The income effect in personal shopping value, consumer self-confidence, and information sharing (word of mouth communication) research.

Paridon, Terrence J. ; Carraher, Shawn ; Carraher, Sarah C. 等

ABSTRACT

Theory predicts that income should influence a complex set of relationships involving personal shopping values, consumer self-confidence, and word of mouth communication. Nevertheless, irrespective of income level, findings indicate that social self-confidence mediates the effects of hedonic experiences upon word of mouth communication. Similarly, the direct effects of utilitarian value upon personal self-confidence are also invariant across income groups. Findings involving the effects of personal confidence upon word of mouth communication are inconclusive. However, income does influence the nature of the relationships between hedonic experiences and social self-confidence as well as the significance of the effects involving utilitarian value and social self-confidence. The implications for these findings are discussed from the perspectives of retail management, site location decisions, and the direction of future research efforts.

THE INCOME EFFECT IN PERSONAL SHOPPING VALUE, CONSUMER SELF-CONFIDENCE AND INFORMATION SHARING RESEARCH

The marketing discipline's sustained stream of research into the nature of hedonic and utilitarian constructs has produced a significant body of literature that has enhanced knowledge about effective marketing practices. For example, Kozinets et al. (2002) emphasized that retail environments should incorporate hedonic and epistemic (utilitarian) designs (cf. Titus & Everett, 1995) when formulating retail themes. Such recommendations rested on solid empirical grounds. To be more specific, research indicates that the level of the shopping environment's hedonic value contributed significantly to explaining differences in shopper's reactions to mall and store atmospheres (Babin, Darden & Griffin, 1994; Michon, Chebalt & Turley, 2005; Wakefield & Baker, 1998). Similarly, findings indicate that utilitarian value is a factor in clarifying the nature of purchasing behavior (Babin & Attaway, 2000; Babin, Darden & Griffin, 1994).

The effects of hedonic and utilitarian experiences upon shopper's responses to marketing activities extend beyond immediate reactions to the shopping milieu. That is, hedonic experiences and utilitarian outcomes have been researched as antecedent variables in word of mouth communication. A significant causal relationship between hedonic atmosphere and patronage intentions included the shopper's willingness to recommend the store to a friend (Grewal, Baker, Levy, &Voss, 2003). Other research (Paridon, 2005a, 2005b) suggests the viability of a model in which utilitarian value directly affects personal outcomes confidence while hedonic value influences social outcomes confidence. However, only social outcomes confidence has been found to mediate the effects of hedonic value upon information sharing.

The aforementioned contributions to the discipline's understanding of interpersonal influence notwithstanding, research suggest the possibility of extending the significance of those initial results involving word of mouth communication. To be more specific, interest in the potential for demographics to moderate findings in the area of consumption and interpersonal influence is not unknown. For example, Holbrook and Hirschman (1982) postulated an income effect and an income difference was expected but not confirmed in research on hedonic shopping motivations (Arnold & Reynolds, 2003). Other research has hypothesized and confirmed the interaction effects of hedonic/functional usage, social/personal consumption context, and income upon purchasing behavior (Wakefield & Inman, 2003). Since purchasing behavior frequently precedes word of mouth, this study examines income as a possible moderator in research involving personal shopping value, consumer self-confidence, and word of mouth communication.

CONCEPTUAL DEVELOPMENT

Building upon a number of studies about the availability of product information, merchandising practices, and store design, Titus and Everett (1995) proposed the consumer retail search process model. It postulates that the search for product information may be guided by epistemic and hedonic constructual systems. The epistemic system, representing the shopper's system of logic, gives rise to a need for the design of in-store information displays and merchandising practices, the sine qua non of the epistemic system. In other words, such displays and practices enable the shopper to focus upon utilitarian needs. In turn, fulfillment of these task oriented needs leads to a satisfactory utilitarian experience, an experience that has been defined as the " ... work ... " (Babin, Darden, and Griffin, 1994, p. 646) associated with shopping. The emotionally laden hedonic system is the sensate orientation that accompanies the shopping experience. This sensate orientation enables the shopper to experience hedonic value, which is defined as the emotional response attributable to the " ... potential entertainment and emotional worth ... " of the shopping event (Babin, Darden, and Griffin, 1994, p. 646). The retail search model suggests that the individual and the combined effects of these two constructual systems may lead to an efficient, pleasurable, and satisfying shopping experience.

Mall and store research supports the aforementioned hedonic and epistemic (utilitarian) postulates of the model. In a mall study, a sensate environmental factor was causally related to the excitement associated with the shopping trip and a desire to continue shopping (Wakefield & Baker, 1998). Similarly, olfactory impressions generated a favorable response to the mall environment as well as product related effects (Mattila & Wirtz, 2001; Michon, Chebat & Turley, 2005). In store research, a positive emotional state explained one's satisfaction with the shopping experience (Babin & Darden, 1996). In other store research, shopper's hedonic reaction to, and utilitarian orientation towards, the shopping experience were associated with their overall satisfaction with the marketplace offerings (Babin, Darden & Griffin, 1994; Griffin, Babin, & Modianos, 2000). In other mall research, Shim and Eastlick (1998) report a composite measure of knowledge about the attributes of one's most frequently shopped mall, arguably utilitarian based. Emotional variables contributed to one's overall frequency of shopping and average monthly mall expenditures. Knowledge based and emotional variables contributed also to an increase in the amount of time and money spent in shopping (Babin, Griffin & Boles, 1997).

Efficient, pleasurable, and satisfying shopping experiences are thought also to contribute to the consumer's personal and social confidence in making decisions (Bearden, Hardesty & Rose, 2001). According to the authors, consumer self-confidence is the " ... extent to which an individual feels capable and assured with respect to his or her marketplace decisions and behaviors" (p. 122). Since the decision to shop involves explicit marketplace behaviors, one's hedonic and utilitarian shopping experiences should influence the consumer's personal and social self-confidence. Stated somewhat differently, a shopper who feels capable and assured in her shopping experiences will experience a minimal level of doubt about her consumption related social and personal self-confidence (cf. Folkes & Kiesler, 1991). Thus, the first research hypothesis is:

H1: Hedonic and utilitarian shopping experiences will positively influence personal and social self-confidence.

Consumer Self-Confidence

When consumers interact successfully with shopping environments, their acquisition of personal and social information is grounded in their hedonic experiences and their utilitarian outcomes. In turn, this information, in the form of word of mouth communication (Feick, Price & Higie, 1986; Paridon, 2005a; Reynolds & Darden, 1971; Summers, 1970), is passed along to their friends and acquaintances (Titus & Everett, 1995; see also Higie, Feick & Price, 1987). Nevertheless, this transmission should not be thought of as automatic. That is, since the personal and social nature of the shopping experience (Tauber, 1972) leads to the acquisition of information that will influence one's personal outcomes and social outcomes self-confidence (Bearden et al., 2001), and since self-confidence influences word of mouth communication (Reynolds & Darden, 1971; Summers, 1970), one's personal outcomes and social outcomes self-confidence should influence one's word of mouth communication. Accordingly, the following research hypotheses appear tenable:

H2: Word of mouth communication will be positively influenced by consumer social outcomes and personal outcomes confidence.

H3: Hedonic value and utilitarian value operating through social outcomes confidence and personal outcomes confidence will positively effect word of mouth communication.

Income Effect

Theorists and researchers have postulated a differential income effect involving hedonic and functional consumption and usage. For example, Holbrook and Hirschman (1982) recognized the possibility that income might contribute to explaining differences in hedonic consumption. Similarly, in discussing hedonic consumption situations, Caldwell (2001) and Richards (1999) emphasize the potential for an income effect. Support for this effect exists. To be more specific, Blanchflower and Oswald (2004) report that income influences one's happiness in life, with happiness generally considered one indicator of an orientation towards hedonic consumption (Batra & Ahtola, 1990; Voss, Spangenberg, and Grohmann, 2003).

The relationship between income and utilitarian orientation is not so straightforward. For example, in discussing the complex nature of relationships between income and experiential as well as utilitarian orientations, Okada (2005, p. 50) states, " ... it might be predicted that financial constraints will increase the need for justification of choice and therefore amplify the magnitude of the reversal in relative preferences between hedonic and utilitarian alternatives." Specifically, any effect of income upon hedonic and utilitarian choices might be subject to the purchase consumption situation. Direct evidence for such a relationship exists. Wakefield and Inman (2003) hypothesized a hedonic/functional usage, a consumption related social and personal self-confidence, and an income effect upon purchasing behavior. In discussing their findings, the authors report that the social/personal consumption context, the intended usage--hedonic or functional, one's income, and the quantity purchased interact in contributing to understanding differences in the price's shopper's pay. An income effect was associated with one's intended usage--hedonic/functional--as well as the social personal context of the consumption situation. When income was taken into consideration, the causal nature of the findings was altered.

While the preceding research involving an income effect is conceptually significant, the findings do not enable the formulation of specific income effect hypotheses in this study. Nevertheless, the research indicates that income has the potential to contribute to understanding shopping behavior. This potential to contribute appears to be related to income's role as a moderator or " ... a qualitative (e.g. sex, race, class) or quantitative (e.g. level of reward) variable that affects the direction and/or the strength of the relation between an independent or predictor variable and a dependent or criterion variable" (Baron & Kenny, 1986, p. 1174). Accordingly, the final research hypothesis is:

H4: Income will moderate the effects of personal shopping value and consumer self-confidence in word of mouth communication research.

METHODOLOGY

Data Collection

Marketing students enrolled in an advanced marketing course at a southwestern regional university were trained and given instructions in personal interviewing procedures and techniques. Since female shoppers account for approximately 80 % of total retail expenditures, women were the subjects in the study. In two surveys spaced approximately six months apart, students contacted their adult non-student female friends and acquaintances and asked them to complete a self-administered structured questionnaire about their department store shopping behaviors. Respondents were asked also to provide a first name and telephone number for verification purposes. Students were instructed to remain content neutral and answer only questions about the instructions for completing the questionnaire.

For each survey, a quota-sampling plan was adopted in an attempt to obtain a representative demographic cross section of respondents (cf. Bearden, Hardesty, & Rose, 2001). An analysis of the demographic data for the 458 respondents, 258 from the first study and 200 from the second study, revealed that the typical respondent is white, has at least one child, is employed at least part time, and is thirty-nine years old. She resides in a household with an annual average income of $40,000.00 dollars and has attained at least a partial college education. A convenience sample of 60 respondents, two per student interviewer, was contacted by telephone and asked to verify their participation. All interviewees responded that they had completed the questionnaire.

Testing for an income effect required partitioning the sample of 458 into two or more groups. Since structural equation modeling was the preferred mode of analyzing the data, statistical considerations associated with the technique influenced the partition decision. When the general rule of thumb for a sample of size 200 for structural equation modeling is considered along with research suggesting that the minimum acceptable size to evaluate SEM fit statistics is 150 (Hu and Bentler, 1998), the respondent base of 458 could be partitioned into either two or three groups. Since three groups would involve a partition of approximately 150 per group, the bare minimum to assess the fit statistics, the decision was made to split the respondents into two groups.

A review of Census Bureau statistics for the MSA indicates that approximately fifty percent of the households reported an annual household income of less than thirty five thousand. Accordingly, this value was used as the partitioning point. The partitioned database consisted of 222 respondents with income less than thirty five thousand, and 236 with income equal to or greater than thirty five thousand.

Scale Indicators

Recent research in measuring latent constructs has documented acceptable values of reliability and validity coefficients for unipolar scales (Bearden et al., 2001; Clark & Goldsmith, 2005; Dabholkar, Thorpe & Rentz, 1996; Darden & Babin, 1994; Feick & Price, 1987; Kim & Jin, 2001). Preliminary research involving the measurement of information sharing was consistent with those findings: unipolar construct measurement generated comparable internal consistency and construct reliability values. Similarly, initial research into the viability of the proposed model indicated that improvements in construct reliability could be attained by using shortened unipolar measures (Paridon, 2005a; 2005b). The original sets of construct indicators for the model were modified accordingly. The first change was the deletion of the "gift giving," and the "agonizing over purchasing" measures from the original personal and social self-confidence scales, respectively (Bearden et al., 2001). Responses to the remaining four indicators for each self-confidence construct were obtained by asking interviewees to indicate on a seven point Likert-type scale the extent to which they agreed that the statement characterized them.

Similar issues involving reliability resulted in selecting four hedonic orientations, and four utilitarian value indicators (Babin et al., 1994; Griffin et al., 2000) that demonstrated, by their standardized factor loading, an acceptable level of construct validity. However, previous research by Paridon (2005a; 2005b) suggested that deletion of a negatively worded hedonic construct indicator would result in improved reliability and variance extracted values. Further research into utilitarian measurement (Voss, Spangenberg & Grohmann, 2003) suggested an improvement in the original four indicator utilitarian scale might be realized by incorporating satisfaction as a new measure. The decision to include satisfaction was guided in part by findings indicating that satisfaction was significantly and positively related to utilitarian value (Babin, Darden, & Griffin 1994). Other research suggests that satisfaction may also be a viable indicator when measuring market place experiences. Bagozzi, Gopinath, and Nyer (1999) state (p. 201), "We suspect that previous studies finding discriminant validity for measures of satisfaction can be explained by the way the items were presented on the questionnaire (e.g. separation of measures of satisfaction from measures of other positive emotions) or the lack of a sufficient number of positive emotions." A context effect (cf. Podsakoff, MacKenzie, & Lee, 2003) can be expected when using satisfaction as an indicator, and the findings of Babin, Darden, and Griffin (1994) confirm the potential for such an effect, thereby suggesting the viability of satisfaction as an indicator for utilitarian measurement.

One additional change was incorporated into the set of indicators used to measure utilitarian value. Research involving retail information sharing (Paridon 2005a; 2005b) indicated that deleting the negatively worded utilitarian statement "I couldn't buy what I really needed," a product related measure, could lead to improved measurement statistics. Thus, utilitarian value was assessed using two original utilitarian value indicators and two substitute measures of satisfaction and usefulness. For these eight personal shopping value measures, using a conventional seven point Likert type format, respondents were asked to indicate the extent to which they agreed with the item.

The final set of construct measures emerged from research in retail information sharing (Paridon, 2004), and word of mouth communication (Feick et al., 1986; Higie et al., 1987). Three of the ten indicators from the retail information sharing scale were adopted and supplemented with one item that explicitly addressed the altruistic nature of word of mouth communication (Feick et al., 1986; Higie et al., 1987; Paridon, 2006). For each of the four indicators, participants were asked to rate the extent to which they agreed, on a seven point Likert type scale, with the statement.

Although not developed as a conceptual complement to or as a theoretical foundation in this study, the questionnaire also contained, for comparison purposes, the complete market maven scale (Feick & Price, 1987), an accepted measure of one's propensity to engage in general marketplace conversations. In addition to the previously discussed, standard semantic anchors, the indicators for all five constructs and the market maven measures contained numeric anchors with one representing the least favorable interpretation of the statement and seven indicating the most favorable interpretation of the statement. Standard demographic measures--gender, age, marital status, employment status, annual household income, education, and ethnicity--were included also.

ANALYSIS AND RESULTS

As an initial analysis for evaluating multiple indicator scales, Hair, Anderson, Tatham, and Black (1998) recommend using the Bartlett test of sphericity and the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy. For each of the five constructs in both income groups, the significance level for the Bartlett test was less than .01. In the lower income (upper) group, the KMOs were as follows: hedonic orientation, .82 (.83); utilitarian value, .81 (.84); information sharing, .79 (.80); social confidence, .80 (.73); personal confidence, .75 (.69). (Comparable statistics characterize the original two data sets.) Sets of indicators whose KMO value is .80 or greater are characterized as meritorious while KMO values that exceed .70 are considered adequate.

For both income groups, the significance levels for the Bartlett tests for the market maven scale were less than .01. The KMO values were .82 and .83, lower income group and upper income group, respectively. Accordingly, an initial statistical comparison of the five constructs and the market maven scale made use of Pearson product moment correlations. The resulting pattern of coefficients, summarized in Table 1, suggests an acceptable level of agreement between the market mavenism and the constructs of interest in this study. One would expect that shoppers who share information about their shopping experiences; exhibit social confidence; enjoy shopping; and shop for what they want would also score high on the market maven scale. Furthermore, the correlation between personal confidence and market mavenism is in partial agreement with expectations: the negatively worded personal confidence should correlate negatively with market mavenism. This pattern holds for the lower income group but not the higher income group. The remaining correlations of Table 1 are offered as descriptive measures of the relationships between the five constructs of interest in this study.

Structural Equation Model Analysis

Non-nested model testing is appropriate for comparing covariance structure models across groups (Joreskog & Sorbom, 2001; Rust, Lee, & Valente, 1994). Accordingly, the conceptual model depicted in Figure 1 was subjected to LISREL 8 (Joreskog & Sorbom, 2001) group analyses.

[FIGURE 1 OMITTED]

Prior to testing the hypotheses, a SEM analysis focused upon comparing the two original data sets using the group analysis procedures advocated by Joreskog and Sorbom (2001). The analysis involved freely estimating the structural equation coefficients for each group. It generated an overall minimum fit function chi square of 918.32 with 326 degrees of freedom, p < .01, indicating a lack of exact fit. The goodness of fit statistics for this analysis were: a comparative fit index (CFI) .94; a root mean square error of approximation (RMSEA) of .09; and a standardized root mean square residual (SRMR) of .07. Comparing the observed CFI and SRMR values with findings that reasonable true (nonzero factor covariance), simple (indicators loading only on one factor), and robust (measured variables not multivariate normally distributed) models of sample sized 150 to 5,000 (True, simple, and robust characterizes all models and data sets used in this study.), should generate CFI values of .90 to .96 and SRMR values of .06 to .11 (Hu & Bentler 1999, p. 16), leads to accepting the model as reasonable. This conclusion is buttressed by Hu and Bentler's finding that RMSEAs and SRMRs of .09 and .07 for true, simple, and robust models lead to acceptable Type 1 error rates (1999, p. 26).

Accordingly, each structural equation coefficient was examined for invariance across the two survey groups. The procedure required a series of LISREL 8 SEM analyses that constrain each structural equation coefficient to be equal across the two groups (cf. Babin and Darden, 1995). The chi square value of the freely estimated analysis is then subtracted from the chi square value of the constrained analysis. If this chi square value with one degree of freedom is statistically significant, the groups differ with respect to the constrained coefficient. (Acceptable values of CFIs, RMSEAs, and SRMRs should also characterize the constrained analyses. As expected, each of the six analyses involved in this modeling generated CFIs, RMSEAs, and SRMRs of approximately .94, .09, and .07.) The chi square difference values are summarized in Table 3. It should be noted that the survey groups differ with respect to the effects of social confidence upon information sharing, and utilitarian value upon social confidence. Since no major changes in the macro environment occurred during the six month interval between administering the surveys, these differences are most likely attributable to variation between the two survey groups.

The modeling procedures for testing the hypotheses were similar. In the first LISREL 8 analysis of the income group data, the structural equation coefficients for both income groups were constrained to be equal. The minimum fit function chi square was 829.13 with 332 df, p < .01. In the second income group analysis, the six structural equation coefficients for both groups were freely estimated. This analysis produced a chi square value of 805.68 with 326 df, p < .01, with a CFI of .95, a RMSEA of .08, and a SRMR of .07. The chi square difference of 23.45 with 6 degrees of freedom is significant beyond .01. Since this chi square difference test indicated that income effects the magnitude of the structural equation coefficients, six additional structural equation models were analyzed in order to identify the invariant structural relationships.

As before, the analysis for income effects requires testing of chi square difference values derived from the constrained and the freely estimated structural models. However, since the measurement properties of the indicator variables influences the confidence that may be placed in a SEM analysis, an evaluation of these measurement properties prior to structural analysis appears warranted. The standardized factor structures and loadings of the freely estimated baseline model contained in Table 2 were used to evaluate the reliability and discriminant validity of the constructs. Using the commonly accepted standard that factor loadings of an indicator serve as a measure of its construct validity, an evaluation of the loadings in Table 2 suggests each indicator attains an acceptable (greater than or equal to .50) level of construct validity. Construct reliability estimates were calculated using standardized factor loadings (Fornell & Larcker, 1981). The reliability estimates (coefficient alpha in parentheses) for the lower/upper income groups are: hedonic shopping orientation, .86/87 (.85/.87); utilitarian value, .88/.91 (.88/.90); social outcome confidence, .86/.83 (.86/.82); personal outcome confidence, .90/.84 (.90/.85); and information sharing, .81/85 (.81/.84). For both income groups, the more conservative average variance extracted values of the five constructs exceeded the commonly accepted value of .50. The specific lower/upper income values are: hedonic shopping orientation, .60/.62; utilitarian value, .65/.71; social outcome confidence, .60/.55; personal outcome confidence, .69/.58; and information sharing, .52/.58.

The average variance extracted values were used to evaluate the discriminant validity of the constructs. The structural equation model test for discriminant validity requires the square of the construct correlations to be less than both variance extracted values (Fornell & Larcker, 1981). For each income group, the results of this comparison test indicated the existence of discriminant validity for all ten pairwise comparisons.

Structural Equation Effects

Hypotheses one and two require consideration of direct structural equation effects, while hypothesis three requires indirect hypothesis testing. These three hypotheses were tested by using the structural equation coefficient values generated by the freely estimated LISREL 8 income group analysis or the baseline model. Hypothesis four involved the aforementioned chi-square difference test for moderation. Testing the latter hypothesis required the use of the baseline model as well as six additional LISREL 8 models.

The direct standardized structural equation coefficients and the statistical significance of the LISREL 8 estimates obtained from the separate income groups in the aforementioned freely estimated baseline analysis are contained in Figure 1. Numerous similarities and differences across the income groups are evident. First, a hedonic orientation effect upon both social confidence and personal confidence characterizes the upper income group while the lower income group exhibits significant hedonic orientation and utilitarian value effects upon social confidence. Both income groups are characterized by a significant relationship between utilitarian value and personal outcomes confidence. The negative sign between the utilitarian and the personal outcomes constructs is consistent with the wording conventions of both constructs. Thus, research hypothesis one is supported in part.

Hypothesis two is supported in part also. The effects of social confidence upon information sharing are strong and significant for both income groups. However, the nonsignificant lower income effect of personal confidence upon information sharing does not support hypothesis two. For the upper income group, the magnitude of the effect of personal value upon information sharing, while statistically significant with a value of .10, may not contribute to the managerial importance of this specific relationship.

For research hypothesis three, only the indirect social confidence effects were confirmed. For the lower income group, the indirect effects of utilitarian value and hedonic orientation operating through social confidence upon information sharing behavior, based upon the direct maximum likelihood LISREL estimates and calculated in accordance with Bollen's (1987) guidelines, were .25 and .13 respectively, attaining statistically significant Sobel (1982, 1986) test statistics of 4.04 and 2.90, respectively (p < .01). For the upper income group, the indirect effect of one's hedonic orientation operating through social confidence upon information sharing, magnitude .24 and a Sobel statistic of 4.84, attained a similar level of significance.

Finally, for research hypothesis four, the results of the chi-square difference tests appear in Table 3. For each reported chi square difference statistic, the baseline model chi square value of 805.68 with 326 degrees of freedom was subtracted from the chi square value obtained when testing for equality of the specific structural equation coefficient. Each of the latter chi squares possessed 327 degrees of freedom and generated CFIs, RMSEAs, and SRMRs that did not differ by more than .01 from the baseline values of .95, .08, and .07, respectively. Although each model qualifies as a reasonable model (Hu & Bentler, 1998) the analysis indicates that income acts as a moderator for the effects of utilitarian value upon social confidence, and for the effects of hedonic value upon social confidence. All other tests for income as a moderator are not significant.

DISCUSSION

In addition to similarities across the income groups, the results confirm the existence of a differential income effect that influences or at the very least accompanies one's orientation towards shopping. The observed similarities and differences possess the potential to further the discipline's understanding of shopper's experiences.

The results indicate that both the lower and the upper income groups gain personal self-confidence when they are able to purchase the functional items they require. Both groups engage also in favorable word of mouth communication as a result of a positive hedonic experience that is mediated by one's favorable social self-confidence. In other words, irrespective of one's level of income, one's confidence in their ability to elicit favorable responses from their friends depends upon their enjoyment of the shopping event, and the effects of this pleasure induced social confidence will translate into an increase in information sharing behavior.

Despite these similarities, the income groups differ when considering the magnitude of the effects of hedonic orientation upon social confidence and utilitarian value upon social confidence. When compared to the lower income group, the upper income group's social confidence exhibits the effect of being influenced more by hedonic shopping experiences. On the other hand, when compared to the upper income group, the lower income group's social confidence is influenced by utilitarian value. The upper income group shows no such effect. This latter dependency in the lower income group involving social confidence and utilitarian value contributes to the emergence of word of mouth communication. Thus, in the lower income group, social self-confidence mediates the effects of hedonic and utilitarian shopping experiences upon word of mouth communication. For the upper income group, social confidence mediates the effect of hedonic experiences only upon word of mouth communication.

When customers' shopping and confidence profiles match the profiles of the shoppers in this study, managers should plan merchandising efforts in order to foster specific positive outcomes. First, managers should focus upon the utility needs of the shopper. This practice should lead to an increase in the personal outcomes confidence associated with shopping decisions for both income groups. In addition, when lower income group shoppers believe that their utility needs are met, their consumption related social confidence will increase and lead to an increase in favorable word of mouth communication. Second, when retailers manage shopping environments, they should focus also upon creating enjoyable shopping atmospheres. The pleasurable experiences associated with such environments will influence the social confidence of both groups and differentially influence the social confidence of the upper income group. Thus social confidence will mediate the effects of hedonic experiences upon word of mouth communication with the upper income group being more receptive to enhanced hedonic experiences.

The theory and findings apply also to site location decisions. For example, spatial competition is a conceptual foundation of site location decision making (Karande & Lombard, 2005) with two strategies predominating. A proximity strategy is one in which the site is selected in order to capitalize on the principal of cumulative attraction. The result is that when stores are concentrated, with higher store density one outcome of the proximity strategy, higher income shoppers are presumed to be more likely to engage in price comparison shopping. However, they are unwilling to engage in price comparison shopping at the expense of their shopping enjoyment (Marmorstein, Grewal, and Fishe, 1992). Since the findings of this study do not negate the latter findings, it is suggested that developers and designers should insure that merchandise price comparisons are possible for higher income shoppers but not ignore shopping enjoyment.

On the other hand, a distancing strategy geographically disperses stores in an attempt to gain a differential advantage (Karande & Lombard, 2005). The latter emphasizes accessibility and time spent shopping, correlates of a utilitarian shopping orientation (cf. Babin & Attaway, 2000). When this strategy targets lower income shoppers, the the utilitarian value generated by the shopping experience will influence social confidence and lead to word of mouth communication. However, since hedonic experiences also influence the social confidence of lower income shoppers, providing them with a "favorable" hedonic experience will also lead to increased social confidence and word of mouth communication. Developers should recognize these relationships and include the appropriate level of hedonic experience when planning and designing stores.

The preceding findings and recommendations notwithstanding, studies should continue into extraneous variable effects upon personal shopping value, consumer self-confidence, and word of mouth communication. One possible extension could involve large sample studies designed to study the potential moderating effects of social class on word of mouth communication. These additional studies should focus also on identifying and clarifying specific hedonic conditions (e.g. Turley & Milliman, 2000) and other antecedent variable relationships. Future research could investigate the effects of scents, lighting, layouts, and merchandising practices upon hedonic experiences and utilitarian behaviors. Continued study of these cause and effect relationships should further the discipline's understanding of the theoretical and the managerial significance of designing and managing shopping environments because the economically competitive advantages generated by optimal shopping environments can be substantial.

REFERENCES

Arnold, M. J., & Reynolds, K. E. (2003). Hedonic shopping motivations. Journal of Retailing, 79 (2), 77-95.

Babin, B. J., & Attaway, J. S. (2000). Atmospheric affect as a tool for creating value and gaining share of customer. Journal of Business Research, 49 (2), 91-99.

Babin, B. J. & Darden, W. R. (1994). Exploring the concept of affective quality: Expanding the concept of retail personality. Journal of Business Research, 29 (2), 101-109.

Babin, B. J. & Darden, W. J. (1995). Consumer self-regulation in a retail environment. Journal of Retailing, 71 (1), 47070.

Babin, B. J., & Darden, W. R. (1996). Good and bad shopping vibes: Spending and patronage satisfaction. Journal of Business Research, 35 (3), 201-206.

Babin, B. J., Darden, W. R., & Griffin, M. (1994). Work and/or fun: Measuring hedonic and utilitarian shopping value. Journal of Consumer Research, 20 (4), 644-656.

Babin, B. J., Griffin, M., & Boles, J. S. (1997). Two ways to keep your customers: An exploratory investigation of patronage loyalty. In W. M. Pride and G. T M Hult (Eds.), Enhancing knowledge development in marketing (Vol. 8, pp. 251-252), Chicago: American Marketing Association.

Bagozzi, R. P., Gopinath, M., & Nyer, U. (1999). The role of emotions in marketing. Journal of the Academy of Marketing Science, 27 (2), 184-2006.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51 (6), 1173-1182.

Batra, R. & Ahtola, O. T. (1990). Measuring the hedonic and utilitarian sources of consumer attitudes. Marketing Letters 2 (2), 159-170.

Bearden, W. O., Hardesty. D. M., & Rose. R. L. (2001). Consumer self-confidence: Refinements in conceptualization and measurement. Journal of Consumer Research, 28 (1), 121-134.

Blanchflower, D. G. & Oswald, A. J. (2004). Well-being over time in Britain and the USA. Journal of Public Economics, 88 (7/8), 1359-1386.

Bollen, K. A. (1987). Total, direct, and indirect effects in structural equation models. In C. C. Clogg, (Ed.), Sociological methodology: Vol. 17. Washington DC: American Sociological Association.

Caldwell, M. (2001). Applying general living systems theory to learn consumers' sense making in attending performing arts. Psychology and Marketing, 18 (5), 497-511.

Clark, R. A. & R. E. Goldsmith (2005). Market mavens: Psychological influences. Psychology and Marketing, 22 (4), 289-312.

Dabholkar, P. A., Thorpe, D. I., & Rentz, J. O. (1996). A measure of service quality for retail stores: Scale development and validation. Journal of the Academy of Marketing Science, 24 (1), 3-16.

Darden, W. R., & Babin, B. J. (1994). Exploring the concept of affective quality: Expanding the concept of retail personality. Journal of Business Research, 29 (2), 101-109.

Feick, L. F., & Price, L. L. (1987). The market maven: A diffuser of marketplace information. Journal of Marketing, 51 (1), 83-97.

Feick, L. F., Price, L. L., & Higie, R. A. (1986). People who use people: The other side of opinion leadership. In R. Lutz (Ed.), Advances in consumer research (Vol. 13, pp. 301-305). Provo, UT: Association for Consumer Research.

Folkes, V. S., & Kiesler, T. (1991). Social cognition: Consumers' inferences about the self and others. In T. S. Robertson & H. H. Kassarjian (Eds.), Handbook of consumer behavior. Englewood Cliffs, NJ: Prentice-Hall.

Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18 (1), 39-50.

Grewal, D., Baker, J., Levy, M., & Voss, G. B. (2003). The effects of wait expectations and store atmosphere evaluations on patronage intentions in service-intensive retail stores. Journal of Retailing, 79 (4), 259-268.

Griffin, M., Babin, B. J., & Modianos, D. (2000). Shopping values of Russian consumers: The impact of habituation in a developing economy. Journal of Retailing, 76 (1), 33-52.

Hair, J. F. Jr., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis (5th ed.). Englewood Cliffs, NJ: Prentice Hall.

Higie, R. A., Feick, L. F., & Price, L. L. (1987). Types and amount of word-of-mouth communications about retailers. Journal of Retailing, 63 (3), 260- 278.

Holbrook, M. B., & Hirschman, E. C. (1982). The experiential aspects of consumption: Consumer fantasies, feelings, and fun. Journal of Consumer Research, 9 (2), 132-140.

Hu, L. H., & Bentler, P. M., (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6 (1), 1-55.

Joreskog, K., & Sorbom, D. (2001). LISREL 8: User's reference guide. Lincolnwood, IL: Scientific Software International.

Karande, K., & Lombard, J. R. (2005). Location strategies of broad-line retailers: An empirical investigation. Journal of Business Research, 58 (5), 687-695.

Kim, S., & Jin, B. (2001). An evaluation of the retail service quality scale for U. S. and Korean customers of discount stores. In M. C. Gily & J. Meyers-Levy (Eds.), Advances in consumer research (Vol. 28, pp. 169-176), Provo, UT: Association for Consumer Research.

Kozinets, R. V., Sherry, J. F., DeBerry-Spence, B., Duhachek, A., Nuttavuthisit, K., & Storm, D. (2002). Themed flagship brand stores in the new millennium: Theory, practice, and prospects. Journal of Retailing, 78 (1), 17-29.

Marmorstein, H., Grewal, D., & Fishe, R. P. H. (1992). The value of time spent in price-comparison shopping: Survey and experimental evidence. Journal of Consumer Research, 19 (1), 52-61.

Matilla, A. S., & Wirtz, J. (2001). Congruency of scent and music as a driver of in-store evaluations and behavior. Journal of Retailing, 77 (2), 273-289.

Michon, R., Chebat, J-C., & Turley, L. W. (2005). Mall atmospherics: The interaction effects of the mall environment on shopping behavior. Journal of Business Research, 58 (5), 576- 583.

Okada, E. M. (2005). Justification effects on consumer choice of hedonic and utilitarian goods. Journal of Marketing Research, 42 (1), 43-53.

Paridon, T. J. (2004). Retail opinion sharing: Conceptualization and measurement. Journal of Retailing and Consumer Services, 11 (1), 87-93.

Paridon, T. J. (2005a). Antecedents in retail information sharing research: The case for personal shopping values and consumer self-confidence. Journal of Applied Management and Entrepreneurship, X (4), 18-31.

Paridon, T. J. (2005b). The direct and indirect effects of personal shopping values in consumer self-confidence and retail information sharing research. Central Business Review, 24 (1-2), 20-25.

Paridon, T. J. (2006). Extending and clarifying causal relationships in research involving personal shopping value, consumer self-confidence, and word of mouth communication. Marketing Management Journal, XVI (1), 32-43.

Podsakoff, P. M., MacKenzie, S., & Lee, J-Y (2003). Common method bias in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88 (5), 879-903.

Reynolds, F. D., & Darden, W. R. (1971). Mutually adaptive effects of interpersonal communication. Journal of Marketing Research, 8 (4)), 449-454.

Richards, G. (1999). Vacations and the quality of life. Journal of Business Research, 44 (3), 189-198.

Rust, R. T., Lee, C., & Valente, E., Jr. (1994). Comparing covariance structure models: A general methodology. International Journal of Research in Marketing, 12 (4), 279-291.

Shim, S., & Eastlick, M. A. (1998). The hierarchical influence of personal values on mall shopping attitude and behavior. Journal of Retailing, 74 (1), 139-160.

Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhardt (Ed.), Sociological methodology (pp. 290-312). San Francisco: Josey-Bass.

Sobel, M. E. (1986). Some new results on indirect effects and their standard errors in covariance structure models. In N. B. Tuma (Ed.), Sociological methodology (Vol. 16 , pp. 159-186), Washington, DC: American Sociological Association.

Summers, J. O. (1970). The identity of women's clothing fashion opinion leaders. Journal of Marketing Research, 7 (2), 178-185.

Tauber, E. M. (1972). Why do people shop? Journal of Marketing, 36 (4), 46-59.

Titus, P. A., & Everett, P. B. (1995). The consumer retail search process: A conceptual model and research agenda. Journal of the Academy of Marketing Science, 23 (2), 106-119.

Turley, L. W., & Milliman, R. E. (2000). Atmospheric effects on shopping behavior: A review of the experimental evidence. Journal of Business Research, 49 (2), 193-211.

Voss, K. E., Spangenberg, E. R., & Grohmann, B. (2003). Measuring the hedonic and utilitarian dimensions of consumer attitude. Journal of Marketing Research, 40 (3), 310-320.

Wakefield, K. L., & Baker, J. (1998). Excitement at the mall: Determinants and effects on shopping response. Journal of Retailing, 74 (4), 515-539.

Wakefield, K. L., & Inman, J. J. (2003). Situational price sensitivity: The role of consumption occasion, social context, and income. Journal of Retailing, 79 (4), 199-212.

Table 1. Product Moment Correlations.

 Market Information Social
 Maven Sharing Confidence

Market 31.6/31.5
 maven (6.51)/(6.52)
Information 65 (a)/.70 (a) 23.3/27.6
 sharing (3.87)/(5.02)
Social 62 (a)/.62 (a) 54 (a)/.72 (a) 20.8/20.2
 confidence (4.44)/(4.32)
Personal -.21 (a)/-.01 -.18 (a)/.07 -.23 (a)/.01
 confidence
Hedonic .36 (a)/.40 (a) .31 (a)/.44 (a) .40 (a)/.40 (a)
 orientation
Utilitarian .40 (a)/.27 (a) 49 (a)/.20 (a) .44 (a)/.21 (a)
 value

 Personal Hedonic Utilitarian
 Confidence Orientation Value

Market
 maven
Information
 sharing
Social
 confidence
Personal 12.8/11.3
 confidence (6.26)/(5.36)
Hedonic -.05/.04 20.7/19.8
 orientation (5.48)/(5.54)
Utilitarian -.34 (a)/-.26 (a) .47 (a)/.41 (a) 22.8/23.0
 value (4.32)/(4.56)

Note: Main diagonal values are means with standard deviations in
parentheses. Income less than 35k/income equal to or greater than 35k

(a) p < .01, two tail test.

Table 2. Factor Structure and Standardized Loadings.

 Factor Personal Outcomes Factor
Hedonic Orientation Loading Confidence Loading

The shopping trip was .83/.84 I never seem to buy the .95/.59
 truly a joy. right thing for me.
The shopping trip truly .80/.80 Too often the things I .90/.61
 felt like an escape. buy are not
 satisfying.
I had a good time .78/.76 I often wonder if I've .72/.91
 because I was able to made the right
 act on the spur of the purchase selection.
 moment.
I enjoyed the shopping .67/.74 I often have doubts .72/.87
 trip for its own sake, about the purchase
 not just for the items decisions I make.
 I may have purchased

Utilitarian Value Information Sharing

The shopping trip was .87/.88 When my friends give me .78/.78
 useful. shopping advice I can
 use, I usually act on
 it.
I was satisfied with the .87/.83 When I help a friend by .73/.69
 items I purchased. telling her about my
 shopping experiences,
 I feel good about
 myself
I accomplished just what .76/.85 My friends and I enjoy .70/.75
 I wanted to on the talking about the
 shopping trip. styles and fashions we
 see on shopping trips.
While shopping, I found .70/.82 When we find quality .68/.83
 just the item(s) I was service in a store, my
 looking for. friends and I let each
 other know

Social Outcomes Confidence

I impress people with .83/.85
 the purchases I make.
I get compliments from .81/.66
 others on my
 purchasing decisions.
My friends are impressed .79/.87
 with my ability to
 make satisfying
 purchases.
My neighbors admire my .67/.56
 decorating ability.

Note: Income less than 35k values/income equal to or greater than 35k
values.

Table 3: [chi square] Difference Tests for Survey Group and Moderating
Effects

Invariant Original Groups Income groups

 [chi square] sig. [chi square] sig.
Path Difference level Difference level

Hedonic [right arrow] 1.31 ns 4.39 .05
 Social
Hedonic [right arrow] .18 ns .92 ns
 Personal
Utilitarian [right arrow] 4.34 .05 15.38 .01
 Social
Utilitarian [right arrow] 2.90 ns .93 ns
 Personal
Social [right arrow] 4.63 .05 .06 ns
 Sharing
Personal [right arrow] 2.63 ns 2.97 ns
 Sharing

Terrence J. Paridon, Cameron University Shawn Carraher Cameron University Sarah C. Carraher Consolidation Enterprises