The effects of distributive, procedural, and interactional justice on postcomplaint behavior

The effects of distributive, procedural, and interactional justice on postcomplaint behavior

1The Effects of Distributive, Procedural, and ~ Interactional Justice on Postcomplaint Behavior JEFFREY G. BLODGETT University of Mississippi DONNA J...

2MB Sizes 60 Downloads 173 Views

1The Effects of Distributive, Procedural, and ~ Interactional Justice on Postcomplaint Behavior

JEFFREY G. BLODGETT University of Mississippi DONNA J. HILL i Bradley University

STEPHEN S. TAX University

of Victoria

This study examines the effects of distributive,

interactional,

and procedural justice on complainants’

repatronage and negative word-of-mouth intentions. In order to test the conducted; twelve different scenarios were created, each describing a was returning a product to a retail store. In the various scenarios. exchange the product or was given a partial discount on a new product,

hypotheses an experiment was situation in which a customer the customer was allowed to was treated either with tour-

tesy and respect or was treated rudely, and was able to take care of the problem promptly or wlas asked to come back to the store the next day. Subjects were asked to read one of the rwelve scenarios and to imagine that this situation happened to them, they were then asked to imagine how they would have felt and what they subsequently would have done. In order to make the scenarios more vivid, subjects then watched a videotape depicting the same event. Of the three dimensions, interactional justice had the largest impact on complainants repatronage and negative word-of-mouth intentions.

INTRODUCTION A number of factors have made it necessary for retailers to focus greater attention on successfully resolving customer complaints. Increasingly competitive markets point to the importance of preserving loyalty and developing long-term relationships with customers. Furthermore, there is evidence that long-time customers are more profitable because they tend to purchase in greater quantity and more frequently than new customers (Reichheld and Sasser, 1990). Understanding the cost of customer defections and implementing strat-

Jeffrey G. Blodgett, Department of Management and Marketing, School of Business Administration, The University of Mississippi, University, MS 38677; Donna J. Hill, Department of Marketing, Foster College of Business Administration, Bradley University, Peoria, IL 61625; and Stephen S. Tax, School of Business, University of Victoria, Victoria, British Columbia, Canada, VSW 3Pl. Journal of Retailing, Volume 73(2), pp. 185-210, ISSN: 0022-4359 Copyright 0 1997 by New York University. Ail rights of reproduction

185

in any form reserved.

186

Journal of Retailing

Vol. 73, No. 2 1997

egies to lessen their occurrence has been a key element of many quality improvement programs (e.g., Federal Express and Chrysler). Its central role in maintaining customer loyalty clearly positions complaint handling as an important strategic tool for retailers. Many retailers recognize that complaints represent an opportunity to remedy product or service related problems and to positively influence subsequent customer behavior. There is considerable evidence that dealing effectively with complaints can have a dramatic impact on customers’ evaluations of retail experiences (Bitner, Booms, and Tetreault, 1990; Kelley, Hoffman, and Davis, 1993; McCollough and Bharadwaj, 1992) as well as enhance their likelihood of repurchase and limit the spread of damaging negative word-ofmouth (Blodgett, Granbois, and Walters, 1993; Gilly and Gelb, 1982; TARP, 1986). Unfortunately, many customers who complain end up feeling more negative about the business because of the way their problems are addressed {Hart, Heskett, and Sasser, 1990). Since complaint management can have considerable influence on retail sales and profitability (Pomell and Wernerfelt, 19X7), this topic certainly is worthy of further research. Despite the compelling evidence linking complaint handling to subsequent purchase beha~~ior, relatively little progress has been made in de~~e~opin~ a theoretical a~derstanding of how consumers evaluate retailers ’ responses to their complaints. One construct that has shown promise in explaining individuals’ reactions to a variety of conjlict situations.is perceivedjustice. Perceived justice is actually a broad, multifaceted construct, encompassing three di~lensions: dist~butive justice, infractions justice, and procedural justice (Alexander and Ruderman, 1987; Bies and Shapiro, 1987; Clemmer and Schneider, 1996). In a consumer complaint context, distributive justice refers to the perceived fairness of the remedy offered by the seller. Proceduml justice encompasses the perceived fairness of the policies and procedures used by the seller in processing a complaint, while interactional justice focuses on the manner in which the complainant was treated (Blodgett et al., 1993). The concept of perceived justice offers a valuable framework for explaining customers’ reactions to complaint episodes. It is also consistent with the service marketing literature which recognizes the importance of process and interpersonal communication during the service encounter (Bitner et al., 1990; Solomon, Surprenant, Czepiel, and Gutman, 1985). The purpose of this study is to further examine the linkage between perceived justice and postcomplaint behavior. More specifically, this study will assess the effects of dist~butive, interactional, and procedural justice on complainants’ negative word-of-mouth and repatronage intentions. The current research is unique in that we simultaneously investigate the effects of all three dimensions of justice; by doing so, we can determine the relative importance of each dimension. In order to test the hypotheses we conducted a qu~i-expe~ment, in which we manipulated the three dimensions of justice. This study will be of value to researchers who are developing causal models of complaining behavior and to retail managers who are responsible for developing and implementing complaint handling policies and procedures (and who must train retail employees how to effectively interact with complainants). By understanding the impact of each dimension of justice on postcomplaint behavior retailers may be able to focus more precisely on the critical aspects of complaint m~agement. This info~ation, in turn, should help retailers develop more effective and cost efficient methods for resolving customer disputes, thus resulting in higher levels of customer retention and profits.

Postcomplaint

187

Behavior

THEORETICAL

Complaining

DEVELOPMENT

Behavior

Complaining behavior may be viewed in terms of a set of possible customer responses to dissatisfying purchase experiences (Singh, 1988; Richins, 1983). In general, complaint options include seeking redress (i.e., a refund, exchange, repair, or apology, etc.), engaging in negative word-of-mouth (i.e., telling other people about one’s dissatisfaction), exiting (i.e., vowing never to repatronize the seller), and contacting third parties (e.g., the Better Business Bureau, writing a letter to a newspaper, taking legal action, etc.). These options are not mutually exclusive and any dissatisfied customer may engage in multiple responses. Blodgett et al. (1993) describe complaint behavior as a dynamic process, in which one’s initial response(s) (i.e., whether or not one seeks redress or exits, and whether or not one engages in negative word-of-mouth) is based upon factors such as the likelihood of success, one’s attitude toward complaining, and the importance of the product. However, once a dissatisfied customer seeks redress, subsequent word-of-mouth behavior (both negative and positive), repatronage intentions, and third party complaints are primarily dependent upon the complainant’s perception of justice. There is considerable evidence that if a firm handles complaints well it tends to reduce the incidence of negative word-of-mouth and third party complaints, and increases the likelihood of repurchase (Blodgett et al., 1993; Kelly et al., 1993; McCollough and Bharadwaj, 1992). Importantly, complainants may subsequently engage in positive word-of-mouth (thus spreading goodwill), and may even become more loyal customers, if they perceive a high degree of justice. Despite the empirical evidence linking complaint handling with subsequent customer responses, limited effort has been expended in developing a theoretical understanding of how different facets of justice affect consumers’ postcomplaint behavior.

Theoretical

Foundations

of Complaining

Behavior

Several authors have noted that there is no single, comprehensive theory of complaining behavior (Blodgett et al., 1993; Goodwin and Ross, 1992; Kelly and Davis, 1994). Rather, “ . . the study of complaining behavior is based upon several different theories from various fields of study” (Blodgett et al., 1993, p. 402). The confirmation/disconfirmation paradigm (Oliver, 1980), along with research on satisfaction/dissatisfaction (see Day, 1984; Day, Grabicke, Schaetzle, and Staubach, 198 1); Hirschman’s (1970) theory of exit, voice, and loyalty; and attribution theory (Folkes, 1984) all help to explain why some dissatisfied consumers seek redress (and thus give the retailer a chance to remedy the problem), while others silently exit, vowing never to shop there again. While these theories provide the theoretical foundation which allows us to better understand dissatisfied consumers’ initial complaining behavior(s), it is the concept of fairness (Clemmer and Schneider, 1996; Greenberg, 1990; Lind and Tyler, 1988) that provides the theoretical framework for the study of dissatisfied consumers’ postcomplaint behavior(s).

188

Journal of Retailing

Vol. 73, No. 2 1997

The concept of fairness, or justice, has its foundations in social psychology, and has been widely and successfully employed to explain individuals’ reactions to a variety of conflict situations. Fairness principles have been applied in the context of pay raises (Folger and Konovsky, 1989), hiring and promotion decisions, labor relations (see Greenberg, 1990), legal settings (Thibaut and Walker, 1975), and more recently in buyer/seller transactions (Clemmer, 1993; Oliver and Swan, 1989). These studies have demonstrated that fairness has both psychological (e.g., satisfaction, loyalty, trust) and behavioral outcomes (e.g., employee turnover, legal appeals, repurchase intentions). In a recent review article covering 30 years of research Clemmer and Schneider (1996) concluded that there are three dimensions of justice: distributive, procedural, and interactional. The distributive dimension focuses on the perceived fairness of outcomes, the procedural component reflects the fairness of the policies and procedures by which the outcome is produced, and the interactional component focuses on the interpersonal treatment people experience during the conflict resolution process. The next sections will explore each of these three dimensions of justice in more detail, and will discuss their applications in the context of consumer complaining behavior.

DistriSutive

Justice

Distributive justice refers to the perceived fairness of the tangible outcome of a dispute, negotiation, or decision involving two or more parties. The concept of distributive justice has its origins in social exchange theory (Blau, 1964; Adams, 1965), which emphasizes the role of equity in shaping subsequent exchanges. The equity principle defines a fair exchange as one in which each party to an exchange receives an outcome in proportion to one’s contributions to the exchange (Messick and Cook, 1983). Although by far the most pervasive, equity is not the only distributive rule that has been applied; two others are need and equality (Deutsch, 1985). Need refers to whether the outcome meets the requirements of the recipient, while equality demands that all parties the receive the same outcome regardless of contributions. Because of its anchoring in exchange theory, marketing studies have focused almost exclusively on equity as the appropriate distributive rule. Distributive justice issues are pervasive throughout society, existing in all situations where individuals or groups enter into exchanges (Deutsch, 1985). In a consumer complaint context, distributive justice centers on the perceived fairness of the redress offered to consumers to resolve their complaints. Types of redress include refunds, exchanges, repairs, discounts on future purchases, store credits, etc., or some combination thereof (see Kelly et al., 1993). It is important to recognize that perceptions of distributive justice rest with individual complainants and reflect their impressions of tangible outcomes. Several marketing studies provide insight into the effects of distributive justice. There is considerable evidence indicating that equity evaluations influence customer satisfaction, perceived service quality, and repurchase intentions (Fisk and Coney, 1982; Mowen and Grove, 1983; Oliver and DeSarbo, 1988; Oliver and Swan, 1989; Huppertz, Arenson, and Evans, 1978). Complaint handling incidents which are rated favorably usually include compensation in line with the perceived costs experienced by the customer (Kelly et al., 1993), thus supporting an equity-based evaluation of complaint outcomes.

Postcomplaint

Procedural

Behavior

189

justice

Procedural justice refers to the perceived fairness of the policies, procedures, and criteria used by decision makers in arriving at the outcome of a dispute or negotiation (Thibaut and Walker, 1975; Lind and Tyler, 1988; Alexander and Ruderman, 1987). Fair procedures are consistent, unbiased and impartial, representative of all parties’ interests, and are based on accurate information and on ethical standards (Leventhal, Karuza, and Fry, 1980). Fair procedures also allow focal parties to provide input into the decision (Greenberg and Folger, 1983; Goodwin and Ross, 1992). Studies in courtroom and organizational settings have provided support for the importance of these particular elements of procedural justice (Barrett-Howard and Tyler, 1986; Folger and Konovsky, 1989; Tyler, 1987). In a study of service encounter incidents, Clemmer (1993) also identified flexibility, waiting time/responsiveness, and efficiency as dimensions of procedural justice. These particular criteria reflect many concepts that also have been associated with customer satisfaction and service quality (see Parasuraman, Zeithaml, and Berry, 1985; Bitner et al., 1990; Hui and Bateson, 1991). While procedural justice appears to be a rather complex concept, a focal issue that appears particularly relevant to retail complaints is the speed with which the conflict is resolved (Hart et al., 1990; Kelley et al., 1993; Clemmer and Schneider, 1996). This dimension, in effect, is reflective of the timeliness, responsiveness, and convenience of the complaint handling process. A number of studies have cited the negative emotional states and resulting dissatisfaction caused by the perceived unfairness of waiting too long in service situations (Katz, Larson, and Larson, 1991; Venkatesan and Anderson, 1985). Time loss is seen as both aggravating and expensive to the customer (Maister, 1985). Waiting appears to be especially disconcerting when the customer is angry and uncertain to begin with, and believes that the seller has some control over the delay (Taylor, 1994). Such is likely to be the case in consumer complaint situations.

Interactional

justice

Interactional justice refers to the manner in which people are treated during the conflict resolution process; for example, with courtesy and respect or rudely (Bies and Moag, 1986; Bies and Shapiro, 1987). Previous studies, conducted across a variety of situations (e.g., service encounters, job performance evaluations, job recruitment) have identified a number of other elements associated with interactional justice, such as truthfulness, the provision of an explanation (Bies and Moag, 1986) politeness, friendliness, sensitivity, interest, honesty (Clemmer, 1993) empathy and assurance (Parasuraman et al., 19X5), directness and concern (Ulrich, 1984), and effort (Mohr, 1991). In a consumer complaint context, two other important factors are acceptance of blame (Goodwin and Ross, 1989) and the offering of an apology (Goodwin and Ross, 1992; Bies and Shapiro, 1987; Folkes, 1984). Because of the importance of communication in the resolution of complaints (Jacoby and Jaccard, 1981) the concept of interactional justice appears particularly relevant to understanding consumers’ postcomplaint behavior.

190

Journal

of Retailing Vol. 73, No. 2 1997

Limited empirical research has investigated the concept of interactional justice in a marketing context. In general, it has been found that fair interpersonal treatment contributes to satisfaction with service encounters (Bitner et al., 1990; Tyler, 1987), enhanced evaluations of service quality (Parasuraman et al., 1985), better overall evaluations of complaint handling (Goodwin and Ross, 1989,1992), and more favorable repurchase intentions (Blodgett and Tax, 1993).

Overall Perceptions

of justice

Although it is generally accepted that the three dimensions of justice are independent, it is the combination of these three dimensions that determines complainants’ overall perceptions of justice and hence their subsequent behavior. For example, Greenberg and McCarty (1990) note that giving constructive, informative comments when delivering a performance review makes a poor review more acceptable than when the same review is given in a rude manner and without an accompanying explanation. Likewise, providing explanations and mitigating circumstances strengthens the perceived fairness of layoffs (Brockner and Greenberg, 1990). Thus, it appears that high levels of interactional justice may offset (at least partially) lower levels of distributive justice. Similarly, higher levels of procedural justice may compensate for lower levels of distributive justice; for example, research in social psychology has shown that even if concerned parties do not receive the outcome they desire, they may still be satisfied with the overall result if they perceive that the policies and procedures used to determine the outcome were fair (Lind and Tyler, 1988). In a marketing context, Goodwin and Ross (1989) found that consumers were willing to return to an offending service provider when only a token remedy was offered, if they perceived that the procedures used to resolve the complaint were fair. It also appears that interactional justice and procedural justice may complement one another. For example, it has been demonstrated that judgments of procedural fairness are influenced by the adequacy with which formal decision making procedures are explained and by the manner in which the focal party is treated during the conflict resolution process (Bies and Moag, 1986; Tyler and Bies, 1989). In summary, it appears that complainants use a compensatory model to arrive at an overall, global perception of justice. Their overall perceptions of justice, in turn, directly influence their subsequent behaviors.

HYPOTHESES

AND RESEARCH QUESTIONS

This study will examine the effects of distributive, interactional, and procedural justice on complainants’ (i.e., those dissatisfied consumers who have sought redress) subsequent repatronage and negative word-of-mouth intentions. In order to test the various hypotheses an experiment was designed, in which we manipulated the levels of distributive, interactional, and procedural justice (the experiment will be described in detail in the methodology section). For purposes of this experiment, distributive justice was made operational by varying the amount of the exchange/discount offered to the complainant, while interac-

Postcomplaint

191

Behavior

tional justice was derived by varying the manner in which the complainant was treated (i.e., courteous vs. rudely). As previously mentioned, one aspect of procedural justice that is particularly relevant in a retail setting is the speed in which the complaint is resolved. Accordingly, procedural justice was made operational by varying the number of trips that the complainant had to make to the store in order to get the problem resolved. For the sake of parsimony, and to reduce potential confounds, other aspects of procedural justice (such as the store’s return policy) were held constant.’ It should be noted, then, that our measure of procedural justice is indeed more narrow than the broader conceptual definition, and that any subsequent findings pertain only to the timeliness aspect of the procedural construct. Theory and previous research (e.g., Blodgett et al., 1993; Clemmer, 1993) indicate that higher levels of distributive, interactional, and procedural justice (i.e., timeliness) will lead to more favorable repatronage intentions and a decreased likelihood of negative word-ofmouth. We therefore hypothesize that: Hla:

Distributive justice will have a positive effect on complainants’ tronage intentions.

repa-

Hlb:

Distributive justice will have a negative effect on complainants’ ative word-of-mouth intentions.

neg-

H2a:

Interactional justice will have a positive repatronage intentions.

H2b:

Interactional justice will have a negative effect on complainants’ ative word-of-mouth intentions.

H3a:

Procedural justice (i.e., timeliness) will have a positive effect on complainants’ repatronage intentions.

H3b:

Procedural justice (i.e., timeliness) will have a negative complainants’ negative word-of-mouth intentions.

effect on complainants’

neg-

effect on

In addition to these hypotheses other relevant research questions are: (1) Which dimension of justice has the largest impact on postcomplaint intentions?, (2) Are there any interactions among the three dimensions of justice, and if so, what is the nature of these interactions?, and (3) Which combinations of distributive, procedural, and interactional justice result in the most favorable postcomplaint intentions, and which combinations result in the least favorable postcomplaint intentions? Answers to these exploratory research questions would allow retailers to develop strategies to more effectively respond to customer complaints, and thus increase long-term sales and profits. These answers would also allow researchers to advance their understanding of fairness/justice theory, and to more precisely model the consumer complaining behavior process. We would like to point out that the exploratory nature of these research questions does not allow for specific, directional hypotheses.2 As Clemmer and Schneider (1996, p. 111) note, “To date, there has been little research on the relative contribution of [distributive] procedural and interactional fairness.. . .” Given the general lack of knowledge regarding these issues, an examination of these exploratory research questions could provide retail

192

Journal of Retailing Vol. 73, No. 2 1997

managers and marketing researchers alike with a greater understanding of the relationships between the three dimensions of justice and consumer complaint behavior.

METHODOLOGY

Experimental

Design and Procedure

In order to test the hypotheses a 3 x 2 x 2 between groups, quasi-experimental design (Cook and Campbell, 1979) was employed, with three levels of distributive justice (high, medium, and low), two levels of interactional justice (high and low), and two levels of procedural justice (high and low). Subjects were first asked to read a scenario describing a situation in which a customer was dissatisfied with a product (a pair of athletic shoes) and subsequently sought redress from the retailer. The customer had worn the shoes for only a couple of months but they were already starting to fall apart. The customer did not have a receipt, but had purchased several other pairs of shoes at that particular retail store previously. In the different scenarios, the dissatisfied customer was offered a full exchange (high distributive justice), a 50% discount on another pair of shoes (medium distributive justice), or a 15% discount on a new pair of shoes (low distributive justice). The customer was treated either with courtesy and respect and was offered a sincere apology (high interactional justice) or was treated rudely (low interactional justice). In all twelve scenarios the salesperson referred the customer to the store manager; in one-half of the scenarios the manager was in and was immediately available (high procedural justice), while in the other scenarios the customer was informed that the manager was not in and was asked to return the next day (low procedural justice). Prior to reading the scenario subjects were asked to imagine that this situation actually happened to them, and to imagine how they would have felt and what they subsequently would have done. (See the Appendix for a sample of the different scenarios.) In order to make the scenarios more vivid, and to better communicate some of the more subtle aspects of the service encounter (such as tone of voice, body language, facial expressions, etc.), subjects were then shown a video depicting the same event.3 This procedure also allowed subjects more time to reflect upon the situation, and to experience the same thoughts and emotions that a customer would naturally experience in a similar situation. After watching the video, subjects were given a questionnaire asking about their repatronage and negative word-of-mouth intentions (i.e., assuming that they had actually experienced the described event). An experimental approach was utilized in order to better control the independent variables of interest, and to be able to rule out extraneous variables (Cook and Campbell, 1979). A pair of athletic shoes was chosen as the focal product because it is one that most everyone is familiar with and has purchased at one time or another, either for themselves or for a family member.4 In order to create a certain amount of uncertainty as to the responsibi.lity of the retailer, and thus be better able to manipulate the three dimensions of justice, a situation was described in which product failure occurred after the customer had been using the prod-

Postcomplaint

Behavior

193

uct for a relatively short period and did not have a receipt. (Subjects were told that under present circumstances the shoes should have lasted for approximately one year.) This type of “nonroutine” situation increases the potential for customer/retailer conflict, thus highlighting the need for a greater understanding of the concepts of distributive, interactional, and procedural justice.

Pretesting

The twelve scenarios were developed based on extensive pretesting. We first tested a number of written scenarios to identify levels of redress that appropriately represented high, medium, and low levels of distributive justice, and to determine whether the high and low levels of interactional and procedural justice (i.e., timeliness) were being perceived accordingly. After reading a scenario pretest subjects were given a short questionnaire containing items designed to assess the validity of the manipulations. Subjects were then debriefed and were asked to verbally comment on their perceptions of justice and on the validity of the scenario. Based on these pretests the scenarios were modified and tested again. This process continued through several iterations, until we were reasonably certain that the manipulations possessed both discriminant and convergent validity. Once the written versions of these scenarios were judged satisfactory, we then filmed “rough’ versions of the scenarios. These filmed versions were then pretested using a sample of full- and part-time evening MBA students, and with staff members at two universities. As with the written versions, these subjects responded to a series of items assessing the validity of the manipulations. After completing the pretest questionnaire, subjects were also asked to comment on the believability and vividness of the scenarios. Based on these pretests additional modifications were made to the scripts. The final set of videotaped scenarios was tilmed in an actual retail store with three theater majors playing the roles of customer, salesperson, and store manager. The scenarios were professionally filmed, edited, and narrated by full-time staff members at a university teleproductions/public television department. The scenarios were edited so that everything was held constant except for the manipulations. At the beginning of each scenario, the narrator reminded subjects to imagine that the situation actually happened to them, and to imagine how they would have felt. Post-experiment interviews with subjects indicated that the actors were credible, the videos were of professional quality, and the scenarios were realistic.

Subjects

Data were collected via a self-report questionnaire, at three different locations: a midsouth university, a midwestern university, and a Canadian university. Subjects were recruited from faculty and staff employees at each university and from local church groups, and were given $5 in exchange for their participation. Each subject was randomly assigned to one of the 12 scenarios. At the midsouth and midwestern universities each subject read the written version of the scenario, watched the videotape, and was administered the ques-

194

Journal of Retailing

Vol. 73, No. 2 1997

tionnaire on an individual basis. In the Canadian sample the experiment was conducted in small groups of five people or fewer (again, subjects read the written scenario first, and then were shown the videotaped version).” A total of 265 useable responses were collected, resulting in cell sizes ranging from 20 to 23. Of the 265 respondents, 61% were female and 39% were male. Eighty-six percent of the respondents were white, 10% were African-American, and 4% were from other minority groups. Twenty-seven percent of the respondents classified themselves as clerical workers, 32% held professional jobs, 13% worked in white collar occupations, and 28% worked in blue collar or skilled trades occupations. Respondents were fairly evenly divided between the different age levels; 15% were between the ages of 18-24, 30% were between 25-34, 24% were in the 35-44 age group, 27% were age 45-64, and 4% were age 65 or older. Sixteen percent of the respondents reported that their highest educational level was high school, 28% had attended some college, 27% had completed college, and 29% had attended graduate school. Nineteen percent of the respondents had household income of less than $20,000,26% reported household income between $20,000-$34,999,15% earned between $35,000-$44,999, 20% earned $45,000-$64,999, 15% earned $65,000-$99,999, and 5% reported income in excess of $100,000.

Measures

Multiple item scales were created to measure each construct. The items were based on previous research (e.g., Blodgett et al., 1993; Singh, 1988; Day et al., 1981), and were modified somewhat to better fit the context of the study. All of the items were measured using 7-point Likert-type scales; most of the items were anchored with “strongly agree/strongly disagree”, while others were anchored with descriptors such as “very likely/very unlikely” or “more than expected/less than expected.” Prior to conducting the main experiment the questionnaire was extensively pretested. Based on this pretesting many of the items were slightly reworded and the scales were refined. Subjects’ repatronage intentions and negative word-of-mouth intentions were each measured with three items. The resulting scales were highly reliable; Cronbach’s alpha for repatronage intentions was .91 and for negative wordof-mouth intentions alpha equaled .87. In order to perform manipulation checks and to rule out potential confounds the questionnaire also contained items measuring the three different dimensions of justice and subjects’ attitudes toward complaining. Distributive justice and interactional justice were each measured with four items, resulting in alphas of .92 and .95, respectively. Procedural justice and attitude toward complaining were each measured with three items, with alphas equaling .85 and .75. (See Table 1 for a listing of all the items.)

Manipulation

Checks: Convergent

and Discriminant

Validity

In order to assess the convergent and discriminant validity of the three dimensions of justice manipulation checks were performed (see Cook and Campbell, 1979; Perdue and Summers, 1986). In this study, convergent validity would be established if it is shown that

Postcomplaint

195

Behavior

TABLE 1

List of Items for Each Construct (Including

Means,

Standard

Deviations,

and Cronbach

Repatronage Intentions l

Alphas)

Mean

S.D.

Alpha

3.30

1.92

.91

4.26

1.90

.87

3.47

1.82

.92

4.31

1.93

.85

3.95

2.28

.95

4.79

1.71

.78

What is the likelihood that you would shop at this retail store in the future?’

l

If this situation had happened to me

I would

never shop at this store

again. l

If this had happened to me

I would

still shop at this store in the

future.

Negative Word-of-Mouth l

Intentions

How likely would you be to warn your friends and relatives not to shop at this retail store!’

l

If this had happened to me

I would

complain to my friends and

relatives about this store. l

If this had happened to me I would make sure to tell my friends and relatives not to shop at this store.

Distributive Justice* l

Compared to what you expected, the “discount” offered was:l

l

Taking everything into consideration,

the manager’s offer was quite

fair. l

The customer did not get was deserved (i.e., regarding a refund or exchange).

l

Given the circumstances,

I feel

that the retailer offered adequate

compensation.

Procedural Justice* The customer’s complaint was handled in a very timely manner. l

l

The customer’s complaint was not resolved as quickly as it should

l

The customer had to make too many trips to the store in order to

have been. resolve the problem.

Interactional Justice* l

The customer was treated with courtesy and respect.

l

The employees seemed to care about the customer.

l

The employees listened politely to what the customer had to say.

l

I feel that the customer was treated rudely.

Attitude Toward Complaining* l

I am usually reluctant to complain to a store regardless of how had a product is.

l

I

am less likely to return an unsatisfactory product than are most

people l

I know.

If a defective produr-t is inexpensive

I usually keep it rather than ask

for a refund, or to exchange it. Nole~: All items were measured Unless otherwise

on 7.point

indicated,

‘Thl5 item was anchored

scales.

all items were

anchored

with

“very unlikcly/vcry

IThIT item was wchored

with

“lcs

These measures were

included

with

“strongly disagree/strongly

agree.”

likely.”

than expected/more

on the survey in order

than expected.” to perform

manipulation

and confounding

checks.

196

Journal of Retailing

Vol. 73, No. 2 1997

subjects did indeed perceive significant differences between each level of the various experimental conditions (i.e., between high, medium, and low distributive justice; high and low interactional justice; and high and low procedural justice). At the same time, discriminant validity would be established if it is shown that none of the manipulations were confounded by one another. Three 3 x 2 x 2 ANOVAs (with distributive, interactional, and procedural justice, respectively, as the dependent variable) were used to assess convergent validity. (See Table 2.) The first ANOVA revealed that the level of distributive justice was significantly different across the high (i = 4.99), medium (X = 3.29), and low (X = 2.30) distributive conditions (F2,253 = 76.83, p = .OOO), as expected. Furthermore, a Newman-Keuls test showed that each of the three group means was significantly different from one another. The second ANOVA revealed that subjects exposed to the high interactional condition did indeed perceive higher levels of interactional justice (X = 5.99) than did those subjects exposed to the low interactional justice condition (i = 1.96; F1,253 = 992.76, p = .OOO).Likewise, the third ANOVA showed that the level of procedural justice was significantly different across the high (X = 5.45) and low (X = 3.24) procedural conditions (F1,2s3 = 194.09, p = .OOO).These results indicate that subjects did indeed perceive significant differences between each level of the various experimental conditions, thus establishing the convergent validity of the manipulations. The three ANOVAs also allowed us to assess the discriminant validity of the manipulations. For the most part, the manipulations were “clean”; however, a slight amount of confounding was present. (See Table 2.) As expected, subjects’ perceptions of interactional and procedural justice were not affected by the level of distributive justice. Likewise, subjects’ ratings of distributive and interactional justice were not affected by the procedural manipulation. However, subjects’ perceptions of both distributive and procedural justice were affected somewhat by the level of interactional justice. Subjects who were exposed to the high interactional (i.e., courteous) condition perceived higher levels of both distributive and procedural justice (Xdist = 3.92, $,.,, = 5.19) than did subjects in the low interactional (i.e., rude) condition (Xdjsl = 3.13, .?,,,, = 3.53). When this type of confounding occurs Perdue and Summers (1986, p. 323) suggest that further analysis be done to determine whether the degree of confounding is severe enough to “impair an unambiguous evaluation of the results of the main experiment.” They state that the appropriate indicator to assess the degree of confounding is w2. Calculation of o2 indicated that the distributive manipulation accounted for 35.07% of the variance of distributive justice, while the interactional manipulation accounted for only 4.52%. In the case of procedural justice, the procedural manipulation accounted for 33.92% of the variance of procedural justice, while the interactional manipulation accounted for 19.78% of the variance. Clearly, the interactional manipulation had only a minor effect on distributive justice, and only a moderate effect on procedural justice. Based on these findings, we feel that the degree of confounding is such that the results of the main experiment can be interpreted in a straight-forward manner. Finally, in order to rule out another potential confounding variable subjects were also asked about their attitudes toward complaining. (Differences in subjects’ attitudes toward complaining could be a plausible alternative hypotheses for any significant differences in their subsequent repatronage and negative word-of-mouth intentions.) A 3 x 2 x 2 ANOVA revealed that subjects’ attitudes toward complaining were not significantly different across

Behavior

Postcomplaint

197

TABLE2 Manipulation

High

Check

Variable Dependent

Variable:

Distributive Justice Interactional Justice Procedural Justice

Distributive

Manipulation Confounding

Dependent

Interactions1 Justice Procedural Justice

Interactional

Confounding Manipulation

Dependent

Variable:

Distributive Justice Interactional

Justice

Procedural

Confounding Confounding

2.30 (1.40)

3.92

3.13

(1.91)

(1.72) 3.44 -

3.99

4.06

3.85

(2.35)

(2.05)

(1.05)

-

(1.09)

4.03

-

Dependent

Variable:

Distributive Justice

1.96

Interactional

Justice

Attitude

4.38

4.42

4.26

(1.95)

(1.88)

~

(1.75)

-

(1.77)

Confounding

‘d.f.

= 2,253

‘d.f.

= 1,253

4.52%

20.67] 1.142

,287

0.00%

1.44’

,238

0.00%

992.76*

.ooo

78.12%

3.492

,063

0.20%

0.72’

,488

0.00%

113.64*

,000

19.78%

3.24

194.092

,000

33.92%

1.32’

.269

0.25%

0.202

.654

0.00%

1.58*

,210

0.22%

Complaining 4.29

5.68

(1 .50)

(1.66)

(I .58)

-

(1.57)

-

(1.59)

4.53

4.45

4.62 (1.58)

Notes:

3.53

4.52

(1.61) Procedural Justice

3.90

(1.97)

Toward

Confounding

35.07%

(2.37)

5.45

Confounding

.ooo

Justice

(1.39) D.

02

(1.83)

(2.44)

5.19

Manipulation

P

Justice

(1.74) Procedural Justice

76.83’

3.29 (1.69)

(2.20) C.

F

4.99

5.99

Confounding

Low

(1.39)

3.60

Confounding

Variable:

Distributive Justice

Med

Justice

(1.89) B.

Checks

Type of

independent

A.

and Confounding

4.37

the various levels of distributive,

interactional, and procedural justice (see Table 2). This finding demonstrates homogeneity across groups and thus provides additional evidence as to the internal validity of the experiment (Cook and Campbell, 1979).

RESULTS

A 3 x 2 x 2 full factorial MANOVA was used to assess the effects of distributive, interactional, and procedural justice on complainants’ repatronage and negative word-of-mouth intentions. A summary of all cell and group means is given in Table 3. All main effects and interactions were tested for significance. See Table 4 for MANOVA summary statistics.

Lfl =N 8Z’P = X 9E’E = x

!H

CL9= “) OO’f = x 9s.t = :

0-I

(P9 = “) f9.S = x 1L.Z =;

JJ!lSll(

/e"O!PPJa]",

Pt‘l = N PL’P =

LE’E =

x x

(69 = “) 91’S = x 61.2 = i cs9 = “1 EO’E = -Y S9’P = ?

-ualui qlnow-Jo-plom a~ye5au pur? aSmo1leda1 ,waCqns uo lceduq ~saXnq aql p~q aysnr 30 uoy~auup qqq~ au!uualap pm ‘uopsanb qsmasa~ droleloldxa ls.~g aql lsalol IapJo UI .pamafa.~ a.ra~ argsn~ lempaDold ~u~.I~xIo:, sasaqlodLq 0~1 aql aI!qM ‘pavoddns alarn asgsn[ ~euo!l~~a~u! pur! agnq!Ils!p %up~8a_t sasaqlodLq mo3 aql ‘Lmu -urns UI ‘(28~’ = d ‘PE’I = ZiZZJ SyI!M) SUO!lUalU~ qlUOUI-JO-p.IOM Z4~ldkXl Jraql UO .lOll suogualu! a%suo.uedaJ ,wa[qns uo lcm33a uyu 1um3~u2Qs e amq lou p!p (ssaugatug ‘.a.!) mlsnfpmpa~o.Id ‘X~Ouypd.ms ‘(000’ = d ‘~8.6~1 = EG‘~d) suoyualu! qlnour-30-pIorlh a12!lv -%?aU_I!aql uo 13a33a aA!]v%aU e pue (()()()’= d '~5361= tsz‘rd) suoyualu! a%mogeda~ ,sl3a[ -qns uo 13a33a aA!l!sod 13peq acysn[ [euogmalu! ieqi pap?aAaJ slsai ay.n+un :(oo()’ = d ‘6p’gj = Zsz‘Zd q!&) suoy~aiu~ qmow3o-proh ah@!au pue a%mo.wda~ ‘slDa[qns uo ~1~3y@3;rs I? peq oslr! acgsn[ [euopmaluI ‘(000’ = d ‘o~y,l = EQ’ZJ) suoy.~a~u~ aAg&aU .I!aql uo 13aJja aAy?%au prre (000 = d ‘~5'81 = csz'zd) suo!lualu!

lDa33au!m

qlnow-Jo-pJOM

a%.Io.wdar ‘slDa[qns uo iDa33a aA!i!sod e peq aD!lsnf aylnqys!p leql paleaaal sasL@m alegen!un .(ooo. = d ‘OE.OI = @OS‘@ .g sq!~) suogualy qlnow30-prom a+leZau pm a2?c -uo.mdaJ ,slsa[qns uo va33a U+WI 1ue~~uB~s e peq acylsn[ aylnq~lsrp ‘paz!saqloddq sv L66 1 Z ‘ON ‘EL ‘1% ~U!I!w

10 1”U.w

861

Postcomplaint

199

Behavior

TABLE4 MANOVA Summary Statistics MainEffects & A.

Dependent Variable

heractions

Multivariate

d.f.

Wilks F

p-value

Statistics

Distributive Justice

RepatronageiNWOM

4,504

10.30

,000

Interactional

RepatronageiNWOM

2,252

95.49

,000

Procedural Justice

RepatronageiNWOM

2,252

1.34

,382

Dist by Inter

RepatronageiNWOM

4,504

4.70

,002

Dist by Proc

RepatronageiNWOM

4,504

0.79

,390

Inter by Proc

RepatronageiNWOM

2,252

1.86

,088

Dist by Inter by Proc

RepatronageiNWOM

4,504

0.26

,950

6.

Justice

Univariate

Statistics

Distributive Justice

Repatronage

2,253

18.52

,000

Distributive Justice

NWOM

2,253

14.30

,000

Interactional Justice

Repatronage

1,253

195.57

,000

Interactional Justice

NWOM

1,253

i 79.84

.OOO

Dist by Inter

Repatronage

2,253

8.99

,000

Dist by Inter

NWOM

2,253

4.82

,009

multivariate

test was significant.

MAP:

*Univariate

statistics are reported

nnly if thr mrresponding

tions, o2 was calculated. The results show that interactional justice explained 38.5% of the variance of subjects’ repatronage intentions and 37.5% of the variance of their negative word-of-mouth intentions, while distributive justice explained just 7% of the variance of repatronage and only 5.6% of the variance of negative word-of-mouth. These results clearly indicate that the major determinant of respondents’ repatronage and negative word-ofmouth intentions was interactional justice. In order to answer the second exploratory research question we tested for all possible interactions among the three dimensions of justice. Interestingly, there was a significant interaction between distributive and interactional justice (Wilks F4,504 = 4.70, p = .002). Univariate tests revealed that the interaction was significant both for repatronage intentions (F2,253 = 8.99, p = .OOO)and negative word-of-mouth intentions (F2,2s3 = 4.82, p = .009). In order to better understand the nature of this interaction the simple main effects were calculated. Analyses of these simple effects revealed that the main effect of distributive justice was significant only within the high interactional (i.e., courtesy and respect) condition; in the low interactional (i.e., rude) condition the main effect of distributive justice was nonsignificant. Calculation of w2 indicated that the interaction of distributive and interactional justice had only a minor effect on postcomplaint behavior, accounting for only 3.1% of the variance of subjects’ repatronage intentions and another 1.7% percent of the variance of negative word-of-mouth. (All other possible two-way and three-way interactions were nonsignificant; see Table 4.) Finally, to assess the third exploratory research question and determine which combinations of justice resulted in the most favorable postcomplaint intentions (and which combinations resulted in the least favorable intentions), a Newman-Kuels planned comparisons test was performed. Since procedural justice did not have any significant effects on postcomplaint intentions we collapsed the data accordingly, thus creating six different combi-

Journal of Retailing Vol. 73, No. 2 7997

200

nations of distributive and interactional justice. Table 5 rank-orders the six different combinations according to their mean scores for repatronage and negative word-of-mouth intentions, and indicates all significant differences between these means. The same pattern was found both for repa~onage and negative won-of-mouth in~ntions. Obviously, the most favorable combination was the “high distributive/high interactional” condition. More interesting, however, is the fact that subjects who were in the “medium distributive/high interactional” and “low distributive/high interactional” conditions reported higher repatronage intentions and lower negative word-of-mouth intentions than those subjects who were exposed to the “high dis~butive~ow interactional” condition. In other words, subjects who “received” either a 50% or 15% discount, and who were ‘“treated” with courtesy and respect, were more likely to repatronize the retailer and were less likely to engage in negative wordof-mouth than those subjects who “received” a full exchange but were ““treated”rudely. Table 5 also highlights the previously discussed interaction between distributive and interactional justice. The Newman-Kuels test revealed that subjects’ repatronage and negative word-of-mouth intentions were statistic~ly different across the “high dist~butive/ high interactional, ” “medium dis~butive~igh interactional,” and “low dist~butive~igh interactional” conditions; however, there were no significant differences across the *‘high distributive/low interactional, ” “medium distributive/low interactional,” and “low distributive/low interactional conditions.” It appears, then, that the low interactional manipulation created a “floor effect,” and that if subjects were “treated” rudely the amount of the exch~ge/~scount did not matter. Overall, these results lend further support for the earlier TABLE5

Newman-Kuels Planned Comparison Test for Differences in Means Across Distributive and Interactional Justice D;scribut~~e~ustice High

Experimental Condition

LOW

Med

Repat

Nwom

I

1) hi distihi inter 2) med d&z/hi inter

5.72 4.61

2.04 3.11

*

3) low dist/hi inter

3.4’)

3.89

*

* *

*

2

3

4) hi distJow inter

2.37

5.10

*

5) mud dist/low inter

2.12

5.49

*

*

*

6) low distilow inter

1.95

5.67

*

*

*

Notes:

The data was rollapsed An “*’

indicates

over pmredural

a significant

differences

pstice. between

The same results hold for both repatmnage

means,

and negative

hased on the Newman-Ku~ls word-of-mouth

intentions.

test.

4

5

Postcomplaint

201

Behavior

finding that the main determinant of complainants’ mouth intentions is interactional justice.

DISCUSSION

repatronage

and negative

word-of-

AND IMPLICATIONS

In general, these findings indicate that complainants who experience higher levels of distributive and interactional justice are more likely to repatronize the retailer and are less likely to engage in negative word-of-mouth behavior (and vice-versa). Contrary to expectations, we found that procedural justice (or more specifically, timeliness) had no effect on subjects’ repatronage intentions or on their negative word-of-mouth intentions. Although the high and low procedural justice conditions were perceived accordingly, having to come back the next day to talk to the manager apparently was of little consequence compared to the manner in which the complainant was treated and the amount of the exchange or discount that was offered. One possible explanation for this finding is that complainants may understand the need to speak to the store manager (especially in nonroutine situations), and realize that the manager cannot always be at the store. Hence, having to come back to the store the next day is not so unreasonable as to cause complainants to engage in negative word-of-mouth or to vow never to repatronize the retailer. The key finding of this study is that interactional justice explained a significantly greater percentage of the variance of subjects’ postcomplaint intentions than did distributive justice. This finding indicates that higher levels of interactional justice can compensate for lower levels of distributive justice. In other words, complainants may be willing to repatronize the retailer when only a partial refund, exchange, or discount is given, provided that they are treated with courtesy and respect. However, even a full exchange is not enough to overcome the ill will due to being treated rudely. Complainants who are treated rudely are more likely to vow never to shop there again (i.e., exit), and to warn others not to patronize the retailer, regardless of the amount of redress that is offered. These findings are consistent with service encounter research that has observed the importance of interpersonal elements in shaping customers’ satisfaction with service experiences (Bitner et al., 1990), and with other research that has demonstrated the impact of emotions on postcomplaint behavior (Westbrook, 1987). Another interesting finding concerns the interaction between distributive and interactional justice. As reported in the results section, there was a significant difference in the repatronage and negative word-of-mouth intentions-across the three levels of distributive justice-of those subjects who experienced a high degree of interactional justice. However, there was no significant difference in the repatronage and negative word-of-mouth intentions-across the three levels of distributive justice-of those subjects who were exposed to a low level of interactional justice. Based on these findings, it appears that complainants may use a two-stage decision making rule in determining their overall perceptions of justice and their subsequent responses. In this two-stage process interactional justice first acts as a “cutoff’ that determines whether the secondary criteria (i.e., distributive justice) is even taken into consideration. When complainants experience a high level of interactional justice their overall perceptions of justice, and their subsequent behaviors, are then based on

202

.Journal of Retailing Vol. 73, No. 2 1997

the amount of redress that was offered. However, if complainants experience a low level of interactional justice the level of distributive justice is not even taken into consideration, In other words, when complainants are treated rudely the level of redress offered by the retailer is irrelevant. It is only when complain~ts are treated with courtesy and respect that distributive justice has a positive effect on postcomplaint behavior. These results reinforce previous research emphasizing the importance of front-line employee attitude in retail service recovery (Bowen and Lawler, 1995). These findings point to the importance of training retail employees how to respond to customer complaints. Accordingly, we recommend that all retail employees, both full- and part-time, be provided with an in-depth unders~nding of the different dimensions of justice, particularly the interactional component. In particular, retail employees should be taught to always respond to customer complaints with courtesy and respect. Retail employees should be instructed to give the customer an opportunity to explain the problem, especially in nonroutine situations. Employees should also offer an apology, and thank the customer for bringing the problem to the retailer’s attention. Since interactional justice is the key factor that determines whether a complainant will repatronize the retailer (and possibly become a more loyal customer) or whether that person will exit and engage in negative word-of-mouth behavior, training programs that emphasize the importance of this dimension of justice can have a substantial impact on the long-term profitability of a retail firm. Retailers that do not give this topic the attention it deserves may unnecessarily be losing customers, and hence sales and profits. These findings also have implications for retailers’ return policies. Some retailers have very liberal return policies, guaranteeing satisfaction with “no questions asked.” However, a full refund or exchange may not always be deserved (or even expected). Interestingly, our findings indicate that retailers do not necessarily have to offer full refunds or exchanges to all complainants in order to guarantee satisfaction. Rather, in nonroutine situations (i.e., those in which the customer has used the product for a fair mount of time, has clearly abused the product, or is partially to blame for the problem, etc.), some retailers may want to ask complainants what they think would be a fair outcome, and to proceed accordingly (i.e., within reason). Instead of offering a full refund or exchange, retailers may be able to satisfy these complainants by providing some type of partial redress. Indeed, our findings show that if complainants are treated in a courteous manner they may be satisfied with this type of remedy. At the same time, retailers that do not offer any refunds or exchanges (except maybe under strict conditions) may want to rethink this policy. When complainants are treated with courtesy and respect even a partial refund, exchange, or discount can have a favorable impact on their repatronage behavior. Since the opportunity cost of losing a customer is many times greater than the cost of an exchange (Blodgett, Wakefield. and Barnes, 1995) these retailers would be wise to adopt a more flexible policy.

LIMITATIONS Because of the experimental nature of this study the findings should be inte~reted accordingly. Subjects were asked to read, and view, a complaint scenario and to imagine them-

203

Postcomplaint Behavior

selves in that situation; they were then asked as to their repatronage and negative word-ofmouth intentions. Although subjects’ comments indicated that the scenarios were very realistic, we cannot say with certainty that the same pattern of findings would emerge from a field study of subjects’ actual complaint experiences. Indeed, a field study would provide an interesting complement to the current study, and would provide additional insight as to the relative impact of each dimension of justice. Another limitation of this study concerns our measure of procedural justice. For the sake of parsimony, and to avoid possible confounding, the only aspect of procedural justice that was manipulated was timeliness. Even though we successfully manipulated this aspect of procedural justice, the manipulation may have been, in effect, somewhat weak. This dimension was probably less tangible than the distributive justice dimension, and most likely was less vivid than the interactional dimension. As a result, subjects may not have fully appreciated the inconvenience of having to return to the store the next day to resolve the problem. Had this manipulation been more vivid or more tangible the results might have been different. In retrospect, a field study is probably more appropriate to investigate this particular aspect of procedural justice. Furthermore, since this study focused solely on the timeliness aspect of procedural justice, the results do not apply to other dimensions of the construct. Considering that our measure of procedural justice was somewhat narrow, additional studies could broaden our understanding of this construct. Indeed, this construct presents a potentially fertile area of future research. Issues such as the degree to which employees should be empowered to resolve complaints), and whether store refund/exchange policies should be strictly enforced or whether they should be flexible, represent two such topics. Finally, because of the slight degree of confounding between distributive and interactional justice, some of the findings should be interpreted with caution. Although it is clear that interactional justice had the largest effect on subjects’ postcomplaint behavior, these effects may be slightly overstated, while the effects of distributive justice (and procedural justice) may be somewhat understated. At this point, a replication experiment would be of only minimal value; therefore, we recommend that a field study be conducted in order to more precisely ascertain the relative importance of each dimension of justice.

SUMMARY

Because of its impact on customer loyalty, and hence on long-term sales and profits, the concept of perceived justice has emerged as a critical element in retailers’ marketing programs. In order to effectively implement successful complaint handling policies and procedures retailers need to have a more complete understanding of the different dimensions of justice. Our results indicate that-even if a retail store has a liberal return policy-the endresult of a customer complaint could be disastrous if a retail employee acts rudely. Retailers that focus on interactional issues will have the greatest chances of building long-term relationships with their customers.

204

Journal of Retailing

APPENDIX:

Med Distributive

l

TWO

Hi Interactional

Vol. 73, No. 2 1997

EXAMPLES OF THE SCENARIOS

l

Hi Procedural

Imagine that you purchased a pair of moderately priced, name-brand athletic shoes a couple of months ago (for yourself, your spouse, or a son or daughter) and just recently noticed that they are already starting to fall apart. Under normal conditions the shoes should have lasted much longer (probably about a year); therefore, you decide to take the shoes back to the retailer and ask for a refund or an exchange. Upon entering the store you are greeted by a salesperson, who appears to be very friendly and helpful. “Hi, I’m Laurie,” she says, “How can I help you today?’ You reply, “Well, I bought these shoes a couple of months ago and they are already wearing out pretty badly.” You proceed to take the shoes out of the bag and show the salesperson. “These shoes should normally last about a year before they wear out. So, I would like to either get my money back or exchange them for another pair,” The salesperson takes a good look at the shoes, and says, sounding somewhat apologetic, “You are right, these shoes haven’t held up very well. I’m really sorry. They should last much longer than just a couple of months.” She then asks, very politely, “Do you happen to have your receipt with you?’ “No, I don’t,” you reply, “Since these are brand-name that they might wear out so quickly.”

shoes it did not even occur to me

“I understand,” says the salesperson, “I hardly ever save my receipts for more than a few days, either. I am going to need to get the manager though. If you will excuse me for a few seconds I will get Adam and let you speak with him.” “OK’, you reply. The salesperson then exits, and returns a few moments later with the manager. As the salesperson and the manager are approaching, the salesperson says, “Adam, this person is returning these shoes. He’s (or she’s) had them for only a couple of months and they are worn out already.” Adam very cordially introduces himself, “Hi, I’m Adam. How are you?” “I’m fine,” you respond. “Let me take a look at these shoes,” says Adam, pausing for a moment while looking at the shoes. “I really am very sorry that you had a problem with these shoes. They normally hold up very well, but these are definitely worn out.” “I agree,” you reply. “I have worn this brand many times before and have never had a problem. But for some reason this pair just did not last very long.” “I understand,” says Adam, very politely. “We really do want to take care of our customers, but unfortunately, our store policy is no cash refunds after thirty days, or without a receipt. However, I will be glad to give you a 50% discount on another pair of shoes. Would that be OK?’

Postcomplaint

205

Behavior

“I guess so.” “Good, we really do appreciate your business. Why don’t you look around for a while and see if there is anything you like. We’ve got a lot of new shoes in. If you want to try any on I’ll be glad to get them for you.”

Hi Distributive

l

Lo Interactional

l

Lo Procedural

Imagine that you purchased a pair of moderately priced, name-brand athletic shoes a couple of months ago (for yourself, your spouse, or a son or daughter) and just recently noticed that they are already starting to fall apart. Under normal conditions the shoes should have lasted much longer (probably about a year); therefore, you decide to take the shoes back to the retailer and ask for a refund or an exchange. Upon entering the store you are greeted by a salesperson, who appears to be very friendly and helpful. “Hi, I’m Laurie,” she says, “How can I help you today?’ You reply, “Well, I bought these shoes a couple of months ago and they are already wearing out pretty badly.” You proceed to take the shoes out of the bag and show the salesperson. “These shoes should normally last about a year before they wear out. So, I would like to either get my money back or exchange them for another pair.” At this point the salesperson’s attitude changes, as she responds very matter of factly: “Do you happen to have the receipt with you? “No, I don’t,” you reply. “Since these are brand-name that they might wear out so quickly.”

shoes it did not even occur to me

At this point the salesperson becomes a bit abrupt, and sounding irritated, says “Well, I’m sorry (but not really), but since you don’t have a receipt, and since you have been wearing the shoes, we cannot give you a refund or let you exchange the shoes.” You respond emphatically, “Look, I’ve purchased several pairs of shoes from this store in the past, and I’ve never had a pair of shoes wear out this quickly.” “I’m sorry,” says the salesperson, sounding quite sarcastic, “but our policy is that we need to have a receipt, otherwise we can’t even be sure that you bought the shoes at our store!” You are now somewhat angry. “But I just bought these shoes two months ago! I thought this store stood behind its products! Let me speak to the manager!” The salesperson replies, “I’m sorry,” sounding irritated. “But he’s not here right now. He’s left today already. You’ll just have to come back tomorrow.” “I’ve got to come in tomorrow?!” in?’

you say, sounding incredulous.

“What time will he be

“He’ll be in all day tomorrow. Just come in anytime,” says the salesperson, “OK, I’ll be back tomorrow,” “OK,” says the salesperson,

flippantly.

you respond, sounding very frustrated and angry. as the customer is leaving the store.

206

Journal of Retailing

Vol. 73, No. 2 1997

The next day you make a special trip back to the store. You park your car, grab the shoes, and walk to the storefront. As you enter the store you are greeted by the same salesperson, “Hi, can I help you?’ “Yes,” replies the customer. return these shoes and... “

“I need to speak to the manager. I came in yesterday

to

“Oh yeah. Wait here just a minute and 1’11get him,” says the salesperson, with a grim sounding voice. The salesperson then exits and returns a few moments later with the manager. As the salesperson and the manager are approaching, the salesperson says, in a sarcastic tone of voice, “Adam, this lady wants to return these shoes. She’s been wearing the shoes for a while and doesn’t have a receipt.” As the salesperson and manager are approaching, the salesperson says, in a sarcastic tone of voice, “Adam, this person wants to return these shoes. He (or she) has been wearing the shoes for a while and doesn’t have a receipt.” The manager greets you in a somewhat cold tone of voice, “Hi, I’m Adam. What can I do for you?” You proceed to explain what happened, “I bought these shoes from your store a couple of months ago, and they are falling apart already. They really should have lasted much longer. So, I would like a refund.” In a very condescending tone of voice, the manager replies, “I understand you don’t have a receipt. Is that right?’ “No, I don’t,” you reply, sounding irritated. “Since these are brand-name not even occur to me that they might wear out so quickly.”

shoes it did

The manager then coldly states, “ Well, our policy is that we cannot give any refunds after 30 days or without a receipt.” “What!” you exclaim, sounding very surprised. Then, in a very assertive you state, “Look, I’ve purchased several pairs of shoes from this store in have never had a problem, but for some reason this pair of shoes just did really think that you should give me a refund, or at least let me exchange

tone of voice the past and I not hold up. I the shoes!”

The manager reiterates, rudely, “Like 1 said, since you don’t have a receipt I can’t refund your money. That’s our policy.” You respond, very emphatically, give me my money back!”

“I can’t believe this! Any other store in town would

The manager then seems to get your message, and starts to back off a little. Still sounding a bit sarcastic the manager says, “OK. OK. We don’t normally do this, but .. . because of your inconvenience I guess I could let you exchange the shoes for another pair. Would that be OK?’ You start to reply, somewhat reluctantly, “Well, I . ....” when you are cut off by the manager. “Good,” he says, again in a condescending tone of voice. ‘Why don’t you look around for a while and see if there is anything you like. We’ve got a lot of new shoes in, if you want to try any on we’ll be glad to get them for you.”

Postcomplaint

207

Behavior

This research was supported, in part, by a grant provided Acknowledgment: Chapter of the University of Mississippi Alumni Association.

by the Business

NOTES

1. Although procedural justice is a separate and distinct dimension of justice, as a practical matter it can have implications for the distributive component. For example, whether a retailer has a liberal return policy or a restrictive policy may affect the amount of refund, discount, or exchange offered to the complainant. In order to ensure against this type of confound we decided to manipulate only the timeliness aspect of procedural justice, while holding all other aspects constant. We thus acknowledge that our findings are limited to only the timeliness aspect of the construct. 2. As noted, previous research does not allow us to posit a more specific set of hypotheses. Although prior research does indicate that the three dimensions of justice complement one another, and that it is the combination of these three dimensions that determines ones’ subsequent behaviors, it does not indicate whether one dimension of justice might have a larger impact than the others, nor does it discuss any possible interactions. Previous research is also too general in nature to allow us to predict that specific combinations of justice will result in more favorable postcomplaint behaviors than other combinations. 3. During pretesting, subjects’ remarks gave us the impression that the written versions were not nearly as vivid as a “real life” complaint episode; as a result, subjects did not become very emotionally involved. In order to ensure the external validity of the experiment we felt that it was necessary to make the scenarios “come to life,” hence, we created the videotaped versions. During subsequent pretesting subjects remarked that the videotaped versions did indeed make the scenarios more vivid and emotionally involving. Although they were more vivid, we chose not to show the videotaped versions only, just in case a subject missed out on something that was said, either by the customer, the salesperson, or the manager. This procedure also allowed subjects more time to internalize their thoughts and emotions. 4. In choosing a product we felt that we should use a product category that both males and females, and younger and older consumers, and so forth, could relate to. Pretests indicated that the athletic shoe product category was one that most everyone could relate to. More importantly, we do not feel that the product category had any effect on subjects’ responses because the focus of this study was on the effect of perceived justice on postcomplaint behavior. The product category may have implications for whether or not a dissatisfied consumer initially seeks redress; however, once a dissatisfied consumer seeks redress the focus is shifted off of the product and onto the three dimensions of justice. 5. This was done purely for administrative convenience; it simply was not possible to schedule these subjects on an individual basis. The experimenter ensured that there was no talking among subjects. A MANOVA revealed that “method of administration” (individual vs group) had no effect on subjects’ responses.

REFERENCES Adams, J.S. (1965). “Inequity in Social Change.” Pp. 267-299 in L. Berkowitz Experimental Social Psychology, Vol. 2. New York: Academic Press.

(ed.), Advances

in

208

Journal of Retailing

Vol. 73, No. 2 1997

Alexander, S. and M. Ruderman. (1987). “The Role of Procedural and Distributive Justice in Organizational Behavior,” Social Jusrice Research, 1: 177- 198. Barrett-Howard, E. and T.R. Tyler. (1986). “Procedural Justice as a Criterion in Allocation Decisions,” Journal of Personality and Social Psychology, 50: 296-304. Bies, R.J. and J.S. Moag. (1986). “Interactional Communication Criteria of Fairness.” Pp. 289-3 19 in R.J. Lewicki, B.H. Sheppard, and M.H. Bazerman (eds.), Research in Organizational Behavior. Greenwich, CT: JAI Press. Bies, R.J. and D.L. Shapiro. (1987). “Interactional Fairness Judgements: The Influence of Causal Accounts,” Social Justice Research, 1: 199-2 18. Bitner, M.J., B.M. Booms, and M.S. Tetreault. (1990). “The Service Encounter: Diagnosing Favorable and Unfavorable Incidents,” Journal ofMarketing, 54(January): 7 l-85. Blau, P.M. (1964). Exchange and Power in Social Life. New York: Wiley. Blodgett, J.G., D.H. Granbois, and R.G. Walters. (1993). “The Effects of Perceived Justice on Negative Word-of-Mouth and Repatronage Intentions,” Journal of Retailing, 69(Winter): 399428. Blodgett, J.G. and S.S. Tax. (1993). “The Effects of Distributive and Interactional Justice on Complainants’ Repatronage Intentions and Negative Word-of-Mouth Intentions,” Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior, 6: 100-I 10. Blodgett, J.G., K.L. Wakefield, and J.H. Barnes. (1995). “The Effects of Customer Service on Consumer Complaining Behavior,” Journal of Services Marketing,9(4): 3 l-42. Bowen, D.A. and E.E. Lawler. (1995). “Empowering Service Employees,” Sloan Management Review, 36(Summer): 73-84 Brockner, J. and J. Greenberg. (1990). “The Impact of Layoffs of Survivors: An Organizational Justice Perspective.” Pp. 45-75 in J. Carroll (ed.), Advances in Applied Social Psychology: Business Settings. Hillsdale, NJ: Erlbaum. Clemmer, E.C. (1993). “An Investigation into the Relationships of Justice and Customer Satisfaction with Services.” In R. Cropanzano (ed.), Justice in the Workplace: Approaching Fairness in Human Resources Management. Hillsdale, NJ: Erlbaum. Clemmer, E.C. and B. Schneider. (1996). “Fair Service.” Pp. 109-126 in S.W. Brown. D.A. Bowen, and T. Swartz (eds.), Advances in Services Marketing and Management, Vol. 5. Greenwich, CT: JAI Press. Cook, T.D. and D.T. Campbell. (1979). Quasi-Experimentation: Design & Analysis Zssuesfor Field Settings. Boston: Houghton Mifflin Company. Day, R.L. (1984). “Modeling Choices Among Alternative Responses to Dissatisfaction.” Pp. 496499 in T.C. Kinnear (ed.), Advances in Consumer Research, Vol. 1 I. Ann Arbor, MI: Association for Consumer Research. Day, R.L., K. Grabicke, T. Schaetzle, and F. Staubach. (1981). “The Hidden Agenda of Consumer Complaining,” Journal of Retailing, 57(Fall): 86-106. Deutsch, M. (1985). Distributive Justice. New Haven, CT: Yale University Press. Fisk, R.P. and K.A. Coney. (1982). “Postchoice Evaluation: An Equity Theory Analysis of Consumer Satisfaction/Dissatisfaction with Service Choices.” Pp. 9-16 in H. Keith Hunt and Ralph L. Day (eds.), Conceptual and Empirical Contributions to Consumer Satisfaction and Complaining Behavior. Bloomington, IN: Indiana University Press. Folkes, V.S. (1984). “Consumer Reactions to Product Failure: An Attributional Approach,” Journal of Consumer Research, 10: 398-409. Folger, R. and M.A. Konovsky. (1989). “Procedural Justice, Distributive Justice, and Reactions to Pay Raise Decisions,” Academy of Management Journal, 32: 851-866. Fomell, C. and B. Wernerfelt. (1987). “Defensive Marketing Strategy by Customer Complaint Management: A Theoretical Analysis,” Journal of Marketing Research, 24(November): 337-346.

Postcomplaint

209

Behavior

Gilly, M.C. and B.D. Gelb. (1982). “Post-Purchase Consumer Processes and the Complaining Consumer,” Journal of Consumer Research, 9: 323-328. Goodwin, C. and I. Ross. (1992). “Consumer Responses to Service Failures: Influence of Procedural and Interactional Fairness Perceptions,” Journal ofBusiness Research, 25: 149-163. (1989). “Salient Dimensions of Perceived Fairness in Resolution of Service Complaints,” -. Journal

of Consumer

Satisfaction/Dissatisfaction

and Complaining

Behavior,

2: 87-92.

Greenberg, J. and R. Folger. (1983). “Procedural Justice, Participation and the Fair Process Effect in Groups and Organizations.” Pp. 235-256 in P.B. Paulus (ed.), Basic Group Processes. New York: Springer-Verlag. Greenberg, J. and C. McCarty. (1990). “The Interpersonal Aspects of Procedural Justice: A New Perspective in Pay Fairness,” Labor Law Journal, 41(August): 580-585. Greenberg, J. (1990). “Organizational Justice: Yesterday, Today, and Tomorrow,” Journal of Management, 16(June): 399-432. Hart, C.W.L., J. L. Heskett, and W.E. Sasser Jr. (1990). “The Profitable Art of Service Recovery,” Harvard Business Review, 68(July_August): 148-156. Hirschman, A. 0. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations and States. Cambridge, MA: Harvard University Press. Hui, M.K. and J.E. Bateson. (1991). “Perceived Control and Consumer Choice on the Service Experience,” Journal of Consumer Research, lS(September): 174-185. Huppertz, J.W., S.J. Arenson, and R.H. Evans. (1978). “An Application of Equity Theory to BuyerSeller Exchange Situations,” Journal qfMarketing Research, lS(May): 250-260. Jacoby, J. and J.L. Jaccard. (1981). “The Sources, Meaning, and Validity of Consumer Complaint Behavior: A Psychological Analysis,” Journal @Retailing, 57(Fall): 4-24. Katz, K., B. Larson, and R. Larson. (1991). “Prescription for the Waiting in Line Blues: Entertain, Enlighten, and Engage,” Sloan Management Review, (Winter): 44-53. Kelly, S.W., and M.A. Davis. (1994). “Antecedents to Customer Expectations for Service Recovery,” Journal of the Academy of Marketing Science, 22(Winter): 52-61. Kelly, S.W., K.D. Hoffman, and M.A. Davis. (1993). “A Typology of Retail Failures and Recoveries,” Journal of Retailing, 69(Winter): 429-452. Leventhal, J., J. Karuza, and W.R. Fry. (1980). “Beyond Fairness: A Theory of Allocation Preferences.” Pp. 167-218 in G. Mikula (ed.), Justice and Social interaction. New York: SpringerVerlag. Lind, E.A. and T.R. Tyler. (1988). The Social Psychology of Procedural Justice. New York: Plenum Press. McCollough, M.A. and SC. Bharadwaj. (1992). “The Recovery Paradox: An Examination of Consumer Satisfaction in Relation to Disconfirmation, Service Quality, and Attribution Based Theories.” In Chris T. Allen et al. (eds.), Marketing Theory and Applications. Chicago: American Marketing Association, Maister, D.H. (1985). “The Psychology of Waiting Lines.” Pp. 113-123 in John D. Czepiel, Michael R. Solomon, and Carol F. Surprenant (eds.), The Service Encounter. Lexington, MA: Lexington Books. Messick, D.M. and KS. Cook. (1983). “Psychological and Sociological Perspectives on Distributive Justice: Convergent, Divergent and Parallel Lines.” Pp. l-13 in David M. Messick and Karen S. Cook (eds.), Equity Theory: Psychological and Sociological Perspectives. New York: Praeger. Mohr, L.A. (1991). Social Episodes and Consumer Behavior: The Role of Employee Effort in Satisfaction with Services. Unpublished doctoral dissertation, Arizona State University. Mowen, J.C. and S.J. Grove. (1983). “Search Behavior, Price Paid and the Consumption Other: An Equity Theory Analysis of Post-Purchase Satisfaction.” Pp. 57-63 in Ralph L. Day and H. Keith

210

Journal of Retailing

Vol. 73, No. 2 1997

Hunt (eds.), International Fare in Consumer Satisfaction and Complaining Behavior. Bloomington, IN: Indiana University Press. Oliver, R.L. (1980). “A Cognitive Model of the Antecedents and Consequences of Satisfaction Decisions,” Journal of Marketing Research, 17(November): 460-469. Oliver, R.L. and W.S. DeSarbo. (1988). “Response Determinants in Satisfaction Judgment,” Journal of Consumer Research, 14(March): 495-507. Oliver, R.L. and J.E. Swan. (1989). “Consumer Perceptions of Interpersonal Equity and Satisfaction in Transactions: A Field Survey Approach,” Journal of Marketing, 53(April): 21-35. Parasuraman, A., V.A. Zeithaml, and L.L. Berry. (1985). “A Conceptual Model of Service Quality and its Implications for Future Research,” Journal of Marketing, 49(Fall): 41-50. Perdue, B.C. and J.O. Summers. (1986). “Checking the Success of Manipulations in Marketing Experiments,” Journal of Marketing Research, 23(November): 317-326. Reichheld, F.F. and E.W. Sasser, Jr. (1990). “Zero Defects: Quality Comes to Services,” Harvard Business Review, 68(September_October): 105-l 11. Richins, M.L. (1983). “Negative Word-of-Mouth by Dissatisfied Consumers: A Pilot Study,” Journal of Marketing, 47(Winter): 68-78. Singh, J. (1988). “Consumer Complaint Intentions and Behavior: Definitional and Taxonomical Issues,” Journal of Marketing, 52(January): 93-107. Solomon, M.R., C. Surprenant, J.A. Czepiel, and E. Gutman. (1985). “A Role Theory Perspective on Dyadic Interactions: The Service Encounter,” Journal of Marketing, 49(Winter): 99-l 11. Taylor, S. (1994). “Waiting for Service: The Relationship Between Deals and Evaluations of Service,” Journal of Marketing, 58(April): 56-69. Technical Assistance Research Programs. (1986). Consumer Complaint Handling in America: An Update Study. Washington, DC: U.S. Office of Consumer Affairs. Thibaut, J. and L. Walker. (1975). Procedural Justice: A Psychological Analysis. Hillsdale, NJ: Erlbaum. Tyler, R.T. (1987). “Conditions Leading to Value-Expressive Effects of Procedural Justice: A Test of Four Models,” Journal of Personal@ and Social Psychology Research, 52: 333-344. Tyler, R.T. and R.J. Bies. (1989). “Beyond Formal Procedures: The Interpersonal Context of Procedural Justice.” Pp. 77-98 in Applied Psychology and Organizational Settings. Hillsdale, NJ: Erlbaum. Ulrich, W.L. (1984). “HRM and Culture: History, Ritual and Myth,” Humnn Resource Management, 23(Summer): 117-128. Venkatesan, M., and B.B. Anderson. (1985). “Time Budgets and Consumer Services.” Pp. 52-55 in Thomas M. Bloch, Gregory D Upah, and Valerie Zeithaml (eds.), Services Marketing in a Changing Environment. Chicago: American Marketing Association. Westbrook, R.A. (1987). “Product/Consumption-Based Affective Responses and Postpurchase Processes,” Journal of Marketing Research, 24(August): 258-270.