attributes are likely to come to mind first and therefore appear early in selfdescriptive inventories (see Higgins, 1996 for a review). In other words: Hypothesis 2a. Typical self-relevant attributes have output primacy in low extraordinariness retail environments. Extraordinary retail environments are likely to activate atypical self-relevant attributes, thus acting as a temporary source of accessibility for those attributes. However, atypical self-relevant attributes are likely to have low levels of chronic accessibility. Thus, the accessibility of the atypical self-relevant attributes elicited by extraordinary environments primarily arises from a single source (the environment), and therefore, such attributes are not likely to have especially high levels of accessibility. Formally: Hypothesis 2b. Atypical self-relevant attributes do not have output primacy in high extraordinariness retail environments. The next section describes two experiments testing the hypotheses.

V.D. Kaltcheva et al. / Journal of Business Research 64 (2011) 551–557

553

2. Experiment 1 2.1. Participants and procedure The participants were 60 undergraduate students from a Southeastern university who volunteered to participate in the experiment in return for extra-credit toward business courses. The experiment involved two sessions held two days apart. In the first session, the participants described their self-concept during a typical meal. In the second session, the same participants saw either an extraordinary or an ordinary restaurant ambience. The instructions asked the respondents to imagine themselves in the restaurant, describe their self-concept in the environment, and then complete a manipulation check. The study concluded with a hypothesis-guessing question, which asked the respondents to describe their beliefs about the objectives of the study. No participant was able to identify the hypotheses. 2.2. Experimental stimuli Two restaurant ambiences were downloaded from VisualStore's web-site (www.visualstore.com). An interior view from the Fantasea Reef restaurant in Atlantic City served as the high-extraordinariness environment, while an interior view from TOPZ in Sherman Oaks, California, was the low-extraordinariness environment. A pretest showed that the two restaurant ambiences significantly differed on extraordinariness (t(39) = 3.20, p b .01, 2-tailed, MEXT = 5.48 vs. MORD = 4.03). Each participant saw only one restaurant environment on an individual computer screen. The picture of the restaurant remained on the screen for three minutes. Participants could look at the screen as frequently as desired in the three minutes and start completing their questionnaire at any time. The names and locations of the restaurants were not revealed to the participants. No participant indicated that s/he was able to identify the restaurant. 2.3. Measure of self-concept atypicality The measure used to evaluate the atypicality of participants' working self-concept in the treatment environment involved comparing participants' self-concept in that environment to a baseline measure of their typical meal-related self-concept. Participants described their typical self-concept in the first session. In adjectives and/or short phrases, they responded to the question, “Who are you during a typical meal on an ordinary day?” In the second session, after viewing the treatment restaurant environment, the participants responded to the question, “Who are you in this context?” Aron et al. (1995) developed an index that measures the difference in an individual's working self-concept across two different occasions. This self-concept index was used to compare participants' working self-concept in the treatment environment to their typical self-concept. Two coders who were unfamiliar with the objectives of the study and the assignment of participants to conditions independently coded participants' self-descriptions into self-relevant attribute categories. (For example, “fast-paced” and “hurried” were classified into one attribute category “in a rush.”) First, each coder independently coded a subset of self-descriptions to determine a set of relevant attribute categories. Next, through discussion, the two coders established a common set of attribute categories that each coder then used independently to code all participants' selfdescriptions. Intercoder reliability was 88% (Perreault and Leigh, 1989). Disagreements were resolved through discussion. The index shown below was calculated for each participant: Total number of self‐descriptions in the participant’s two inventories −Number of self‐descriptions common across the participant ’s two inventories Total number of self‐descriptions in the participant’s two inventories

Self-descriptions were considered common across a participant's two inventories if the self-descriptions were classified into the same attribute category. Index scores ranged from 0 to 1. A higher score indicated a greater difference between participants' working self-concept in the treatment environment and their typical self-concept. 2.4. Manipulation check The effectiveness of the extraordinariness manipulation was evaluated with two seven-point Likert-scaled items: “Being in this restaurant makes me feel as if I were out of the ordinary world,” and “Being in this restaurant makes me feel as if I were in another world.” The correlation between the two items was .68 (p b .001, n = 60). Thus, the two items were combined into a single scale by averaging participants' ratings. The combined scale had a mean of 4.71 and a standard deviation of 1.81. The extraordinariness manipulation was successful. The participants who saw the high-extraordinariness restaurant rated the environment as significantly more extraordinary than the participants who saw the lowextraordinariness restaurant (t(59) = 6.58, p b .001, 2-tailed, MEXT = 5.88 vs. MORD = 3.53). 3. Results Hypothesis 1, which proposes that retail environment extraordinariness will increase the atypicality of the working self-concept, is supported. The mean value of the self-concept index is significantly higher in the high-extraordinariness condition than in the lowextraordinariness condition (t(59) = 3.21, p b .005, 1-tailed, MEXT = .71 vs. MORD = .54) suggesting that, compared to the participants in the ordinary environment, the working self-concept of the participants in the extraordinary environment is significantly more dissimilar from their typical self-concept (see Fig. 1). Table 1 shows the self-relevant attributes most frequently listed in each of the three contexts: (1) the typical self-concept, (2) the low-extraordinariness restaurant environ-

ment, and (3) the high-extraordinariness restaurant environment. The pattern of results is consistent with Hypothesis 1. The same attribute categories emerge with high frequency in the typical self-concept and in the low-extraordinariness environment. In contrast, four of the seven self-relevant attributes emerging with high frequency in the highextraordinariness environment (amazement, curiosity/exploration, self-indulgence, and interest) do not appear frequently in the typical self-concept. Thus, the content analysis of participants' self-relevant attributes lends further support to Hypothesis 1. Hypothesis 2a proposes that typical self-relevant attributes will be more likely to be listed first in low extraordinariness environments, while Hypothesis 2b argues that output primacy is not likely for atypical self-relevant attributes in high extraordinariness environments. Thus,

554

V.D. Kaltcheva et al. / Journal of Business Research 64 (2011) 551–557 Table 2 Typical and atypical self-relevant attributes by order and environment extraordinariness: Experiment 1. Order

1 and 2 3 and 4 5 and 6

Fig. 1. The effect of environment extraordinariness on self-concept atypicality: Experiment 1.

the next analysis examines the order in which the typical and atypical self-relevant attributes were listed in each extraordinariness condition. Typical self-relevant attributes are defined as the attributes appearing both in a participant's typical self-concept (Session 1) and in the same participant's treatment-elicited self-concept (Session 2). Self-relevant attributes appearing only in the participant's treatment-elicited selfconcept (Session 2) are considered atypical. Table 2 shows the percentage of the typical and atypical selfrelevant attributes listed at each order level in the two extraordinariness conditions. Each two adjacent order levels (1 and 2, 3 and 4, 5 and 6) are combined in order to limit the number of pairwise comparisons. Because fewer than 50% of the participants listed six or more self-descriptions, participants' self-descriptions are analyzed up to order level 6. The percentage of the atypical self-relevant attributes is 100% minus the percentage of the typical self-relevant attributes. For example, in the low-extraordinariness condition, 58.3% of the attributes listed first or second in participants' self-description inventories are typical self-relevant attributes, while 41.7% (100%– 58.3%) are atypical self-relevant attributes. In the low-extraordinariness condition, the typical self-relevant attributes are more likely to appear earlier. Significantly more typical self-relevant attributes are listed at the 1–2 level than at the 5–6 level (Z = 3.02, p b .005, 1-tailed, Table 1 Participant's self-relevant attributes in Experiment 1 (restaurants). Self-relevant attribute Typical self-concept (related to meals) 1) Relaxation 2) Hungry 3) In a rush 4) Excitement/fun 5) Happiness 6) Comfort/contentment 7) Friendly/talkative Low-extraordinariness environment 1) In a rush 2) Excitement/fun 3) Friendly/talkative 4) Happiness 5) Hungry 6) Comfort/contentment 7) Relaxation High-extraordinariness environment 1) Excitement/fun 2) Happiness 3) Amazement 4) Curiosity/exploration 5) Friendly/talkative 6) Self-indulgence 7) Interest

Frequency n (%) 59 46 39 37 36 33 27

(11.9%) (9.2%) (7.8%) (7.4%) (7.2%) (6.6%) (5.4%)

36 32 21 20 18 16 14

(14.0%) (12.4%) (8.1%) (7.8%) (7.0%) (6.2%) (5.4%)

34 27 26 19 15 14 14

(13.7%) (10.8%) (10.4%) (7.6%) (6.0%) (5.6%) (5.6%)

Note: Self-relevant attributes listed less than 5% of the time in a condition are not shown in the table.

Low extraordinariness

High extraordinariness

Typical

Atypical

Typical

Atypical

58.3 50.0 31.6

41.7 50.0 68.4

35.0 25.4 35.8

65.0 74.6 64.2

58.3% vs. 31.6%). Also, significantly more typical self-relevant attributes are listed at the 3–4 level than at the 5–6 level (Z = 2.07, p b .05, 1-tailed, 50.0% vs. 31.6%). The percentage of typical selfrelevant attributes is not significantly different between the 1–2 and 3–4 levels (Z = .92, p N .10, 1-tailed, 58.3% vs. 50.0%). Thus, Hypothesis 2a is supported. Hypothesis 2b also receives support. No significant differences are found in the percentage of atypical self-relevant attributes among any of the order levels in the high-extraordinariness condition (Z ≤ 1.20, p N .10, 1-tailed). The next experiment replicates Experiment 1 in a different type of retail environment (stores). 4. Experiment 2 4.1. Participants, procedure, and measures The participants in Experiment 2 were 40 undergraduate students who were recruited from the same population and for the same compensation as the participants in Experiment 1. The procedure and the measures (reworded for a store environment) were the same as in Experiment 1 with one exception. In Session 2, after viewing the store environment, participants wrote a story about themselves in the store. As in Experiment 1, participants responded to a hypothesisguessing question at the end. No participant identified the hypotheses or the stores used as stimuli. 4.2. Experimental stimuli Two store ambiences were downloaded from VisualStore's website (www.visualstore.com). An interior view from a Discovery Channel Destination store in Washington, D.C., was selected as the high-extraordinariness environment, while an interior view from a Liverpool store in Cancun, Mexico, was the low-extraordinariness environment. A pretest showed that the two store ambiences significantly differed on extraordinariness (t(39) = 7.72, p b .001, 2tailed, MEXT = 5.25 vs. MORD = 1.88). The manipulation was administered in exactly the same way as in Experiment 1. 4.3. Manipulation check The correlation between the two manipulation check items was .80 (p b .001, n = 40). Thus, a combined scale was created by averaging the two items. The combined scale had a mean of 4.25 and a standard deviation of 1.75. The extraordinariness manipulation was successful. The participants in the high-extraordinariness condition rated the shopping environment as significantly more extraordinary than the participants in the low-extraordinariness condition (t(39) = 5.59, p b .001, 2-tailed, MEXT = 5.47 vs. MORD = 3.14). 5. Results Hypothesis 1 is supported. Compared to the participants in the low-extraordinariness condition, the mean value of the self-concept index is significantly higher for the participants in the highextraordinariness condition, suggesting that their working selfconcept in the shopping environment is significantly more different from their typical shopping self-concept (t(39) = 4.79, p b .001, 1-

V.D. Kaltcheva et al. / Journal of Business Research 64 (2011) 551–557

555

Table 4 Typical and atypical self-relevant attributes by order and environment extraordinariness: Experiment 2. Order

1 and 2 3 and 4 5 and 6

Fig. 2. The effect of environment extraordinariness on self-concept atypicality: Experiment 2.

tailed, MEXT = .77 vs. MORD = .43) (see Fig. 2). As in Experiment 1, two coders independently coded participants' self-descriptions into selfrelevant attribute categories. The coding procedure was the same as in Experiment 1. Intercoder reliability was 86% (Perreault and Leigh, 1989). Disagreements between the coders were resolved through discussion. Table 3 shows the self-relevant attributes most frequently listed in each of the three contexts. All five self-relevant attributes appearing with high frequency in the low-extraordinariness environment are listed frequently also as part of the typical self-concept. On the other hand, three of the five self-relevant attributes emerging with high frequency in the high-extraordinariness environment (curiosity/exploration, amazement, and interest) do not appear frequently in the typical self-concept. The next section discusses Hypotheses 2a and 2b. Table 4 shows the percentage of the typical and atypical self-relevant attributes listed at each order level in the two extraordinariness conditions. As in Experiment 1, each two adjacent order levels are combined and participants' self-descriptions are analyzed up to order level 6. As in Experiment 1, Hypothesis 2a receives support. In the low-extraordinariness condition, the percentage of typical self-relevant attributes is significantly higher at the 1–2 level than at the 5–6 level (Z = 1.94, p b .05, 1–tailed, 73.8% vs. 50.0%). The percentage of typical selfrelevant attributes does not differ significantly between the 1–2 and the 3–4 levels (Z = 1.26, p N .10, 1-tailed, 73.8% vs. 61.0%), and between the 3–4 and the 5–6 levels (Z = .86, p N .10, 1-tailed, 61.0%

Table 3 Participant's self-relevant attributes in Experiment 2 (stores). Self-relevant attribute Typical self-concept (related to shopping) 1) Excitement/fun 2) Relaxation 3) Organization/efficiency 4) Happiness 5) Price consciousness 6) In a rush Low-extraordinariness environment 1) Excitement/fun 2) Relaxation 3) Happiness 4) Organization/efficiency 5) Price consciousness High-extraordinariness environment 1) Relaxation 2) Excitement/fun 3) Curiosity/exploration 4) Amazement 5) Interest

Frequency n (%) 36 30 29 24 24 14

(13.1%) (11.0%) (10.6%) (8.8%) (8.8%) (5.1%)

18 15 14 14 6

(16.4%) (13.6%) (12.7%) (12.7%) (5.5%)

19 12 11 10 7

(16.2%) (10.3%) (9.4%) (8.6%) (6.0%)

Note: Self-relevant attributes listed less than 5% of the time in a condition are not shown in the table.

Low extraordinariness

High extraordinariness

Typical

Atypical

Typical

Atypical

73.8 61.0 50.0

26.2 39.0 50.0

23.7 30.6 21.4

76.3 69.4 78.6

vs. 50.0%). Hypothesis 2b also receives support. No significant differences emerge in the percentage of atypical self-relevant attributes among any of the order levels in the high-extraordinariness condition (Z ≤ .84, p N .10, 1–tailed). Empirical evidence in the retail literature shows that the selfconcept is a major determinant of motivation and behavior (Chebat et al., 2006; Sirgy et al., 2000). Thus, this article does not advance a formal hypothesis regarding the relationship between the selfconcept and behavior. The focus of this research is on investigating the causal relationship between environment extraordinariness and the self-concept. However, the relationship between the self-concept and behavior is replicated in Experiment 2. Two independent coders classified the behaviors described in participants' stories. The coders were provided with a tentative list of classification categories representing behaviors of interest, such as exploration (exploring the store's merchandise and/or ambience). The coders were allowed to create and use their own categories as well. On average, participants listed 1.80 behaviors. Intercoder reliability was 80% (Perreault and Leigh, 1989). Disagreements between the coders were resolved through discussion. Significant differences between the high- and low-extraordinariness environments are found for three behavioral categories: (1) price consciousness (the participant takes into account financial constraints while shopping); (2) self-indulgence (the participant considers selfindulgent activities such as going on vacation, escaping, etc.); and (3) exploration (the participant is exploring the store's merchandise and/or ambience). While 18% of the participants in the low-extraordinariness condition consider financial constraints in their stories, only 5% of those in the high-extraordinariness condition mention financial considerations (Z = 1.66, p b .05, 1-tailed, 18% vs. 5%). Compared to the participants in the low-extraordinariness condition, a higher percentage of the participants in the high-extraordinariness condition mention selfindulgence activities (Z = 2.11, p b .02, 1-tailed, 11% vs. 0%) and describe exploring the store's merchandise and/or ambience (Z = 1.61, p = .05, 1-tailed, 26% vs. 12%). These results are consistent with participants' self-concept (see Table 3). Specifically, in their self-descriptions, participants mention price consciousness with high frequency in the ordinary store but not in the extraordinary store. Thus, participants' working self-concept is consistent with the heightened price consciousness and decreased self-indulgence observed in the ordinary store. Consistent with the higher frequency of exploration behaviors found in the extraordinary store, the self-relevant attributes of curiosity/ exploration, amazement, and interest emerge with high frequency in the high-extraordinariness not in the low-extraordinariness store. 6. Discussion 6.1. Theoretical and managerial implications The first objective of this research is to examine whether extraordinary retail environments elicit atypical self-concepts. Two between-subjects experiments manipulate the extraordinariness of the retail environment and evaluate participants' self-concept atypicality. In order to reduce the likelihood of demand effects, both experiments use a free-response measure of self-concept atypicality. The two experiments find consistent results. Compared to low-

556

V.D. Kaltcheva et al. / Journal of Business Research 64 (2011) 551–557

extraordinariness retail environments, high-extraordinariness retail environments elicit a more atypical self-concept. Content analysis of participants' self-concept in the two environments lends further support to the hypothesis. This research also offers insight into how retail environments varying in extraordinariness may activate self-relevant attributes. The results show that the typical self-relevant attributes elicited by low extraordinariness environments are likely to have a very high level of accessibility, which is explained with the additive effect of two sources of accessibility (temporary and chronic). In contrast, the accessibility of the atypical self-relevant attributes elicited by extraordinary environments stems primarily from a single source— the environment. These findings have important implications for retailers. Extraordinary retail environments must make a powerful impact on the senses in order to elicit atypical working self-concepts. Retailers should take care to eliminate any cues that may remind customers of the everyday reality because such cues are likely to activate typical self-relevant attributes. For example, the use of imitation construction materials (such as a linoleum imitation of hardwood floors) is likely to undermine the realism of the retail environment and allow customers to recall the everyday reality. Accordingly, interpretive research suggests that consumers want extraordinary retail environments to look and feel “more real than the real world” (Firat and Venkatesh, 1995, p. 252; see also Belk, 2000; Kozinets et al., 2004). Extraordinary retail environments can be very effective in encouraging customers to depart from their habitual behavior. For example, an extraordinary restaurant ambience may encourage customers to sample exotic or novel dishes. Extraordinary environments in stores that sell home accessories, such as Pier One Imports and Michaels, may make consumers more open to new home decor ideas. On the other hand, a retailer may wish to reinforce customers' habitual behavior. For example, an upscale clothing store may want to reinforce customers' preferences for high quality and elegance. In such cases, the store should create a typical upscale ambience consistent with customers' expectations. Market researchers investigating customers' self-concept in retail environments can use the self-concept measure described in this article. For example, retailers redesigning their venues can implement this self-concept measure to create ambiences that are likely to increase desired customer behaviors. Consider a retailer who is developing a new store concept. The retailer can show one or more different store concepts to the respondents and evaluate the respondents' self-concept in each store. Thus, the management will be able to identify how each of the alternative store concepts is impacting customers' self-concept. 6.2. Limitations A limitation of this research is that the pictures of the retail ambiences used to manipulate environment extraordinariness differed with respect to many aspects of the environment (such as layout, textures, colors, fixtures, etc.). Thus, one or more of the environment features that were not controlled for could have potentially produced or contributed to the results. Another limitation of this research is related to how participants' self-descriptive items were coded into self-relevant attribute categories. The coders used their own interpretation of participants' self-descriptions and the conventional connotation of the self-descriptive words. As a result, the coders may have inaccurately classified some of the participants' self-descriptions. 6.3. Future research directions Compared to atypical working self-concepts, typical working selfconcepts are likely to be richer and more complex cognitive

structures, encompassing more memories, experiences, and behaviors (Higgins, 1987; Schlenker and Weigold, 1989). The activation of typical working self-concepts, therefore, is likely to increase elaboration and the use of central (as opposed to peripheral) processing of new information (Burnkrant and Unnava, 1995; Meyers-Levy and Peracchio, 1996). A potentially promising avenue for future research would be to test whether consumers are more likely to engage in central processing of product and other information in ordinary retail environments (which activate more typical working self-concepts) and more likely to engage in peripheral information processing in extraordinary retail environments (which activate more atypical working self-concepts). Another avenue for further study would be to examine the potentially interactive effect between extraordinary brick-and-mortar retail ambiences and retailer websites. Consider a consumer who visits an extraordinary brick-and-mortar store and then visits the retailer's website. Would the memory of the in-store experience influence the consumer's working self-concept while shopping on the retailer's website? What are the conditions that may strengthen or weaken such an effect? For example, to what extent may the extraordinariness of the retailer's website facilitate or inhibit the impact of the customer's prior in-store experience? A related issue for investigation would be whether extraordinary retailer websites can have the same impact on consumer working self-concept as extraordinary brick-and-mortar retail ambiences.

References Andersen SM, Chen S. The relational self: an interpersonal social-cognitive theory. Psychol Rev 2002;109(4):619–45. Aron A, Paris M, Aron EN. Falling in love: prospective studies of self-concept change. J Pers Soc Psychol 1995;69(6):1102–12. Babin BJ, Babin L. Seeking something different? A model of schema typicality, consumer affect, purchase intentions and perceived shopping value. J Bus Res 2001;54(2): 89–96. Bargh JA. Attention and automaticity in the processing of self-relevant information. J Pers Soc Psychol 1982;43(3):425–36. Bargh JA, Bond RN, Lombardi WJ, Tota ME. The additive nature of chronic and temporary sources of construct accessibility. J Pers Soc Psychol 1986;50(5): 869–78. Baumeister RF. The self. In: Gilbert DT, Fiske ST, Lindzey G, editors. The Handbook of Social Psychology, Vol 1. New York: McGraw-Hill; 1998, p. 680–740. Belk RW. May the farce be with you: On Las Vegas and consumer infantilization. Consumption, Markets and Culture 2000;4(2):101–24. Bellantonio J. Johnny Rockets retools, eyes franchises. Orange County Business Journal 2003;26(7):3, 47 [February 17-February 23]. Burnkrant RE, Unnava HR. Effects of self-referencing on persuasion. J Cons Res 1995;22(1):17–26. Chebat JC, Sirgy MJ, St-James V. Upscale image transfer from malls to stores: a selfimage congruence explanation. J Bus Res 2006;59(12):1288–96. Elan E. Mars 2112 attacks: ‘themer’ shoots for Red Planet style. Nation's Restaurant News 1998;32(50):16, 64 [December 14]. Fazio RH, Effrein EA, Falender VJ. Self-perceptions following social interaction. J Pers Soc Psychol 1981;41(2):232–42. Firat AF, Venkatesh A. Liberatory postmodernism and the reenchantment of consumption. J Cons Res 1995;22(3):239–67. Forehand MR, Deshpande R. What we see makes us who we are: priming ethnic selfawareness and advertising response. J Marketing Res 2001;38(3):336–48. Forehand MR, Deshpande R, Reed A. Identity salience and the influence of differential activation of the social self-schema on advertising response. J Appl Psychol 2002;87 (6):1086–99. Higgins ET. Self-discrepancy: a theory relating self and affect. Psychol Rev 1987;94(3): 319–40. Higgins ET. Knowledge activation: accessibility, applicability, and salience. In: Higgins ET, Kruglanski AW, editors. Social Psychology: Handbook of Basic Principles. New York: Guilford Press; 1996. p. 133–68. Higgins ET, Brendl CM. Accessibility and applicability: some ‘activation rules’ influencing judgment. J Exp Soc Psychol 1995;31(3):218–43. Hong Y, Morris MW, Chiu C, Benet-Martinez V. Multicultural minds: a dynamic constructivist approach to culture and cognition. Am Psychol 2000;55(7):709–20. Johar GV, Moreau P, Schwarz N. Gender typed advertisements and impression formation: the role of chronic and temporary accessibility. J Cons Psych 2003;13(3):220–9. Johnson R. Disney's Epcot Center in Orlando, Fla., may showcase Spain at pavilion. Knight Ridder Tribune Business News 2002:1 [September 28]. Jones RA, Sensenig J, Haley JV. Self-descriptions: configurations of content and order effects. J Pers Soc Psychol 1974;30(1):36–45.

V.D. Kaltcheva et al. / Journal of Business Research 64 (2011) 551–557 Kihlstrom JF, Klein SB. The self as a knowledge structure. In: Wyer RS, Srull TK, editors. Handbook of Social Cognition: Basic Processes, Vol. 1. Hillsdale, NJ: Erlbaum; 1994. p. 153–208. Kozinets RV, Sherry JF, Storm D, Duhacek A, Nuttavuthisit K, Deberry-Spence B. Ludic agency and retail spectacle. J Cons Res 2004;31(3):658–72. Markus H. Self-schemata and processing information about the self. J Pers Soc Psychol 1977;35(2):63–78. Markus H, Wurf E. The dynamic self-concept: a social psychological perspective. In: Rosenzweig MR, Porter LW, editors. Annual Review of Psychology, Vol. 38. Palo Alto, CA: Annual Reviews Inc.; 1987. p. 299–337. Markus H, Kunda Z. Stability and malleability of the self-concept. J Pers Soc Psychol 1986;51(4):858–66. Meyers-Levy J, Peracchio LA. Moderators of the impact of self-reference on persuasion. J Cons Res 1996;22(4):408–23. Morse S, Gergen KJ. Social comparison, self-consistency, and the concept of self. J Pers Soc Psychol 1970;16(1):148–56.

557

Niedenthal PM, Beike DR. Interrelated and isolated self-concepts. Pers Soc Psychol Rev 1997;1(2):106–28. Nurius PS, Markus H. Situational variability in the self-concept: appraisals, expectancies, and asymmetries. J Soc Clin Psych 1990;9(3):316–33. Perreault WD, Leigh LE. Reliability of nominal data based on qualitative judgments. J Marketing Res 1989;26(2):135–48. Rhodewalt FT. Self-presentation and the phenomenal self: on the stability and malleability of self-conceptions. In: Baumeister RF, editor. Public Self and Private Self. New York: Springer; 1986. p. 117–42. Schlenker BR, Weigold MF. Goals and the self-identification process: constructing desired identities. In: Pervin LA, editor. Goal Concepts in Personality and Social Psychology. Hillsdale, NJ: Erlbaum; 1989. p. 243–90. Sirgy MJ. Self-concept in consumer behavior: a critical review. J Cons Res 1982;9(3): 287–300. Sirgy MJ, Grewal D, Mangleburg T. Retail environment, self-congruity, and retail patronage: an integrative model and a research agenda. J Bus Res 2000;49(2):127–38.