Attitude toward the customer: a study of product returns episodes.

Past research in the retailing and services marketing literatures has focused on the importance of creating and maintaining customer satisfaction (Anderson and Mittal, 2000; Berry and Bendapudi, 2003; Fournier and Mick, 1999; Oliver et al., 1997). One particularly important aspect of the customer satisfaction process that has received significant attention recently is service recovery, which is defined as the process by which the firm attempts to rectify a service- or product-related failure (Maxham and Netemeyer, 2002). Retailer responses to failures are shown to reinforce customer relationships when handled properly (Blodgett et al., 1997), and/ or aggravate the negative effects of the failure when mismanaged (Hoffman et al., 1995; Krapfel, 1988). The behaviors and feelings of the retail salesperson are of critical importance during the service recovery process. The retail salesperson frequently represents the final (or only) point of contact between the retail firm and the customer, and the service recovery efforts he/she enacts can have an enormous impact not only on immediate customer satisfaction, but, importantly, on service quality perceptions and future patronage intentions (Pugh, 2001; Bitner et al., 1994, 1990).

One type of service recovery that has emerged as particularly salient to retailers is the handling of product returns. Returns transactions provide a critical point of contact between the retail firm and the customer, and return volumes can represent anywhere from 6-40% of sales for a given retailer. Return transactions also represent a prime opportunity for the retailer to recuperate assets and customers that might otherwise be lost (Stock, 1999; Rogers and Tibben-Lembke, 1999; Dunne et al., 2001).

In general, most research addressing product returns has assumed that the customer has a legitimate reason for returning a product, and/or the retailer should accept returns with little or no questions asked. However, anecdotal accounts and articles in the business press refute this perceived legitimacy perspective (Krapfel, 1988; Steinauer, 1997; Tomlinson, 2002; Industrial Relations Review & Report, 1994). Retail salespeople who handle returns will readily share stories about outrageous customer behavior. For example, some customers return clothing items that are stretched or ripped at the seams, claiming poor workmanship, when in fact it is clear that the article was simply too small (Dacy, 1994). Other retailers and salespeople cite "boomerang shopping" (c.f., Neuborne, 1996) or "renting" as a major issue; that is, the customer purchases items that are intended from the outset to be used and then returned at a later date. One customer practicing boomerang shopping openly described Wal-Mart as being the "best rent-a-center in the country," as he recalled his purchase of a snow-blower during the autumn months and its subsequent return in the spring after the snow thawed (Neuborne, 1996). These and other types of questionable customer behaviors related to returns, such as receipt fraud, price arbitrage, and the return of stolen merchandise (Speights and Hilinski, 2005), are estimated to cost the retail industry over $13 billion per year (Rogers et al., 2005), placing returns abuse in a similar category with shoplifting and returned checks in terms of lost value to retailers.

Unfortunately, the retail salesperson is frequently "caught in the middle," between satisfying customer desires and enforcing store policy. The salesperson may feel pressured to ignore customer malfeasance and provide "service with a smile," even in the face of overt or blatant attempts by the customer to take advantage of policies that are originally enacted with his/her best interests in mind (Tomlinson, 2002). The salesperson may suspect that the customer is blatantly misleading or even providing false information. It stands to reason that illegitimate or devious acts by the customer meant to circumvent store policies could eventually result in a negative emotional reaction by the retail salesperson.

In the marketing domain, consumer attitudes toward products, services, employees, retail settings, and transactions have played a dominant role in research for over 25 years. However, the attitudes of customer contact personnel--salespersons, retail clerks, service representatives, and marketing management--are much less understood. Our study addresses this deficiency in the literature by introducing attitude toward the customer as a new concept of interest. Attitude toward the customer (hereafter, Ac) is defined herein as the customer contact's predisposition to respond favorably or unfavorably toward a particular customer during an exchange episode. This definition is consistent with other attitudinal constructs studied in marketing (e.g., attitudes toward the advertisement or brand, per Mitchell and Olson (1981)).

This study seeks to answer an important question related to the attitude toward the customer concept: Does the behavior of customers during a service recovery episode influence the retail salesperson's attitude toward the customer? Specifically, Ac is examined in this article within a common and extremely important service recovery context: the retail product returns process.

This research explores the salespersons' Ac within the retail returns context. Specifically, this article examines institutional determinants of Ac under conditions where the customer is perceived as being in breach of the implicit contract governing retail exchange. The following sections of the article (1) introduce the Ac construct, (2) test and evaluate a model useful for predicting Ac in an experiment conducted within the retail returns context, based on relational contracting and perceived legitimacy/institutional theory, and (3) provide discussion and implications of the study for retailers and academics interested in further improving the returns-related service recovery effort.

LITERATURE REVIEW AND HYPOTHESES

The Problem of Retail Returns

In the U.S. and some other parts of the world, when customers are dissatisfied with a purchase they have made, the retailer will "guarantee" the product by allowing the customer to return it within reasonable guidelines. The customer can exchange it for a similar or different item, for store credit or, in some cases, for a full refund (Dunne et al., 2001). Liberalized retail returns represent an outgrowth of the focus on customer satisfaction, the relationship value of the customer, and customer relationship management initiatives. Part of this emphasis has resulted from the increased attention to the economic benefits firms accrue with the retention of loyal customers (e.g., Reinartz and Kumar, 2003; Stahl et al., 2003). Clearly, returns are an important issue for retailers; 91% of consumers interviewed for one study considered return policies and processes as an important factor in their decision about where to make a purchase (see Schuman, 2004).

However, some retailers are becoming concerned that they take the service recovery effort too far. There is a trend toward "no-questions-asked," extremely liberalized returns policies where the customer can return nearly anything, nearly anytime, and in nearly any condition (Autry, 2005; Dacy, 1994; Rosenbaum and Kuntze, 2003). Not surprisingly, many customers do just that, to the detriment of both the retailer and the retail salesperson who has worked hard to serve them. While understanding the returns process is of paramount importance, the role of the salesperson in the retail returns process has generally been ignored both in the marketing literature and in retail practice.

Attitude toward the Customer (Ac)

Attitude toward the customer is an underdeveloped area of marketing. However, it is apparent that an employee's reaction to a customer can be influenced by their evaluation of an interaction with a customer, over and above their feelings about the sales transaction itself. We propose that a specific type of attitude object, then, is the customer him/herself and is based on the salesperson's reaction to a particular interaction. Research has demonstrated repeatedly that attitudes toward an ad or brand serve as a significant predictor of product evaluations (c.f., Burton and Lichtenstein, 1988; Cox and Cox, 1988; MacKenzie and Lutz, 1989; MacKenzie et al., 1986; Mitchell and Olson, 1981; Mittal, 1990; Muehling and Laczniak, 1988; Shimp, 1981). Accordingly, it is reasonable to assume that the nature of the particular interaction between the customer contact personnel and the customer would predict an evaluation of the customer. This evaluation is critical for retailer performance, as evaluations of customers have been linked to sales effectiveness (Bowen and Schneider, 1985; Evans et al., 2000).

Theoretical Foundations

The examination of customer behaviors during service recovery and commensurate salesperson attitudes about the customer is grounded in institutional and relational contracting theory. The research framework depicting the focal theoretical relationships is presented in Figure I. Specifically, two postulates are offered and utilized related to the exchange environment encompassing the customer-retail salesperson relationship. First, we postulate that retail exchanges are institutionalized, i.e., governed by structural assumptions stipulating that the actions of either party are desirable, proper, or appropriate within a socially constructed system of norms, values, and/or beliefs (e.g., Nielsen and Rao, 1996; Suchman, 1995). Second, we postulate that retail exchanges are monitored and evaluated via an implied contract between the customer and retailer (e.g., Robinson and Rousseau, 1994; Morrison and Robinson, 1997). Though neither of these postulates is wholly new to the social science literatures, they have rarely been operationalized to examine retail exchange phenomena, and thus their adoption herein represents a secondary contribution of the article to the existing knowledge base. Each postulate is now examined in detail, and hypotheses stemming from their integration are offered for the purpose of predicting retail salesperson Ac as associated with service recovery.

[FIGURE I OMITTED]

Referring to the first postulate, institutional theory has long been used to explain the viability or perceived legitimacy of entities, most often organizations. Through legitimization, organizations are said to offer evocative external symbols (i.e., brands, logos, messages, etc.) to garner societal support (e.g., Ashforth and Gibbs, 1990; Pfeffer and Salancik, 1978; Suchman, 1995) or to become culturally embedded such that common expectations form related to organizational performance (e.g., DiMaggio and Powell, 1983; Zucker, 1987). There is general agreement in prior institutional theory research that credibility and various types of support for entities are determined as a result of their being granted legitimacy status (Suchman, 1995). This status is gained as a result of three linked but distinct processes where entities are verified as being in compliance with regulatory institutions, normative institutions, and cognitive institutions (Scott and Meyer, 1983; Meyer and Zucker, 1989; Berger and Luckmann, 1967). In other words, legitimacy is perceived in proportion to actors' tendency to conform to commonly-accepted rules, social norms, and/or cognitions (Suchman, 1995). Cognitive legitimacy is defined as the appropriateness, interpretability, and comprehensibility of the action (i.e., does the behavior "make sense"). Normative legitimacy is concerned with whether the action conforms to commonly-held social values and norms of acceptable behavior. Regulative legitimacy refers to the conformance of the action to established rules, policies, and procedures (i.e., whether behavior falls within pre-set parameters).

Institutional theory, to date, has generally been used to explain the perceived legitimacy of organizations or firm processes; it has rarely, if ever, been employed in the marketing context. The current study is therefore an initial endeavor adopting institutional theory for the purposes of explaining the dynamics of the exchange process. Herein, legitimacy is operationalized as the expected norms, rules, and cognitions governing the actors within retail exchange episodes. This operationalization is important in that for retail returns, there are behavioral expectations grounded in rules, social norms, and logic that the parties to the exchange are expected to follow.

The second postulate addresses the possibility that actors in the retail exchange relationship may not act according to social expectations as set forth by institutional theory. As a result of the institutionalization of retail exchanges, expectations for behaviors within exchange episodes are established, and in the case of the current research, returns transactions may or may not be handled by either party according to those expectations. The mutual expectations that the exchange partners (in this case, the retail salesperson and the customer) are bound by take the form of an implied psychological agreement that each will act within established institutional guidelines. Psychological contracting theory has been commonly used to examine violation of such agreements, most commonly in the domains of human resource management (i.e., the employer-employee relationship) (c.f., Brockner et al., 1992), or marketing channel arrangements (c.f., Lusch and Brown, 1996).

Two elements of relational contracting theory are particularly salient to the discussion of the contract governing the individual level retail exchange relationship. The first, breach, refers to the cognition or realization that an implied or explicit promise made by one party to another has gone unfulfilled (Robinson, 1996; Robinson and Morrison, 2000; Rousseau, 1989). The second, violation, represents a frequent outcome of breach at the affective level, defined as feelings of betrayal and deeper psychological distress whereby the victim feels a sense of injustice or wrongful harm (Rousseau, 1989).

Considering the two postulates together, relationships between perceived legitimacy and contracting in the retail realm can be established: breach and violation are important to the discussion of retail returns when grounded in institutional theory. Specifically, when norms, rules, and cognitive expectations are adhered to by customers during their initiation of the service recovery episode, the attitudinal reaction of the salesperson is normalized with respect to that exchange (i.e., the salesperson may possess negative affect related to the transaction for other reasons such as personality conflict or perceived low "effort-to-return" ratio, but is more likely to invoke an internal attribution (e.g., become self-frustrated) because the customer is not in violation of the psychological contract) (Folkes, 1984). However, if the customer acts inconsistently with social norms governing retail returns transactions (1), abuses or appears to abuse store rules or return policies (e.g., returns deadlines or product types), and/or cites nonsensical or illogical reasons in support of the return request (e.g., returns for reasons that are not functionally related to the product's common usage), the salesperson is likely to believe that the customer is breaking the good-faith agreement between the retailer and the customer (i.e., behaving opportunistically; Williamson, 1975). In simplified terms, when the customer behaves in a way perceived by the retail salesperson as illegitimate (consistent with any of the three conceptualizations of the term), the retail salesperson is likely to perceive that the implied exchange contract has been breached.

These ideas are stated formally as our first set of research hypotheses:

H1a: The perceived cognitive legitimacy of the return request is negatively associated with the salesperson's perception that the implied customer-salesperson contract is breached.

H1b: The perceived normative legitimacy of the return request is negatively associated with the salesperson's perception that the implied customer-salesperson contract is breached.

H1c: The perceived regulative legitimacy of the return request is negatively associated with the salesperson's perception that the implied customer-salesperson contract is breached.

As a result of perceived illegitimate behavior on the part of the customer, the salesperson who values legitimacy in the exchange process may also experience a negative emotional reaction. Rafaeli and Sutton (1989) note that an important factor determining emotive responses within job role performance is the emergence of and/or evocation of social, occupational, and organizational norms by others. From a regulative perspective, Ortony et al. (1988) find support for the premise that feelings of disappointment, frustration, and distress typically emanate from the disconfirmation of a process or event that is perceived to be desirable, especially if the focal actor (in this case, the retail salesperson) has previously invested effort toward the process or event. Therefore, assuming that the retail salesperson believes that the original sales transaction was enacted by both parties in good faith, the illegitimate nature of a return transaction may result in a negative attitudinal effect on the part of the salesperson. Alternatively, from the cognitive perspective, DiMaggio and Powell (1983) suggest that a negative emotional reaction is a possible outcome when actors must struggle to reconcile other's accounts of experiences with their larger belief systems and/or experienced realities. In other words, actors presented with arguments that conflict with their commonly-held beliefs regarding reasonable or logical behavior tend to experience stressful and negative feelings related to the source of conflict. This finding is also consistent with the foundations of employee role stress theory (c.f., Kahn et al., 1964). In the retail exchange episode, customers making illogical statements related to their desire to return products will be likely to generate internal turmoil in the salesperson, who must attempt to balance the customer's story with his/her own judgments of the situation.

Taken together with the regulative and normative evidence, this forms the basis for the second set of hypotheses:

H2a: The perceived cognitive legitimacy of the return request is negatively associated with the salesperson's Ac.

H2b: The perceived normative legitimacy of the return request is negatively associated with the salesperson's Ac.

H2c: The perceived regulative legitimacy of the return request is negatively associated with the salesperson's Ac.

The second hypothesis predicts a direct relationship between legitimate customer behavior and the retail salesperson's Ac. However, the premises of psychological relational contracting (Morrison and Robinson, 1997; Rousseau, 1989; MacNeil, 1985) suggest that in addition to a direct effect of legitimate behavior on Ac, there are also indirect relationships via breach to violation to Ac. To complete this portion of the puzzle, it is notable that violation is the natural outcome of breach, and that violation of the contract should predict Ac in the service recovery context. The direct linkage between breach and violation has been shown in the literature (e.g., Morrison and Robinson, 1997; Rousseau, 1989), where violation is positioned as the natural emotional outcome resulting from having been cheated or betrayed as a result of another party's unfulfilled promise. In the retail exchange episode, the salesperson believes that the customer will keep the product once bought, and this represents to the salesperson an implied promise on the part of the customer that is violable only under legitimate conditions (e.g., defective, wrong size, unusable). If the customer breaches the contract, having enacted the return process based on reasons the salesperson perceives to be illegitimate, the salesperson sustains a negative emotional reaction to the customer as a result of the betrayal. Furthermore, if the violation is significant, i.e., passes a critical threshold of inequitable or unacceptable behavior (Huseman et al., 1987), it becomes embedded in the salesperson's social characterization of the customer, and serves to alter the salesperson's emotive state related to that particular customer.

Thus, the remaining hypotheses present Ac as predictable by contract violation:

H3: A perceived breach of the implied customer-salesperson contract is positively associated with feelings of violation related to the implied contract.

H4: Feelings of violation related to the implied contract are negatively associated with the salesperson's Ac.

METHODOLOGY

The context chosen to examine the hypotheses is that of retail salespeople. This context was chosen for a number of reasons. Foremost among these reasons is that retail salespeople are more likely than other salesperson types to be integral in the exchange process from both seller to buyer and, as is pertinent in this study, from buyer back to seller. Other sales force positions skew the attention of the salesperson more toward the outbound selling process than the return aspect. Second, retail salespeople sometimes receive explicit training toward the aim of completing returns, and the return process often takes a significant portion of the salesperson's time while on duty. Finally, this research addresses the salesperson's attitude as it relates to providing the overall service encounter within a retail context; the same would not always be possible for business-to-business transactions, where post-sale service functions are sometimes handled by support staff.

Prior to data collection for the purposes of testing, focus groups were conducted at a large southwestern university using undergraduate and graduate students who had retail sales or retail management experience. The results from these focus groups were used to develop the scenario manipulations utilized in the survey.

The study utilized a quasi-experimental design where respondents first read a scenario in which a customer requests to return an item previously bought. Respondents were then asked to answer questions related to their perceptions of the scenario. The aim of the scenario was to create a service encounter that would provide a common basis for judgment. The scenario was embedded in a full factorial 2x2x2 design where each level of perceived legitimacy (normative, cognitive, and regulative) was manipulated to be either high or low. An example of the scenario for design (high rules, high norms, high cognitive) is included in Appendix A. A convenience sample of retail salespeople working for 314 area merchants listed in a local business directory was contacted in person and asked to participate in the study. In order to verify that the results of the study are generalizable, care was taken to sample retailers from multiple different retail formats, and subsequent testing of all study measures indicated no significant differences across retail formats. A total of 238 questionnaires were completed and entered into a spreadsheet for analysis; 76 participants declined to participate, frequently citing store policy as the reason.

The average age of the respondents was 25. The average length of time with the current employer was just over three years; however, a majority of those responding had been with their current employer 18 months or less. Additionally, the respondents indicated an average of over five years retail selling experience, with half the sample having at least three years. Despite the respondents' significant experience, only 25% of the sample considered themselves full-time employees. The respondents were predominantly (64%) female. Most importantly, however, is the fact that 85% reported having received training on what to do when a customer asked to return an item, indicating that the sample was appropriate for the purposes of the current study. (2)

Measures

Perceived Legitimacy. Three items measured the efficacy of the scenario manipulations, one for each type of legitimacy perception under investigation (i.e., Regulative, Normative, and Cognitive). As reflected in Appendix B, the Regulative measure asked if the respondent believed the customer in the scenario was following the store's return policy, while the Normative measure asked if they thought the customer was acting in a socially acceptable manner, and the Cognitive measure asked if the respondent thought that the reason for the request made sense. Likert-type scales were used for these manipulation checks.

Since these measures were embedded in a 2x2x2 design, the data were dummy coded to reflect the high and low scenario conditions for each level of perceived legitimacy. Subsequently, ANOVA tests were conducted to test the manipulation. For the Regulative measure, the high manipulation condition had a mean of 4.67 and the low a mean of 1.74, indicating significance (F = 142.76, 1df, p < .01), and was therefore considered a successful manipulation. The results for the Normative measure (high = 5.25, low 2.69, F = 118.70, 1df, p < .01), and the Cognitive measure (high = 4.68, low 2.32, F = 78.01, 1df, p < .01) were similar. These results not only indicate the efficacy of the manipulation, but also serve as a proxy measure for the discriminant validity of each of the dimensions of return legitimacy as each were measured under experimental control of the other two dimensions.

Breach and Violation of Implied Contracts. The Breach and Violation of Implied Contracts scales were adapted from work that utilized them successfully (Robinson and Morrison, 2000). Although these constructs are highly related, the distinction between them is an important one. The Breach measure attempts to assess the underlying cognitive beliefs about the implied contractual agreement, whereas the Violation measure assesses the affective influence of the suggested breach. Reliability analysis indicated Cronbach alphas of 0.79 and 0.65 for the Violation and Breach constructs, respectively. Although the Breach construct has marginal reliability, it was deemed adequate for analysis due to its theoretical and psychometric history within the literature.

Intention to Accept Return. To assess the retail salesperson's intent on accepting the return, a single item was used for measurement. Here, respondents indicated their level of agreement on a seven-point Likert scale (anchored at 1 = completely disagree and 7 = completely agree) to the statement "I would be likely to accept this return from the customer."

Attitude Toward the Customer (Ac). The retail salesperson's Ac was measured by adapting scales that have been previously used in assessing attitude toward the brand and repurchase intention. Together, this scale's six items represent the Ac in both the affective and behavioral sense; the affective measures tap into the respondents' feelings toward the customer following the scenario, and the behavioral element is represented by the retail salesperson's intention to act in a manner that is helpful to the customer in the future. The Cronbach alpha generated for this scale was 0.91.

Given the quasi-experimental design used, we began our hypothesis testing by assessing the influence of the bases of perceived legitimacy upon both Breach of Implied Contract (H1) and Ac (H2). With the 2x2x2 full factorial design, each of the bases of perceived legitimacy was crossed with the other two bases. ANOVAs were used to compare the mean score on the Breach of Implied Contract scale in both the High control cases to the Low control cases for each base, thereby experimentally controlling for the other two bases. This analysis was again repeated for the influence of the perceived legitimacy bases on Ac.

The first analyses were concerned with Breach using the Regulative Legitimacy measure. Here, the high perceived Regulative Legitimacy condition had a mean of 8.96 on the summated Breach scale and the low condition resulted in a mean of 8.11, indicating that the higher the perceived legitimacy, the less likely the respondents perceived a breach in the contract. However, the F test indicates that these differences were not statistically significant (F = 2.31, 1df, p > .10). The results were similarly non-significant for perceptions of Cognitive Legitimacy (High Mean = 8.51, Low Mean = 8.56, F = 0.006, 1df, p > .10). Contrary to these findings were the results of the analysis using the perceived Normative Legitimacy measure, where the results were both significant and in the hypothesized direction (High Mean = 7.21, Low Mean = 9.84, F = 0.006, ldf, p < .01). Accordingly, we find only support for the hypothesized relationship between Normative Legitimacy and Breach of the implied contract.

The procedure described above was followed again for assessing the influence of legitimacy upon Ac. Here, the results mirrored the above analysis where perceptions of both Regulative Legitimacy (High Mean = 4.14, Low Mean = 4.25, F = 0.28, 1df, p > .10) and Cognitive Legitimacy (High Mean = 7.21, Low Mean = 9.84, F = 0.006, 1df, p > .10) were not significant. Again, the influence of the Normative Legitimacy was both significant and in the hypothesized direction (High Mean = 4.93, Low Mean = 3.49, F = 64.76, 1df, p < .01). Thus, we only find support for Normative Legitimacy perceptions influencing Ac.

Model Testing

Provided with the evidence depicted in the initial analysis, the model in Figure I was adapted to reflect both significant and non-significant findings. The second model is presented in Figure II. To test this model via structural equations modeling, we first conducted a confirmatory factor analysis with all of the constructs in the model. A variance-covariance matrix was produced and transported into LISREL 8.72 (Joreskog and Sorbom, 2005) for confirmatory factor analyses and structural equation analyses. Table 1 presents the findings from the psychometric testing. The results indicate that all but three of the items across all constructs have loadings above 0.70, and all items are significant via the t-statistic. Furthermore, the composite reliability of each construct met or exceeded the commonly accepted benchmark of 0.70 (Nunnally, 1978), and two of the three constructs exceeded the benchmark of 0.5 for VE. The results of Chi-square test and goodness of fit indices (Chi-square = 196.03, df = 60, p < 0.05, GFI = 0.89, CFI = 0.93) showed the measurement model had acceptable fit to the data. Further, the Chi-square/df ratio of 3.27 is well below the recommended level of 5.0 (Bollen, 1989), indicating an acceptable model fit. Finally, all of the cross-construct correlations were significantly different from 1.0 (tested via a Chisquare test with 1 degree of freedom when constraining the path to 1.0 rather than allowing free estimation), providing evidence of discriminant validity.

[FIGURE II OMITTED]

The results demonstrate the measures employed in this study possess adequate construct validity. Descriptive statistics concerning the summated variables used in the structural equations model can be seen in Table 2.

ANALYSIS

The model presented in Figure II was tested in LISREL 8.72 (Joreskog and Sorbom, 2005) and, following Anderson and Gerbing (1988), the measurement model was constrained to better assess the structural model. The results of the structural model are reported in Table 3. The Chi-square goodness of fit measure is significant (Chi-square = 197.95, df = 83, p < 0.05), and other fit indices such as the CFI, IFI, and RMSEA (Table 3) met or exceeded recommendations for adequate model fit (Hair et al., 1995). Finally, the Chi-square/ degrees of freedom ratio of 2.38 are well below the recommended levels (Carmines and McIver, 1981; Bollen, 1989).

Results of Testing

Following the results of the ANOVA presented above, the structural equations model was used to assess the research hypotheses. Hypothesis H1b predicted a significant and negative association between Normative Legitimacy perceptions and Breach of the implied contract. The results presented in Table 3 indicate support for the hypothesis ([gamma] 11 = -0.45, t = -5.99). Additionally, the path from Normative Legitimacy to Ac was also significant ([gamma] 31 = 0.62, t = 9.44), thereby supporting H2b.

The strongest path indicated by the results was the influence of Breach of the implied contract upon Violation ([beta] 21 = 0.92, t = 15.93). Accordingly, we find evidence to support H3. Violation was also hypothesized to have a negative and significant influence upon Ac. The results in Table 3 lend strong support in favor of H4 where the path was negative ([beta] 32 = -0.25) and significant (t = -4.05).

It is also important to consider potential effects related to the salesperson's behavioral intentions and personal characteristics. To examine potential moderating effects of the intention to accept the return from the customer, group analysis was performed. Here, group analysis was performed by splitting the sample using the Intention to Accept variable, whether or not the retail salesperson was a full- or part-time employee, compensated on a commission basis or salaried, received training or did not receive training for conducting returns, and the length of retail career experience of the respondent, age, gender, compensation, and education. The results of the group analysis between the estimation where the paths were constrained to be equal and an analysis where the paths were free to be estimated for each group produced a calculated Chi-square, in each instance, which was less than the critical value at 1df. None of the variables in consideration significantly influenced any of the relationships presented in Figure II, providing support for the generalizability of the study.

IMPLICATIONS AND FUTURE RESEARCH

Managerial Implications

The results of this study provide several strong managerial implications. Most striking is the fact that while there was a perceived normative legitimacy effect, there was no effect for perceived violation of either cognitive or regulative legitimacy. This means that as long as the customer acts in a socially acceptable manner, the sales-focused employee will tend to overlook violations of policy and rational explanations and accept the return. Post hoc analyses indicated that this held true regardless of whether or not the employee was full- or part-time, compensated on a commission basis or salaried, received training or did not receive training for conducting returns, and their length of experience.

The retail manager is thus faced with a dilemma: should employees be empowered to accept returns, or should all returns be directed through a centralized center where employees are given intensive training which includes an awareness of the influence of social norms on the return process? Each option poses advantages and disadvantages for the retailer. If the employee is empowered, such empowerment is typically seen to influence job satisfaction (Spreitzer et al., 1997) and commitment to the organization (Liden et al., 2000), and is likely to lead to customers' perceptions of enhanced customer service due to the immediacy of resolution of the potentially negative situation. However, such empowerment will reduce the control the organization has over the delivery of the returns service (Hartline and Ferrell, 1996) and the variability of service, and thus there are clearly advantages to centralizing the process. Regardless of the decision made by retail management, our findings imply a need for better understanding of social norms active in the return process. Training programs for customer-contact employees should emphasize that the employee employ reasonable safeguards to insure that customer requests are "within the rules." During training, store managers should strongly emphasize store returns policies, and given the possibility that sales-oriented employees will tend to "look the other way" in some situations, managers must attempt to generate employee "buy-in" with respect to returns handling. Managers may also want to provide specific training with respect to the most common returns scenarios, even to the extent that recurring customer problems are rehearsed and resolved with pre-planned response mechanisms. The denial of a returns request may generate role conflict for a retail salesperson, and thus it may be necessary to carefully prepare employees for such situations in order to preserve future sales with the affected customer.

This study also indicates that when social norms were perceived to be violated, the employee does indeed develop a negative attitude toward the customer. Both employee and/or customer dissatisfaction are resultant possibilities, and the potential exists for damage to hard-earned perceptions of loyalty on both sides of the relationship. George and Hegde (2004) stress the importance of employee disposition, amongst other things, as being instrumental in determining customer's satisfaction with the service. After all, what customer wants to return to a store to be served by a person who barely can hide their negative attitude? Clearly, it is in the best interest of the company to provide the best possible customer experience, and the maintenance of firm but positive attitudes by retail salespersons is the key to future dealings with the affected customer. Though negative employee attitudes are likely to occur to some extent during returns handling, these may be tempered by managers practicing expectations-lowering procedures (i.e., Buckley et al., 1998) or practicing what others call "realistic recruiting" (Buda and Charnov, 2003). Managers should emphasize to sales employees that returns are an inevitable part of the sales process that will at times be frustrating, and illustrate that via their proper handling, returns transactions in the short term can actually result in longer-term sales successes due to enhanced customer service perceptions. By teaching employees to view sales success cautiously, and illustrating that relational selling and good service work together in the long term to generate enduring retailer and salesperson performance, the salesperson will be more likely to develop a customer orientation and provide higher levels of service, which in turn, enhance retailer reputation.

Additionally, as noted previously, the employee with a negative attitude toward customers can easily become disengaged and may eventually quit. Accordingly, the focus of management should be not only on retaining customers but also on retaining employees. Again, training and empowerment are critical. If employees are empowered to handle transactions, it is necessary to (a) emphasize during training that returns handling is a core element of the job, rather than an afterthought, and (b) train the employees specifically on how to handle returns transactions so that they serve to build customer-firm loyalty rather than to tear it down.

Implications for Research

Two primary implications for research emerge as a result of this study. First, though customer attitudes have been examined in a large volume of literature in the marketing domain, prior to this study little regard was given in the literature to the attitudes of retail employees interacting with customers; employees have, in effect, been assumed to be a constant. However, the findings of this study indicate that beliefs and feelings emanating from the seller side of the dyad are important and may have a lasting effect on exchange relationships. Particularly interesting in this regard is the role of customer norms within buyer-seller interactions. Though decades of research indicate that the violation of norms by employees and firms generates a negative impact on customers, our research shows the attitudes of marketing employees such as retail salespersons are similarly affected by customer violations. Unfortunately, how these effects are manifested remains uncertain at this point. Though the current study illustrates one potentially negative outcome--that the salesperson may elect to decline the customer return, thus undermining the service recovery effort--other more damaging outcomes may emerge in the long term. Irrational or uncontrolled reactions to customer norm violations by employees may result in damage to customer loyalty, in that customers generally know and believe the retail mantra that the "customer is always right." Research should assess alternative outcomes of service failure, for both the customer and retail employee.

From the perspective of the employee, being caught between doing what is right for the firm and that which is desired by the customer they are hired to serve may result in multiple forms of role stress. The salesperson may feel role conflict, which research has shown to be associated with negative job affect, lateness, emotional withdrawal, and employee turnover. Without proper training on returns handling and/or service recovery, role ambiguity may also arise. Though the current study does not address these issues specifically, they are each potentially damaging to the business and are thus worthy of future attention.

Second, due to increasing liberalization of returns policies directed at capturing long-term customer value, retail firms have experienced commensurate increases in return volumes. Though the product returns process has recently enjoyed increased focus as a research topic, in general the extant studies in this area are operationally focused, i.e., the research examines the efficiency and effectiveness of the returns process and its elements rather than the execution of the process. Critical to the execution of the returns process is the role of the employee, and yet this module of the service recovery effort remains somewhat of a mystery in the literature. Specifically, very little is understood about employee functions, duties, tasks, and problematic issues within the returns context. In general, the returns process has been managed and operated in the past on a by-exception basis; little formal policy has traditionally governed employee actions within the process, and the topic represents only a minor focus of most sales-focused customer contact employee training (with the notable exception of specialized returns-handling customer service representatives). When the retail supply chain reverses directions for a brief moment in time as a result of a customer-initiated return request, many retail salespeople are forced to execute a task for which they are not only less prepared but also may be unenthusiastic to accomplish. The role of the customer contact employee in the holistic service recovery context is cloudy at best, and significant further investigation is needed to clarify this role.

Limitations and Future Research

While the current research appears to have strong findings, it is clear that additional research is needed. First, our study should be replicated and constrained within specific retail settings and formats. For example, would the finding of no effect for rules- and cognitive legitimacy perceptions hold for all types of retailers, or would expensive specialty clothing stores (with clientele possibly possessing enhanced social acuity and/or greater financial resources) vary compared to discount chains, where volumes are larger and return transactions are potentially more frequent?

Secondly, employee role conflict and role ambiguity have been shown to play a significant role in customers' perceptions of service quality (Hartline and Ferrell, 1996). Of interest would be a study which examines exactly how these two constructs impact the formation of attitudes toward customers under different legitimacy violations. For example, what are conditions under which role stress or ambiguity would increase the negative emotions of norms violation? One would anticipate that high role ambiguity would lead to an increase in perceptions of the violation and thus an increase in the negative affect.

Additionally, it should be recognized that for the purposes of the current study, scenarios were used to prompt employees related to fictitious service recovery situations. Thus, our assessment of the relevant attitudes and behaviors is made in a purely controlled environment. Additional research using observation of actual returns handling processes may provide additional insight into the issues surrounding Ac in the service recovery context. Studying these and other related issues could give practitioners a stronger basis for making retail returns decisions.

Appendix A

SAMPLE EXPERIMENTAL SCENARIO

High Regulative / High Normative / High Cognitive

You are employed as a part-time employee at The Pag, a clothing store in a large mall in Phoenix. You are alone in the store because the manager has gone on dinner break, and are doing typical work cleaning and straightening the store. The date is January 2 *. A customer comes into the store and asks to return a dress that was bought for her as a Christmas present. Your manager has often reminded you that returns on Christmas items are not accepted after the first day of February. The customer says that the dress is too tight in the back and sleeves area **, and wants to return it. You open the bag and remove the dress. It looks new and clean and still has your store tags on it ***.

* The low regulative manipulation put the date at February 12.

** In the low cognitive manipulation customer says that "the dress is impossible to hang on a wire coat hanger."

*** In the low normative manipulation the scenario states "It looks worn and slightly dirty, and has no tags on it."

Appendix B
MEASUREMENT ITEMS

Variable
Name           Alpha   Label                 Item

Regulative     N/A     RL1     The customer is following the store's
Legitimacy                     return policy.
Normative      N/A     NL2     This customer is acting in a socially
Legitimacy                     acceptable manner in asking for the
                               return.
Cognitive      N/A     CL3     The reason that the customer is making
Legitimacy                     this return makes sense.
Intention      N/A     ACT     I would be likely to accept this return
to Accept                      from the customer.
Return
Breach         0.65    B1      I would believe that the customer broke
                               a silent promise to keep the item.
                       B2      I would think that the customer owed it
                               to me to keep the dress.
                       B3      I understand that the customer was NOT
                               obligated to keep the item. Reverse
                               coded.
Violation      0.79    V1      In asking for this return, I would feel
                               that the customer betrayed me.
                       V2      When the customer asked to return the
                               dress, I would feel badly since she
                               violated our implied contract.
                       V3      I would feel upset that the customer was
                               unfaithful in asking for the return.
Ac             0.91    ATT1    My feelings toward this customer are ...
                               (Measured on seven-point scale where 1 =
                               very unfavorable and 7 = very
                               favorable.)
                       ATT2    The likelihood that I would want to help
                               this customer if she
                               came into my store is ...
                               (Measured on seven-point scale where
                               1 = very unlikely and 7 = very
                               likely.)
                       ATT3    I like this customer ...
                               (Measured on seven-point scale where 1 =
                               not at all and 7 = very much.)
                       ATT4    This customer left a positive impression
                               on me.
                       ATT5    If I were to encounter this customer, I
                               would want to help her
                               select a product.
                       ATT6    If this customer entered my store, I
                               would rather NOT assist her. Reverse
                               coded.

* Measured on seven-point Likert scale where 1 = completely disagree
and 7 = completely agree unless otherwise noted.

References

Anderson, E. W. and V. Mittal. 2000. "Strengthening the Satisfaction-Profit Chain." Journal of Service Research 3: 107-120.

Anderson, J. C. and D. W. Gerbing. 1988. "Structural Equation Modeling in Practice: A Review and Recommended Two-step Approach." Psychological Bulletin 103: 411-423.

Ashforth, B. E. and B. W. Gibbs. 1990. "The Double-edge Sword of Organizational Legitimization." Organization Science 1:177-194.

Autry, C. W. 2005. "The Formalization of Reverse Logistics Programs: A Strategy for Managing Liberalized Returns." Industrial Marketing Management 34: 384-391.

Berger, P. L. and T. Luckmann. 1967. The Social Construction of Reality: A Treatise in the Sociology of Knowledge. New York, NY: Doubleday.

Berry, L. L. and N. Bendapudi. 2003. "Clueing In Customers." Harvard Business Review 81: 100-107.

Bitner, M.J., B. H. Booms and L. A. Mohr. 1994. "Critical Service Encounters: The Employee's Viewpoint." Journal of Marketing 58: 95-106.

--, --, and M. S. Tetreault. 1990. "The Service Encounter: Diagnosing Favorable and Unfavorable Incidents." Journal of Marketing 54: 71-84.

Blodgett, J. G., D.J. Hill and S. S. Tax. 1997. "The Effects of Distributive, Procedural, and Interactional Justice on Post-complaint Behavior." Journal of Retailing 73: 185-210.

Bollen, K. A. 1989. Structural Equations with Latent Variables. New York, NY: Wiley and Sons.

Bowen, D. E. and B. Schneider. 1985. "Boundary-spanning Role Employees and the Service Encounter: Some Guidelines for Management and Research." Chapter in The Service Encounter: Managing Employee/Customer Interaction in Service Businesses. Eds. J. A. Czepiel, M. R. Solomon and C. F. Suprenant. Lexington, MA: D.C. Heath and Company. pp. 127-147.

Brockner, J., T. R. Tyler and R. Cooper-Schneider. 1992. "The Influence of Prior Commitment to an Institution on Reaction to Perceived Fairness: The Higher They Are the Harder They Fall." Administrative Science Quarterly 37: 241-261.

Buckley, M. R., J. G. Veres, D. B. Fedor, D. S. Wiese and S. M. Carraher. 1998. "Investigating Newcomer Expectations and Job-related Outcomes." Journal of Applied Psychology 83: 452-461.

Buda, R. and B. H. Charnov. 2003. "Message Processing in Realistic Recruitment Practices." Journal of Managerial Issues 15 (3): 283-316.

Burton, S. and D. R. Lichtenstein. 1988. "The Effect of Ad Claims and Ad Con-text on Attitude toward the Advertisement." Journal of Advertising 17:3-11.

Carmines, E. G. and J. P. McIver. 1981. "Analyzing Models with Unobserved Variables: Analysis of Covariance Structures." In Social Measurement: Current Issues. Eds. G. W. Bohrnstedt and E. F. Gorgatta. Beverly Hills, CA: Sage.

Cox, D. S. and A. D. Cox. 1988. "What Does Familiarity Breed? Complexity as a Moderator of Repetition Effects in Advertisement Evaluation." Journal of Consumer-Research 15: 111-116.

Dacy, J. 1994. "When Shoplifters Ask for Refunds." Nation's Business 82: 27.

DiMaggio, P.J. and W. W. Powell. 1983. "The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields." American Sociological Review 48: 147-160.

Dunne, P. M., R. F. Lusch and D. A. Griffith. 2001. Retailing. Hinsdale, IL: Dryden.

Evans, K. R., R. E. Kleine, T. D. Landry and L. C. Crosby. 2000. "How First Impressions of a Customer Impact Effectiveness in an Initial Sales Encounter." Journal of the Academy of Marketing Science 28: 512-526.

Folkes, V. S. 1984. "Consumer Reactions to Product Failure: An Attributional Approach." Journal of Consumer Research 10: 398-409.

Fournier, S. and D. G. Mick. 1999. "Rediscovering Satisfaction." Journal of Marketing 63: 5-23.

George, B. P. and P. G. Hegde. 2004. "Employee Attitude toward Customers and Customer Care Challenges in Banks." International Journal of Bank Marketing 22: 390-406.

Hair, J. F., R. E. Anderson, R. L. Tatham and W. C. Black. 1995. Multivariate Data Analysis with Readings, 4th Edition. Englewood Cliffs, NJ: Prentice Hall.

Hartline, M. D. and O. C. Ferrell. 1996. "The Management of Customer-Contact Service Employees: An Empirical Investigation." Journal of Marketing 60: 52-70.

Hoffman, K. D., S. W. Kelley and H. M. Rotalsky. 1995. "Tracking Service Failures and Employee Recovery Efforts." Journal of Services Marketing 9: 49-61.

Huseman, R. C., J. D. Hatfield and E. W. Miles. 1987. "A New Perspective on Equity Theory: The Equity Sensitivity Construct." Academy of Management Review 12: 222-234.

Industrial Relations Review & Report. 1994. "Misconduct: The Customer is Always Right." 15.

Joreskog, K. and D. Sorbom. 2005. LISREL 8.72 for Windows. New York, NY: Scientific Software International, Inc.

Kahn, R. L., D. M. Wolfe, R. P. Quinn, J. D. Snoek and R. A. Rosenthal. 1964. Occupational Stress: Studies in Role Conflict and Ambiguity. New York, NY: John Wiley and Sons.

Krapfel, R. E. 1988. "Customer Complaint and Salesperson Response: The Effect of the Communication Source." Journal of Retailing 64: 181-198.

Liden, R. C., S. J. Wayne and R. T. Sparrowe. 2000. "An Examination of the Mediating Role of Psychological Empowerment on the Relations between the Job, Interpersonal Relationships, and Work Outcomes." Journal of Applied Psychology 85: 407-416.

Lusch, R. F. and J. R. Brown. 1996. "Interdependence, Contracting, and Relational Behavior in Marketing Channels." Journal of Marketing 60: 19-38.

MacKenzie, S. B. and R.J. Lutz. 1989. "An Empirical Examination of the Structural Antecedents of Attitude Toward the Ad in an Advertising Pre-testing Context." Journal of Marketing 53: 48-61.

--, --, and G. E. Belch. 1986. "The Role of Attitude Toward the Ad as a Mediator of Advertising Effectiveness: A Test of Competing Explanations." Journal of Marketing Research 23: 130-143.

MacNeil, I. R. 1985. "Relational Contract: What We Do and Do Not Know." Wisconsin Law Review:. 483-525.

Maxham, J. G. III and R. G. Netemeyer. 2002. "Modeling Customer Perceptions of Complaint Handling Over Time: The Effects of Perceived Justice on Satisfaction and Intent." Journal of Retailing 78: 239-252.

Meyer, M. W. and L. G. Zucker. 1989. Permanently Failing Organizations. Newbury Park, CA: Sage.

Mitchell, A. A. and J. C. Olson. 1981. "Are Product Attributes the Only Mediator of Advertising Effects on Brand Attitude?" Journal of Marketing Research 18: 318-332.

Mittal, B. 1990. "The Relative Roles of Brand Beliefs and Attitude Toward the Ad as Mediators of Brand Attitude: A Second Look." Journal of Marketing Research 27: 209-219.

Morrison, E. W. and S. L. Robinson. 1997. "When Employees Feel Betrayed: A Model of How Psychological Contract Violation Develops." Academy of Management Review 22: 226-256.

Muehling, D. D. and R. N. Laczniak. 1988. "Advertising's Immediate and Delayed Influence on Brand Attitudes: Considerations across Message-Involvement Levels." Journal of Advertising 17: 23-34.

Neuborne, E. 1996. "Retail Return Fraud: Merchants Feel Used, Abused; Retailers are Fed Up, Clamping Down." USA Today (June 3), lb.

Nielsen, E. H. and M. V. H. Rao. 1996. "The Strategy-Legitimacy Nexus: A Thick Description." Academy of Management Review 12: 523-533.

Nunnally, J. C. 1978. Psychometric Theory, 2nd Edition. New York, NY: McGraw-Hill.

Oliver, R. L., R. T. Rust and S. Varki. 1997. "Customer Delight: Foundations, Findings and Managerial Insight." Journal of Retailing 73:311-336.

Ortony, A., G. Clore and A. Collins. 1988. The Cognitive Structure of Emotions. Cambridge, UK: Cambridge University Press.

Pfeffer, J. and G. Salancik. 1978. The External Control of Organizations: A Resource Dependence Perspective. New York, NY: Harper and Row.

Pugh, S. D. 2001. "Service With a Smile: Emotional Contagion in the Service Encounter." Academy of Management Journal 44:1018-1027.

Rafaeli, A. and R. I. Sutton. 1989. "The Expression of Emotion in Organizational Life." In Research in Organizational Behavior, Vol. 11. Eds. L. L. Cummings and B. M. Staw.

Reinartz, W.J. and V. Kumar. 2003. "The Impact of Customer Relationship Characteristics on Profitable Lifetime Duration." Journal of Marketing 67: 77-99.

Robinson, S. L. 1996. "Trust and Breach of the Psychological Contract." Administrative Science Quarterly 41: 574-599.

--and E. W. Morrison. 2000. "The Development of Psychological Contract Breach and Violation: A Longitudinal Study." Journal of Organizational Behavior 21: 525-546.

--and D. M. Rousseau. 1994. "Violating the Psychological Contract: Not the Exception but the Norm." Journal of Organizational Behavior 15: 245-259.

Rogers, D. S. and R. S. Tibben-Lembke. 1999. Going Backwards: Reverse Logistics Trends and Practices. Reno, NV: Reverse Logistics Executive Council.

Rogers, K., B. Guffey and A. Markle. 2005. "The Retail Loss Prevention Tool Bag." Security Info Watch (April 20).

Rosenbaum, M. S. and R. Kuntze. 2003. "The Relationship Between Anomie and Unethical Retail Disposition." Psychology and Marketing 12: 1067-1093.

Rousseau, D. M. 1989. "Psychological and Implied Contracts in Organizations." Employee Responsibilities and Rights Journal 2: 121-139.

Schuman, E. 2004. "The War Against Retail Return Abuses." eWeek.com. Ziff Davis Media Inc.

Scott, W. R. and J. W. Meyer. 1983. "The Organization of Societal Sectors." In Organizational Environments: Ritual and Rationality. Eds. W. R. Scott and J. W. Meyer. Newbury Park, CA: Sage Publications.

Shimp, T. 1981. "Attitude toward the Ad as a Mediator of Consumer Brand Choice." Journal of Advertising 10: 3-11.

Speights, D. and M. Hilinski. 2005. "Return Fraud and Abuse: How to Protect Profits." Texas A &M University Retailing Issues Letter 17: 1-5.

Spreitzer, G. M., M. A. Kizilos and S. W. Nason. 1997. "A Dimensional Analysis of the Relationship between Psychological Empowerment and Effectiveness, Satisfaction, and Strain." Journal of Management 23: 679-704.

Stahl, H. K., K. Matzler and H. H. Hinterhuber. 2003. "Linking Customer Lifetime Value with Shareholder Value." Industrial Marketing Management 32: 267-79.

Steinauer, J. M. 1997. "The Face of the Sale." Incentives 171: 48-50.

Stock, J. 1999. Development and Implementation of Reverse Logistics Programs. Oak Brook, IL: Council of Logistics Management.

Suchman, M. C. 1995. "Managing Legitimacy: Strategic and Institutional Approaches." Academy of Management Review 20: 571-610.

Tomlinson, A. 2002. "Being Nice Takes Its Toll." Canadian HR Reporter 15: 1-13.

Williamson, O. E. 1975. Markets and Hierarchies. New York, NY: The Free Press.

Zucker, L. G. 1987. "Institutional Theories of Organizations." Annual Review of Sociology 13: 443-644.

(1) Such as returning an obviously used item or one used for a special event; for example, one studied retailer cited the Monday immediately following Super Bowl Sunday as the most likely day of the year for the return of a large-screen television. Another noted that off-the-rack dresses were often returned just after high school prom season.

(2) The profile of the sample frame is relatively consistent with means and standard deviations of data collected from the National Retail Merchants Association, the National Retail Federation, and the Bureau of Labor Statistics.

Chad W. Autry

Assistant Professor of Supply Chain Management

Texas Christian University 

Donna J. Hill

Professor of Marketing

Bradley University 

Matthew O'Brien

Assistant Professor of Marketing

Bradley University

Table 1 Across-construct Measurement Validity Assessment

                                                               Average
             Standardized    Std.       t        Composite     Variance
Indicator      Loading      Error               Reliability   Extracted

Normative                                           N/A          N/A
Legitimacy
  NL1            0.89        0.20       **
Breach                                             0.70          0.47
  B1             0.85        0.10    14.91 *
  B2             0.77        0.10    13.01 *
  B3             0.31        0.14     4.46 *
Violation                                          0.79          0.56
  V1             0.73        0.11    12.20 *
  V2             0.67        0.11    10.97 *
  V3             0.84        0.11    14.74 *
Ac                                                 0.91          0.65
  ATT1           0.90        0.10    17.44 *
  ATT2           0.84        0.11    15.66 *
  ATT3           0.90        0.09    17.66 *
  ATT4           0.85        0.10    16.06 *
  ATT5           0.77        0.11    13.75 *
  ATT6           0.47        0.12     7.44 *

* p <.01

** Single indicator fixed to reliability of 0.80.

Table 2
Model Variables, Descriptive Statistics and Correlations

(a) Variable summated and averaged where applicable.

                                               Correlations

     Model                 Standard       1         2         3
   Variables    Mean (a)   Deviation

1. Cognitive      3.54       2.37       1.0
   Legitimacy
2. Normative      3.97       2.22       0.560     1.0
   Legitimacy
3. Regulative     3.18       2.38       0.236     0.532     1.0
   Legitimacy
4. Breach         2.85       1.42      -0.241    -0.403    -0.309
5. Violation      3.22       1.54      -0.246    -0.336    -0.263
6. Ac             4.21       1.56       0.592     0.612     0.326

                        Correlations

     Model         4         5         6
   Variables

1. Cognitive
   Legitimacy
2. Normative
   Legitimacy
3. Regulative
   Legitimacy
4. Breach        1.0
5. Violation     0.621     1.0
6. Ac           -0.405    -0.447    1.0

Table 3
Estimated Structural Parameters of Proposed Model with
Normative Legitimacy as Antecedent

                                                          Attitude
                                                           Toward
Independent                   Breach       Violation     Customer
Latent Variable              ([eta] 1)     ([eta] 2)    (Ac)([eta] 3)

Normative Legitimacy         -0.45 (a)        --            0.62
  ([epsilon] 1)             (-5.99) (b)                    -9.44
Breach ([eta] 1)                --           0.92            --
                                            -15.93
Violation ([eta] 2)             --            --           -0.25
                                                          (-4.05)
Goodness of Fit Measures
Chi-square = 197.95         GFI   = 0.88   NFI  = 0.89
df = 83                     CFI   = 0.94   PNFI = 0.95
P = 0.0                     RMSEA = 0.07   NNFI = 0.94
Chi-square/df = 2.38        IFI   = 0.94   RFI  = 0.90

(a) Completely standardized solution

(b) Numbers in parentheses are t values (two-tailed).

All paths significant at the 0.05 level (directional).