Received 15 April 2016; accepted 25 June 2016; published 28 June 2016
From the perspective of relationship marketing theory, relationship quality reflects a general assessment of relationship strength (Huang and Chiu, 2006)  . Enhancement of relationship quality is thus important. Prior studies find that justice influences relationship quality; however, findings regarding the relative effects of various types of justice (distributive justice, procedural justice and interactional justice) on relationship quality are inconsistent. Some researchers believe that distributive justice has a larger impact on relationship quality, while others believe that procedural justice has a greater impact (Humphreys and Williams, 1996; Leung, Smith, Wang, and Sun, 1996; Tang and Sarsfield-Baldwin, 1996; Tax, Brown, and Chandrashekaran, 1998; Teo and Lim, 2001)  -  . This study proposes two possible reasons for this discrepancy below.
1.1. Most Prior Studies Neglect Individual Differences in Relationship Demand
First, as some retailers are economic oriented (transactional), whereas others are social oriented (relational), retailers may care about different types of justice. Economic-oriented (transactional) retailers may care about distribution results, leading them to attach greater importance to distributive justice. However, social-oriented (relational) retailers care more about the exchange process and thus focus more heavily on procedural justice and interactional justice. Therefore, the relative effects on relationship quality of distributive justice, procedural justice and interactional justice may vary with different retailer demands. Although the fact that service involves both outcome and process has been recognized by many researchers for some time, researchers have traditionally focused only on service processes and have seldom discussed outcome and process at the
1.2. Most Prior Studies Regard Satisfaction and Commitment as Unitary Constructs
Second, with respect to “the relative impacts of distributive justice, procedural justice and interactional justice on relationship quality”, most prior causal studies have viewed satisfaction, trust and commitment as unitary constructs (Leung, Smith, Wang, and Sun, 1996; Teo and Lim, 2001)   . However, this position is debatable, as satisfaction, trust and commitment have multiple dimensions. Taking satisfaction as an example, channel members’ working relationships with other channel members are influenced by both economic and non-eco- nomic issues. If studies of satisfaction adopt only economic issues as their barometer, the study results will significantly differ from those that adopt only social issues as their barometer (Geyskens, Steenkamp, and Kumar, 1999; Geyskens and Steenkamp, 2000)   . Therefore, in exploring the relationships of distributive, procedural and interactional justice with relationship quality (including satisfaction, trust and commitment), this study treats satisfaction, trust and commitment as a single construct with multiple dimensions.
2. Theoretical Framework
Although justice researchers in different fields (such as organizational behavior, organizational reconstruction, service justice, service complaints, customer satisfaction and channel relationships) have suggested that there are many types of justice, research has tended to focus on three primary types: distributive justice, procedural justice and interactional justice.
Tang and Sarsfield-Baldwin (1996)  illustrated two types of organizational justice, distributive justice and procedural justice, from the perspective of organizational behavior. Distributive justice emphasizes the outcome or the content of decisions, while procedural justice emphasizes the means or the procedure by which the outcome is achieved (i.e., how a decision is made).
Seiders and Berry (1998)  classified service justice into three categories in their study of the company-customer relationship: distributive justice, procedural justice and interactional justice. While distributive justice concerns the outcomes of decision-making or exchange processes, procedural justice focuses on the process of reaching an outcome, and interactional justice addresses interpersonal treatment of customers in the process.
In their article “Customer Evaluations of Service Complaint Experiences”, Tax, Brown, and Chandrashekaran (1998)  observed that while distributive justice involves outcomes of decision-making processes, procedural justice concerns decision-making processes themselves, and interactional justice involves interpersonal behavior during the process of achieving the outcome.
Szymanski and Henard (2001)  conducted a meta-analysis of the cause and effect of customer satisfaction, noting that distributive justice attaches greater importance to that customers receive what they deserve, given their investment and labor; that procedural justice attaches greater importance to how outcomes are transmitted; and that interactional justice attaches greater importance to the ways in which customers are treated (involving respect, manners and dignity).
The concept of justice is applied only rarely in channel relationship studies. One relevant paper is that by Kumar, Scheer, and Steenkamp (
2.2. Relationship Quality
Nyaga and Whipple (2011)  indicated that the higher the quality of the relationship between buyer and supplier, the higher the level of supply chain operational performance and satisfaction with strategic performance, regardless of the type of relationship that exists (collaborative or arm’s length). Relationship quality is an overarching construct composed of important relational outcomes that reflect the overall nature of the exchange relationship (Skarmeas and Robson, 2008)  . While numerous dimensions of relationship quality are proposed in the literature, there does not appear to be a real consensus regarding the conceptualization of relationship quality (De Búrca, Fynes, and Roche, 2004)  . However, satisfaction, trust and commitment are most commonly cited in the literature as key aspects of relationship quality.
Channel member satisfaction is most commonly defined as a state of positive emotion that results from the evaluation of all aspects of a firm’s working relationship with another firm (Abdul-Muhmin, 2002; Geyskens, Steenkamp, and Kumar, 1999; Lam, Shankar, Erramilli, and Murthy, 2004; Skarmeas and Robson, 2008)     . Mayo, Richardson, and Simpson (1998)  argued that resellers’ satisfaction with suppliers is based on sales support, promotional support, profitability, new product development, product quality, personnel and service. Schellhase, Hardock, and Ohlwein (2000)  observed that retailer satisfaction with a supplier is mainly determined by the following factors: contact staff, cooperation strength, price, condition management, quality and flexibility. Ruekert and Churchill (1984)  evaluated dealers’ satisfaction in four dimensions: product, finance, support and social interaction dimensions. All these studies of channel member satisfaction suggest that channel member satisfaction should include both economic and non-economic aspects (Geyskens, Steenkamp, and Kumar, 1999)  .
Consistent with the relevant scholarly research, this study defines retailer satisfaction as the attitude produced by the retailer’s evaluation of the retailer-supplier relationship with respect to all issues pertaining to the relationship, including economic and non-economic issues.
Geyskens, Steenkamp, and Kumar (1999)  found in their study of marketing channel relationships that definitions of channel satisfaction offered by scholars in the past had lacked clarity. While some researchers have measured channel satisfaction from an economic perspective, others have measured it from a non-economic or psychological perspective, leading to discrepancies. In addition, measurements of satisfaction in some studies have involved differing proportions of economic and non-economic items, which may influence study results. To obtain correct conclusions, channel satisfaction must be split into two variables: satisfaction attributable to economic factors (termed economic satisfaction) and satisfaction attributable to non-economic factors (termed non-economic satisfaction). The present study integrates the two types of channel satisfaction into the research model in an effort to make results comparable. The definitions of economic and non-economic satisfaction are more clearly illustrated as follows:
Economic satisfaction: the positive affective reactions of channel members to economic aspects of partnerships, for example, sales volume, sales profits, target achievement, partnership effectiveness, productivity and financial outcomes.
Non-economic satisfaction: the positive affective reactions of channel members to non-economic and psychological aspects of partnerships, for example, fair interactions leading to satisfaction in exchanges and trust in partners as caring, respectful and willing to exchange ideas.
The outcomes of customer satisfaction can be divided into two categories, attitude (e.g., trust and commitment) and behavior (e.g., behavioral intention, purchase behavior), both of which play an important role in long-term customer relationships. However, the literature currently available on satisfaction is primarily concerned with behavioral intention with respect to the company and/or products. In other words, past research on satisfaction is oriented toward behavior rather than attitude (Tax, Brown, and Chandrashekaran, 1998)  . To address this research deficiency, the present study considers two attitude outcome variables―trust and commitment―which are also important variables in relationship development.
Trust denotes willingness to rely on an exchange partner in whom one has confidence (Skarmeas and Robson, 2008)  . The current business trend has shifted from competition to cooperation. Trust can influence the cooperative willingness of two parties and establish competitive advantages and harmonious channel relationships. Thus, trust plays an important role in channel management and operations (Morgan and Hunt, 1994)  .
Although the dimensions of trust proposed by Kumar, Scheer, and Steenkamp (
Studies of exchange relationships among companies have concluded that commitment is an important concept, referring to the degree to which close and persistent relationships with other parties are established and maintained (Kim and Frazier, 1997)  . Dwyer, Schurr, and Oh (1987)  described commitment as a guarantee to maintain a relationship and the intent to sacrifice short-term benefits for long-term interests. Commitment is a persistent attitude that reflects the positive values of a relationship. Thus, commitment does not change frequently, as people would not make commitments to valueless relationships. As defined by Moorman, Zaltman, and Deshpande (1992)  , commitment is the persistent desire to maintain a valuable relationship. Commitment, which can be viewed as the highest level of relational bonding, constitutes an indispensable part of a successful relationship (Skarmeas and Robson, 2008)  . Therefore, the retailer’s commitment to its supplier can adequately reflect relationship quality of the two parties.
The notion that commitment consists of multiple dimensions has been accepted and supported by many marketing scholars (Gilliland and Bello, 2002; Gundlach, Achrol, and Mentzer, 1995; Kelly, 2004; Kim and Frazier, 1997)   -  . The three major types of commitment are affective commitment, continuance commitment and normative commitment, each of which is impelled by different driving forces. Originally proposed in a study of organizational behavior by Allen and Meyer (1990)  and Meyer and Allen (1991)  , this notion of commitment has now been widely applied in marketing studies. Continuance commitment, which refers to the commitment to maintain a cooperative relationship, is based on cost and economic considerations; thus, it is the product of rational and economic calculation. Some studies have referred to this as calculative commitment (Kelly, 2004)  . In terms of the supplier/retailer relationship, if retailers make commitments after calculating the profit/loss from being in or out of cooperative relationships with suppliers, this is a rational, task-oriented and instrumental relationship connection (Gilliland and Bello, 2002)  . By contrast, both affective and normative commitment falls into the category of loyalty commitment. Affective commitment arises from retailers’ identification with and emotional attachment to suppliers (Kelly, 2004; Monroy and Alzola, 2005)   . Normative commitment arises from a retailers’ sense of obligation to maintain a cooperative relationship with a supplier, while a sense of obligation to maintain such a relationship arises from shared values of the parties, produced after a period of internalization (Allen and Meyer, 1990; Kelly, 2004)   . Relationship connections between retailers and suppliers include economic and social aspects, representing two types of retailer-supplier relationship. Continuance commitment stresses the economic aspect of a cooperative relationship, while affective and normative commitment stress the social aspect of such a relationship (Gilliland and Bello, 2002; Lee, Sirgy, Brown, and Bird, 2004)   .
Brown, Lusch, and Nicholson (1995)  clearly noted that commitment to some channel relationships might be driven by economic or extrinsic concerns, such as the desire for economic rewards or the avoidance of economic harm, a type of commitment that is shallow and short-lived. By contrast, if commitment is based on non-economic or intrinsic concerns, such as identification with partners or internalization of similar values, commitment is comparatively long-lived.
2.2.4. Dual Domains Model
Chinomona (2013)  indicated that non-coercive power positively influences channel relationship outcomes. The maintenance function of justice can foster satisfaction with a collaborative relationship and enhance a partnership (Gu and Wang, 2011)  . Prior research has revealed a positive relationship between justice and satisfaction (Brown, Cobb, and Lusch, 2006; Szymanski and Henard, 2001)   , whereas perceptions of injustice may lead to distress, a key element in dissatisfaction (Patterson, Johnson, and Spreng, 1997)  . Prior research on satisfaction and trust has revealed that satisfaction positively affects trust (Dwyer, Schurr, and Oh, 1987; Geyskens, Steenkamp, and Kumar, 1999)   ; after establishing a trusting relationship, both parties can relatively easily establish the commitment needed to maintain their relationship (Hadjikhani and Thilenius, 2005; Moorman, Zaltman, and Deshpande, 1992)   . In short, prior research has found that justice influences satisfaction, that satisfaction influences trust, and that trust influences commitment.
Although justice, satisfaction, trust and commitment are positively related to one another as a construct, the linkages between dimensions of these constructs are unknown. As depicted in the introduction above, because prior research has neglected individual discrepancies in relationship demands, we believe such research may have led to inconsistent findings regarding the relative effects of distributive justice and procedural justice on relationship quality. The present study adopts a different research approach (the dimensional viewpoint) to explain this problem and clarify that justice and relationship quality have a special dimensional linkage.
The core concept of this study is that justice and relationship quality in an exchange relationship between retailers and suppliers should be viewed in terms of two categories: outcome and process. All study variables, based on their characteristics, are accordingly divided into the two domains of outcome and process. We posit a stronger connection between variables in the same domain and establish a dual domains model to illustrate the connection between justice and relationship quality within given dimensions
First, the core concept of this study, “variables in the same domain have a stronger connection relationship”, is explained by the research results of the service remedy strategy study. Service contact failures have two sources: outcome failure and process failure (representing different types of losses). The service contact outcome dimension pertains to the final results that customers receive, whereas the process dimension concerns the process by which customers receive the service, i.e., how the service is transmitted. Smith, Bolton, and Wagner (1999)  addressed how to use a service remedy to maintain customer satisfaction after service failure by sampling 375 college students who had recently visited restaurants and 602 customers of a chain hotel. The study results revealed that, following a service failure, if the type of service remedy that the customer received was identical to the type of service failure experienced, the customer had comparatively positive feelings, i.e., the service outcome failure should be redressed by outcome attributes, whereas the service process failure should be redressed by process attributes. When service failure occurs, an economic-resource remedy is more effective in the case of service outcome failure, whereas social-resource compensation is more effective in the case of process failure. For example, after a customer is rudely treated by a waiter (process failure), an apology (social resource) would be more effective than a discount (economic resource) as a remedy to the situation (Smith, Bolton, and Wagner, 1999)  .
The theoretical bases of this study are the principle of resource exchange, the principle of mental accounting and prospect theory. According to the principle of resource exchange, people prefer resources that are of the same type as those traded in an exchange. In other words, exchanging resources of the same type leads to a higher degree of satisfaction than exchanging resources of different types. According to the principle of mental accounting, people classify resources into different mental accounts by various intrinsic methods. For example, people categorize economic resources and social resources into different mental accounts. Prospect theory proposes that every resource has a different weight when individuals make decisions based on psychological utility and that resources in short supply have higher weights than resources that are ample. The three theories can be integrated as follows. According to prospect theory, suppliers should be greatly concerned when retailers are short of resources. According to the principle of mental accounting, retailers categorize resources into different mental accounts. Because, according to the principle of resource exchange, resources that retailers want are of the same type as the type of resource that they have lost or are missing, suppliers should know what type of resource retailers are missing.
Which research variables in this study belong to the “outcome domain”, and which belong to “process domain”? This study classifies variables according to their characteristics.
With respect to justice, retailers expect to obtain not only a satisfactory outcome from the core service but also relationship benefits from the service and interactional processes. Through the former, one obtains economic resources; through the latter, one obtains social resources. Distributive justice is concerned with the outcome of decision-making or exchange. Events in the outcome dimension frequently influence retailers’ perceptions of acquired core services, which are often related to factors that belong to service outcomes such as profit, purchase price, promotional support and sales volume. Therefore, distributive justice is a variable belonging to the “outcome domain”. Procedural justice concerns the process of distribution, and interactional justice concerns the interpersonal treatment customers receive in decision-making or exchange processes. Events in the procedural and interactional dimensions frequently influence customers’ perceptions of relationship interests, which are often related to factors belonging to service processes such as participation in discussions, communication and interpersonal attitudes. Thus, procedural justice and interactional justice are variables that belong to the “process domain”.
With respect to satisfaction, economic satisfaction refers to satisfaction with respect to economic issues such as sales volume, sales profits and the accomplishment of objectives related to the outcome of an exchange. Thus, economic satisfaction is a variable that belongs to the “outcome domain”. Non-economic satisfaction refers to satisfaction with respect to non-economic issues, a type of satisfaction that concerns whether the customer is happy or not in the exchange process. Thus, non-economic satisfaction involves affective flows that occur in the exchange process and is thus a variable belonging to the “process domain”.
With respect to trust, ability trust refers to retailers’ trust in suppliers’ management abilities and abilities to create profits, for example, introducing successful goods, selecting correct management strategies, and effectively integrating retailers, all of which are related to beneficial economic outcomes. Therefore, ability trust mainly rests on successful operational outcomes and belongs to the “outcome domain”. By contrast, benevolence trust indicates retailers’ trust in suppliers’ kindness and the sincerity of concern about them. Integrity trust refers to retailers’ trust in the consistency of suppliers’ behavior and their sense of justice, unselfishness and obedience to moral norms. All of these are reflected in daily exchange processes. Therefore, benevolence trust and integrity trust belong to the “process domain”.
Regarding commitment, continuance commitment is a component of interest-oriented commitment. It is a commitment produced by retailers after consideration of “interest” in the cooperative relationship. The purpose of exchange of retailers with suppliers under such commitment is to acquire benefits from cooperation or to avoid disadvantages that retailers may suffer if they leave the cooperative relationship. Therefore, continuance commitment results from an assessment of cooperative results and is thus a variable belonging to the “outcome domain”. By contrast, affective commitment and normative commitment are socially oriented commitments that are based on considerations of affection or obligation to remain in the exchange relationship. These feelings are produced from their interaction; thus, affective and normative commitment are variables belonging to the “process domain”.
To summarize the discussion, distributive justice, economic satisfaction, ability trust and continuance commitment are variables that belong to the outcome domain; procedural justice (and interactional justice), non-economic satisfaction, benevolence trust (and integrity trust) and affective commitment (and normative commitment) are variables that belong to the process domain. Furthermore, based on the core concept of this study, “variables in the same domain have a stronger connection relationship”, this study establishes the “dual domains model”. The model is illustrated as follows.
The relationship connections between variables of the outcome domain are referred to as economic chains; the relationship connections between variables of the process domain are referred to as social chains.
Focusing on the economic chain, if suppliers perform well with respect to distributive justice by stressing outcome distribution and distribute economic values effectively, retailers will experience higher economic satisfaction. With the accumulation of economic satisfaction, retailers gain greater trust in the supplier's abilities and believe that cooperation with the supplier brings positive economic outcomes. Finally, the motivation for retailers to maintain relationships arises mainly from economic considerations―continuance commitment. This exemplifies the economic chain: distributive justice -> economic satisfaction -> ability trust -> continuance commitment.
In contrast, turning to the social chain, if suppliers perform well with respect to procedural and interactional justice, which stress procedure and process in the establishment, maintenance and development of a relationship, retailers experience greater non-economic satisfaction. With the accumulation of non-economic satisfaction, retailers gain greater trust in suppliers’ benevolence and integrity. Finally, the motivation of retailers to maintain relationships mainly rests on social considerations―affective and normative commitment. This exemplifies the social chain: procedural justice (and interactional justice) -> non-economic satisfaction -> benevolence trust (and integrity trust) -> affective commitment (and normative commitment).
Based on the above discussion, we propose the following hypothesis and research framework depicted in Figure 1.
H1: Economic satisfaction is mainly influenced by distributive justice and not by procedural and interactive justice. Non-economic satisfaction is mainly influenced by procedural and interactive justice and not by distributive justice.
H2: Ability trust is mainly influenced by economic satisfaction and not by non-economic satisfaction. Benevolence trust is mainly influenced by non-economic satisfaction and not by economic satisfaction. Integrity trust is mainly influenced by non-economic satisfaction and not by economic satisfaction.
H3: Continuance commitment is mainly influenced by ability trust and not by benevolence and integrity trust. Affective commitment is mainly influenced by benevolence and integrity trust and not by ability trust. Normative commitment is mainly influenced by benevolence and integrity trust and not by ability trust.
3. Research Methodology
This section explains the questionnaire design, operational definitions and measures of variables in addition to the pretest and modification of the questionnaire, the empirical data collection and the reliability and validity of its various measures.
3.1. Questionnaire Design
The questionnaire was designed from the perspective of retailers who are assessing their largest supplier. Retailers thus had a definite target in mind during completion of the questionnaire. Moreover, because high exchange frequency of a retailer with his/her largest supplier leads to a clear understanding of the relationship, the retailer could easily assess the supplier’s behavior (justice) and its relationship with the supplier (satisfaction, trust and commitment).
The questionnaire was designed on the basis of multiple indexes, as suggested by Churchill (1979)  . The measure adopted is the Likert seven-point scale, ranging from strong disagreement to strong agreement, using
Figure 1. Research framework (dual domains model).
scores from one to seven. A high score represents strong positive perceptions of retailers about suppliers regarding justice, satisfaction, trust and commitment.
3.2. Operational Definitions and Measures of Variables
The operational definitions and measures of variables are elaborated as follows:
The operational definition of distributive justice in this study is as follows: retailers’ psychological feelings elicited through comparison between real acquired outcomes and their desired outcomes based on standards such as “efforts and investments made for marketing this supplier’s products”, “the role and obligation undertaken by the retailer for the supplier”, “profits of other retailers in the same industry”, “profits obtained by this supplier through exchange with the retailer” and “the retailer’s contributions to the supplier”.
The operational definition of procedural justice in this study is as follows: retailers’ perceptions of the justice of procedures and processes undertaken by suppliers. Three important dimensions of procedural justice are adopted: first, impartiality, i.e., the degree to which suppliers’ channel policies are equivalent for all retailers; second, refutability, i.e., the degree to which retailers can refute suppliers’ policies; and third, knowledge ability, i.e., the degree to which suppliers are familiar with the local conditions of retailers’ operations.
The operational definition of interactional justice in this study is as follows: retailers’ perceptions of the justice of interpersonal treatment that retailers receive from suppliers. Interactional justice is measured by the following three dimensions: first, explanation, i.e., suppliers’ explanations to retailers of the reasons for their decisions and policies; second, bilateral communication, i.e., engagement of suppliers and retailers in exchanges of opinion; and third, courtesy, i.e., suppliers’ attitudes toward retailers.
To measure suppliers’ justice, the present study adopted the scale used by Kumar, Scheer, and Steenkamp (
The measure of retailer satisfaction in this study includes two parts, economic satisfaction and non-economic satisfaction. The operational definition of economic satisfaction is as follows: retailers’ attitudes derived from retailers’ evaluations of economic issues in exchanges with suppliers, where evaluated factors include sales volume, profit and discount. By contrast, the operational definition of non-economic satisfaction is as follows: retailers’ attitudes derived from retailers’ evaluations of non-economic issues in exchanges with suppliers, where evaluated factors include harmony of relationship, mutual respect and honest treatment.
The measured items of economic and non-economic satisfaction are obtained through modification of the scale used by Geyskens and Steenkamp (2000)  to measure economic and social satisfaction in the marketing channel relationship; the scale used by Gassenheimer, Calantone, and Scully (1995)  to measure dealers’ satisfaction; and the scale used by Crosby, Evans, and Cowles (1990)  to measure satisfaction with respect to salesmen.
The measure of trust used in this study is divided into ability trust, benevolence trust and integrity trust. The operational definition of ability trust refers to the degree of retailers’ belief that suppliers have various relevant abilities; benevolence trust refers to the degree of retailers’ belief that suppliers care about them; and integrity trust refers to the degree of retailers’ belief that suppliers are honest and candid.
The measured items pertaining to ability trust, benevolence trust and integrity trust are obtained through modification of the scale developed by Jarvenpaa, Knoll, and Leidner (1998)  to measure ability, benevolence and integrity.
This study adopts the commitment classifications of Meyer and Allen (1991)  , whereby commitment is divided into three categories: affective commitment, continuance commitment and normative commitment. The operational definitions of the three types of commitment are as follows. Affective commitment: the retailer wishes to maintain its relationship with its supplier on the basis of affective considerations. Continuance commitment: the retailer must maintain its relationship with its supplier, given profit and cost considerations. Normative commitment: the retailer should maintain its relationship with its supplier on the basis of obligation considerations.
The items used to measure commitment are obtained through modification of the scale employed by Meyer, Allen, and Smith (1993)  to measure organizational commitment and modification of the scale used by Verhoef, Franses, and Hoekstra (2002)  to measure customer commitment in the service industry.
3.3. Pretest and Modification of the Questionnaire
3.3.1. Pretest Sampling
The pretest sample consisted of retailers in Tainan City. The main purpose of the pretest was to test the fitness of the questionnaire content and to anticipate possible reactions of respondents to the questionnaire.
3.3.2. Modification of Questionnaire Items
Modifications of questionnaire items were based primarily on information from two sources: first, retailers’ reactions to the pretest questionnaire, some of which were suggestions regarding wording, while others were suggestions of new items that would make the questionnaire more complete; and second, statistical results based on 70 valid pretest questionnaires.
Factor analysis and Cronbach’s alpha reliability test were used to detect improper items, which were then deleted or revised. In the factor analysis, we adopted varimax rotation and an eigenvalue above 1.00 as the standard for selecting factors. Factor loadings must be above 0.6, and differences in factor loadings under different factors for the same item must be above 0.3 to ensure construct validity. The alpha value of the reliability test must be above 0.7 to ensure the consistency of items in a given dimension.
The items for which retailers had opinions, poor factor loading or low consistency within dimensions were reevaluated. It was then decided whether to delete or revise such items. If a variable measure consisted of less than five items after deleting items, new and more suitable items were added. The revised questionnaire was used for our official survey. The items were 35 for justice construct, 17 for satisfaction construct, 17 for trust construct, and 17 for commitment construct.
3.4. Empirical Data Collection
The population of this study consisted of retailers in Taiwan, and personal interviews were used for questionnaire data collection. We adopted purposive sampling to obtain the research sample, and sample selection was based on the following three principles. First, retailers with no suppliers were excluded. Second, retailers must have a definite supplier, which is the object assessed by the questionnaire. Third, suppliers in the study are finished goods suppliers rather than material or semi-finished goods suppliers.
Because exchanges with suppliers are principally handled by bosses or store managers, general staff are not expected to understand relationships with suppliers in an in-depth manner. Thus, the respondents in this study consisted exclusively of bosses or store managers.
We successfully collected 1250 questionnaires. After excluding those questionnaires with incomplete answers and careless errors, 1009 questionnaires were retained. We randomly selected 15% of the retained questionnaires to verify, via a telephone survey, that the interviews had occurred. The results indicated all of the interviews had indeed occurred. Thus, 1009 questionnaires were valid, and the subsequent analysis was conducted.
With respect to the structure of the valid sample, bosses and store managers each composed approximately one-half of the respondents. The respondents were approximately evenly distributed across northern, middle and southern areas of Taiwan. In terms of retail product types, the percentages of the sample were as follows: food products 21.9%, clothing 9.5%, glasses 6.7%, computers and communications 6.5%, books and stationery 7.0%, housing 0.3%, household appliances and electric appliances 5.0%, printing 4.3%, automobile and auto parts 3.1%, motorcycle and bicycle sales 4.3%, household goods 3.9% and other 27.5%.
3.5. Reliability and Validity of the Measures
We tested the reliability of the construct measures using Cronbach’s alpha. The results (Table 1) show that the Cronbach’s alpha for both unitary constructs and dimensions of multi-dimensional constructs exceed 0.7, indicating fairly high reliability.
Table 1. Reliability and convergent validity of measures.
To examine convergent validity, we conducted a factor analysis of each construct measure. Varimax rotation was employed in the factor analysis, and an eigenvalue above one was used as the standard for selecting factors; the results are shown in Table 1. The results of the factor analysis indicate, first of all, that the derived factor structure is identical to the theorized factor structure. In other words, excepting procedural justice and interactive justice, which are multi-dimensional, all of the constructs are unit-dimensional. Second, all the factor loadings and variances extracted for the construct measures exceed 0.6. Thus, our measures possess strong convergent validity (as Table 1).
Regarding discriminant validity, the maximum correlation coefficient between the 11 research variables is 0.77, and the minimum Cronbach’s alpha value is 0.86. Thus, the correlation coefficients between the two variables are smaller than the Cronbach’s alpha values for the two variables themselves, indicating discriminant validity among the eleven variables (Gaski and Nevin, 1985)  .
4. Research Results
Seventeen multiple regression analyses were employed to determine whether there are special linkages between the different types of justice, satisfaction, trust and commitment. The results are shown in Table 2. All seventeen models are significant (p < 0.001) and have significant explanatory power (R2 > 0.2). Additionally, based on the standardized regression coefficients, we assessed linkages among variables.
Economic satisfaction is mainly influenced by distributive justice (β = 0.522, t = 21.551) and not by procedural justice (β = 0.169, t = 4.869) or interactive justice (β = 0.200, t = 6.180). Non-economic satisfaction is mainly influenced by procedural justice (β = 0.347, t = 12.217) and interactive justice (β = 0.547, t = 20.680) and not by distributive justice (β = 0.000, t = 0.009). Ability trust is mainly influenced by economic satisfaction (β = 0.712, t = 31.454) and not by non-economic satisfaction (β = 0.134, t = 5.929). Benevolence trust is mainly influenced by non-economic satisfaction (β = 0.709, t = 31.445) and not by economic satisfaction (β = 0.140, t = 6.213). Integrity trust is mainly influenced by non-economic satisfaction (β = 0.680, t = 28.341) and not by economic satisfaction (β = 0.136, t = 5.689). Continuance commitment is mainly influenced by ability trust (β = 0.550, t = 17.933) and not by benevolence trust (β = 0.018, t = 0.465) or integrity trust (β = 0.035, t = 0.864). Affective commitment is mainly influenced by benevolence trust (β = 0.407, t = 13.472) and integrity trust (β = 0.415, t = 13.667) and not by ability trust (β = 0.030, t = 1.296). Normative commitment is mainly influenced by benevolence trust (β = 0.278, t = 6.586) and integrity trust (β = 0.314, t = 7.393) and not by ability trust (β = −0.116, t = −3.568). Thus, all research hypotheses are supported.
Table 2. Research results.
In addition to the above tests of hypotheses, we also tested other possible relationships. Ability trust is mainly influenced by distributive justice (β = 0.423, t = 15.153) and not by procedural justice (β = 0.236, t = 5.910) or interactive justice (β = 0.107, t = 2.873). Benevolence trust is mainly influenced by procedural justice (β = 0.343, t = 10.102) and interactive justice (β = 0.477, t = 15.051) and not by distributive justice (β = −0.017, t = −0.707). Integrity trust is mainly influenced by procedural justice (β = 0.352, t = 10.030) and interactive justice (β = 0.415, t = 12.689) and not by distributive justice (β = 0.046, t = 1.862). Continuance commitment is mainly influenced by distributive justice (β = 0.418, t = 13.196) and not by procedural justice (β = 0.210, t = 4.631) or interactive justice (β = −0.095, t = −2.251). Affective commitment is mainly influenced by procedural justice (β = 0.168, t = 4.771) and interactive justice (β = 0.606, t = 18.439) and not by distributive justice (β = 0.001, t = 0.027). Normative commitment is mainly influenced by procedural justice (β = 0.087, t = 1.873) and interactive justice (β = 0.413, t = 9.560) and not by distributive justice (β = −0.023, t = −0.705). Continuance commitment is mainly influenced by economic satisfaction (β = 0.707, t = 25.671) and not by non−economic satisfaction (β = −0.089, t = −3.235). Affective commitment is mainly influenced by non−economic satisfaction (β = 0.661, t = 26.422) and not by economic satisfaction (β = 0.126, t = 5.504). Normative commitment is mainly influenced by non−economic satisfaction (β = 0.439, t = 13.320) and not by economic satisfaction (β = 0.017, t = 0.529). The results for all empirical models indicate that the relationships among variables in the same domain are stronger than relationships across different domains.
This study incorporates the concept of justice into research on relationship quality, exploring in detail the relevance of justice to relationship quality. The results of the research demonstrate that the proposed dual domains model (Figure 1) is valid. The dual domains model makes two primary contributions. First, the model clarifies the relationship between different types of justice, satisfaction, trust and commitment. Second, the model can reasonably explain the discrepancies of past research results regarding the relative effects of distributive justice, procedural justice and interactive justice on relationship quality.
Regarding the first contribution, we find that justice and relationship quality have a special linkage that is dependent on dimensional characteristics. Based on the implications of the variables, the research variables in this study can be classified into two categories: outcome domain and process domain. We find that the variables with stronger outcome features have greater effects on variables in the outcome domain than in the process domain and that variables with stronger process features have greater effects on variables in the process domain than in the outcome domain. For example, let us take the relationship between distributive justice and satisfaction. Prior research has shown only that distributive justice positively influences satisfaction. Now, however, we know that distributive justice has its most significant influence specifically on economic, as opposed to non-economic, satisfaction. In addition, retailers’ sense of economic satisfaction mainly affects retailers’ ability trust of suppliers, not their benevolence trust or their integrity trust. Ability trust mainly fosters retailers’ continuance commitment with respect to suppliers but not affective commitment or normative commitment. Thus, we find that distributive justice -> economic satisfaction -> ability trust -> continuance commitment, which we label the economic chain. By contrast, we find that retailers’ perceptions of suppliers’ procedural justice (and interactive justice) mainly influence non-economic satisfaction rather than economic satisfaction, while non-economic satisfaction mainly affects benevolence trust (and integrity trust) rather than ability trust. Finally, benevolence trust (and integrity trust) affects retailers’ affective commitment (and normative commitment) to suppliers rather than their continuance commitment. Thus, we find that procedural justice (and interactive justice) -> non-economic satisfaction -> benevolence trust (and integrity trust) -> affective commitment (and normative commitment), which we label the social chain.
In addition to the above main findings, we find identical results in the branch chains. Thus, relationships between variables within given domains are stronger than relationships between variables across domains. With respect to the outcome domain, distributive justice has a greater effect on ability trust than on benevolence trust (and integrity trust); distributive justice has a greater effect on continuance commitment than on affective commitment (and normative commitment); and economic satisfaction has a greater effect on continuance commitment than on affective commitment (and normative commitment). In the process domain, procedural justice (and interactive justice) has a greater effect on benevolence trust (and integrity trust) than on ability trust; procedural justice (and interactive justice) has a greater effect on affective commitment (and normative commitment) than on continuance commitment; and non-economic satisfaction has a greater effect on affective commitment (and normative commitment) than on continuance commitment.
Regarding the second contribution of the research, the dual domains model advanced here helps explain past research result discrepancies with respect to the relative influences of distributive justice, procedural justice and interactive justice on relationship quality. The clarification is as follows.
Outcome/economy-oriented retailers attach great importance to distributive justice and economic aspects of relationship quality; thus, the relationships identified in the economic chain are more relevant to them. By contrast, process/non-economy-oriented retailers attach great importance to procedural justice (and interactive justice) and the social aspects of relationship quality. Accordingly, the relationships identified in the social chain are more relevant to them. Therefore, the types of justice with the strongest effects on relationship quality are determined by retailers’ relationship orientations. This finding helps explain the inconsistent findings of past studies. In addition, this finding indicates that discrepancies in previous research results may have been caused by incommensurable measurements. In terms of satisfaction, if economic items are used to measure satisfaction, then economic satisfaction is measured. Under this measurement condition, based on the dual domains model, satisfaction is mainly influenced by distributive justice. By contrast, if non-economic items are used to measure satisfaction, then non-economic satisfaction is measured. Under this measurement condition, based on the dual domains model, satisfaction is mainly influenced by procedural justice (and interactive justice). This may explain the inconsistent results of past studies regarding the types of justice with the greatest effects on relationship quality.
6. Implications for Business Marketing Practice
Based on the dual domains model advanced in this paper, different types of justice, satisfaction, trust and commitment fall into two domains, the outcome domain and the process domain; separately, they constitute the economic chain and social chain. This model facilitates discovery by suppliers of the causes (problems related to the outcome domain or problems related to the process domain) that lead to poor relations with retailers. It also helps suppliers understand the possible consequences that poor distributive justice, procedural justice and interactive justice separately produce. To promote relationship quality between suppliers and retailers, we suggest that suppliers hold symposiums and conduct questionnaire surveys on a regular basis to better understand retailers’ perceptions of justice and relationship quality. In particular, understanding which type of justice, satisfaction, trust or commitment that retailers feel they lack can help suppliers take appropriate actions to effectively elevate the quality of their relationships with their retailers. For instance, if retailers’ economic satisfaction is low, suppliers should examine their distributive actions to effectively solve the problem. Conversely, if retailers’ non-economic satisfaction is low, suppliers should examine their procedural and interactional activities.
With respect to promoting retailers’ perceptions of distributive justice, this study provides recommendations for suppliers. When suppliers distribute outcomes to retailers (for example, sales prices, rewards and promotional support), they should consider the efforts, involvement and investment of individual retailers as well as the standard of comparison that is likely to be adopted by retailers when retailers measure distributive justice. For example, suppliers should consider the following: 1) efforts made by retailers to sell the suppliers’ products; 2) investments made by retailers to sell the suppliers’ products; 3) the roles that retailers play; 4) the responsibilities shouldered by retailers; 5) the profits of other retailers in the same industry; 6) the profits of the suppliers themselves; and 7) retailers’ contributions to suppliers.
To promote procedural justice, there are several suggestions for suppliers. Timing should be considered when suppliers address problems related to retailers. Suppliers should also provide opportunities for retailers to participate in certain types of decision making, and they should adopt identical policies and decision procedures for all retailers. Suppliers should take retailers’ interests into consideration rather than simply pursuing their own goals. Suppliers should collect appropriate information, which should be provided to retailers and become the basis for suppliers’ decision-making. Suppliers should be magnanimous enough to correct improper decisions and value retailers’ contrary opinions. In making decisions, suppliers should care about the benefits and requests of retailers. Finally, suppliers’ decision procedures must meet moral standards. Immoral behavior such as deceit, bribery, infringement on privacy and secret monitoring should not be employed.
To promote interactive justice, there are several suggestions for suppliers. To align the expectations of retailers more closely to reality and to obtain the cooperative intentions of retailers, suppliers should explain relevant policies and decisions to retailers and achieve dual communication. With respect to attitude, suppliers should show respect, politeness, honesty and sincerity. They should care about retailers’ operations from retailers’ perspectives and show retailers that they care. Suppliers should make efforts to understand retailers’ problems and treat them as their own, trying hard to help retailers solve them.
7. Suggestions for Subsequent Researchers
Applying the concept of justice to the study of channel relationships is in its early stages. There remains a dearth of relevant studies, a gap that will require continuous efforts of subsequent researchers to fill. Although this study has obtained helpful results, many problems remain to be solved.
First, each retailer may apply different criteria to the evaluation of distributive justice. Some may believe that the benefits should be distributed according to individual effort, i.e., one who invests more effort should obtain greater rewards. Others may believe that when everyone receives equal distribution, justice is achieved, i.e., equal justice. Still others believe that benefits should be distributed according to individual needs, i.e., that one who has greater needs should receive greater benefits. The present study measures distributive justice according to the equity rule, whereby benefits are distributed according to the effort invested. Other rules, such as the equality rule and the need rule (Gassenheimer, Houston, and Davis, 1998; Nowakowski and Conlon, 2005)   , are not included in this study. Distributing benefits according to effort invested, i.e., using the equity rule, has strong motivational effects. Thus, the equity rule of benefit distribution is popular under a profit channel system. However, in other situations, as with relationships among nonprofit organizations in which harmonious relationships are emphasized, the equality rule may be more suitable. Whether distributive justice under different distributive rules would produce differing effects on relationship quality deserves discussion.
Second, the factor analysis applied to measures of continuance commitment in the pretest questionnaire revealed multiple dimensions and that the composite reliability was far less than ideal. To simplify the analysis in the study and to promote the internal consistency of continuance commitment construct measures, this study only measures continuance commitment from the perspective of “benefit” and omits items for other dimensions such as “the number of other suppliers that can be chosen” or “switching cost”. There are great differences between these dimensions and “benefit”. We suggest that subsequent studies compare the results when continuance commitment is expressed by different dimensions.
Third, because this study is a cross-sectional study, the results do not persuasively support causality among variables. If possible, we suggest that subsequent researchers conduct longitudinal studies by collecting data on justice, satisfaction, trust and commitment at different times.
The author is grateful to National Science Council in Taiwan for the financial support (NSC 92-2416-H-251- 010).
 Humphreys, M.A. and Williams, M.R. (1996) Exploring the Relative Effects of Salesperson Interpersonal Process Attributes and Technical Product Attributes on Customer Satisfaction. Journal of Personal Selling and Sales Management, 16, 47-57.
 Leung, K., Smith, P.B., Wang, Z. and Sun, H. (1996) Job Satisfaction in joint Venture Hotels in China: An Organizational Justice Analysis. Journal of International Business Studies, 27, 947-962.
 Tax, S.S., Brown, S.W. and Chandrashekaran, M. (1998) Customer Evaluations of Service Complaint Experiences: Implications for Relationship Marketing. Journal of Marketing, 62, 60-76.
 Teo, T.S.H. and Lim, V.K.G. (2001) The Effects of Perceived Justice on Satisfaction and Behavioral Intentions: The Case of Computer Purchase. International Journal of Retail and Distribution Management, 29, 109-124.
 Dabholkar, P.A. and Overby, J.W. (2005) Linking Process and Outcome to Service Quality and Customer Satisfaction Evaluations: An Investigation of Real Estate Agent Service. International Journal of Service Industry Management, 16, 10-27.
 Geyskens, I., Steenkamp, J.-B.E.M. and Kumar, N. (1999) A Meta-Analysis of satisfaction in Marketing Channel Relationships. Journal of Marketing Research, 36, 223-238.
 Geyskens, I. and Steenkamp, J.-B.E.M. (2000) Economic and Social Satisfaction: Measurement and Relevance to Marketing Channel Relationships. Journal of Retailing, 76, 11-32.
 Szymanski, D.M. and Henard, D.H. (2001) Customer Satisfaction: A Meta-Analysis of the Empirical Evidence. Journal of the Academy of Marketing Science, 29, 16-35.
 Nyaga, G.N. and Whipple, J.M. (2011) Relationship Quality and Performance Outcomes: Achieving a Sustainable Competitive Advantage. Journal of Business Logistics, 32, 345-360.
 Skarmeas, D. and Robson, M.J. (2008) Determinants of Relationship Quality in Importer-Exporter Relationships. British Journal of Management, 19, 171-184.
 Abdul-Muhmin, A.G. (2002) Effects of Suppliers’ Marketing Program Variables on Industrial Buyers’ Relationship Satisfaction and Commitment. Journal of Business and Industrial Marketing, 17, 637-651.
 Lam, S.Y., Shankar, V., Erramilli, M.K. and Murthy, B. (2004) Customer Value, Satisfaction, Loyalty, and Switching Costs: An Illustration from a Business-to-Business Context. Journal of the Academy of Marketing Science, 32, 293-311.
 Mayo, D.T., Richardson, L.D. and Simpson, J.T. (1998) The Differential Effects of the Uses of Power Sources and Influence Strategies on Channel Satisfaction. Journal of Marketing Theory and Practice, 6, 16-25.
 Schellhase, R., Hardock, P. and Ohlwein, M. (2000) Customer Satisfaction in Business-to-Business Marketing: The Case of Retail Organizations and Their Suppliers. Journal of Business and Industrial Marketing, 15, 106-121.
 Ruekert, R.W. and Churchill Jr., G.A. (1984) Reliability and Validity of Alternative Measures of Channel Member Satisfaction. Journal of Marketing Research, 21, 226-233.
 Kumar, N., Scheer, L.K. and Steenkamp, J.-B.E.M. (1995) The Effects of Perceived Interdependence on Dealer Attitudes. Journal of Marketing Research, 32, 348-356.
 Kim, K. and Frazier, G.L. (1997) On Distributor Commitment in Industrial Channels of Distribution: A Multicomponent Approach. Psychology and Marketing, 14, 847-877.
 Moorman, C., Zaltman, G. and Deshpande, R. (1992) Relationships between Providers and Users of Market Research: The Dynamics of Trust within and between Organizations. Journal of Marketing Research, 29, 314-328.
 Gilliland, D.I. and Bello, D.C. (2002) Two Sides to Attitudinal Commitment: The Effects of Calculative and Loyalty Commitment on Enforcement Mechanisms in Distribution Channels. Journal of the Academy of Marketing Science, 30, 24-43.
 Allen, N.J. and Meyer, J.P. (1990) The Measurement and Antecedents of Affective, Continuance, and Normative Commitment to the Organization. Journal of Occupational Psychology, 63, 1-18.
 Lee, D.-J., Sirgy, M.J., Brown, J.R. and Bird, M.M. (2004) Importers’ Benevolence toward Their Foreign Export Suppliers. Journal of the Academy of Marketing Science, 32, 32-48.
 Brown, J.R., Lusch, R.F. and Nicholson, C.Y. (1995) Power and Relationship Commitment: Their Impact on Marketing Channel Member Performance. Journal of Retailing, 71, 363-392.
 Chinomona, R. (2013) Dealer’s Legitimate Power and Relationship Quality in Gaunxi Distribution Channel: A Social Rule System Theory Perspective. International Journal of Marketing Studies, 5, 42-58.
 Gu, F.F. and Wang, D.T. (2011) The Role of Program Fairness in Asymmetrical Channel Relationships. Industrial Marketing Management, 40, 1368-1376.
 Brown, J.R., Cobb, A.T. and Lusch, R.F. (2006) The Roles Played by Interorganizational Contracts and Justice in Marketing Channel Relationships. Journal of Business Research, 59, 166-175.
 Patterson, P.G., Johnson, L.W. and Spreng, R.A. (1997) Modeling the Determinants of Customer Satisfaction for Business-to-Business Professional Services. Journal of the Academy of Marketing Science, 25, 4-17.
 Hadjikhani, A. and Thilenius, P. (2005) The Impact of Horizontal and Vertical Connections an Relationships’ Commitment and Trust. Journal of Business and Industrial Marketing, 20, 136-147.
 Smith, A.K., Bolton, R.N. and Wagner, J. (1999) A Model of Customer Satisfaction with Service Encounters Involving Failure and Recovery. Journal of Marketing Research, 36, 356-372.
 Gassenheimer, J.B., Calantone, R.J. and Scully, J.I. (1995) Supplier Involvement and Dealer Satisfaction: Implications for Enhancing Channel Relationships. Journal of Business and Industrial Marketing, 10, 7-19.
 Crosby, L.A., Evans, K.R. and Cowles, D. (1990) Relationship Quality in Services Selling: An Interpersonal Influence Perspective. Journal of Marketing, 54, 68-81.
 Jarvenpaa, S.L., Knoll, K. and Leidner, D.E. (1998) Is Anybody Out There? Antecedents of Trust in Global Virtual Teams. Journal of Management Information Systems, 14, 29-64.
 Meyer, J.P., Allen, N.J. and Smith, C.A. (1993) Commitment to Organizations and Occupations: Extension and Test of a Three-Component Conceptualization. Journal of Applied Psychology, 78, 538-551.
 Verhoef, P.C., Franses, P.H. and Hoekstra, J.C. (2002) The Effect of Relational Constructs on Customer Referrals and Number of Services Purchased from a Multiservice Provider: Does Age of Relationship Matter? Journal of the Academy of Marketing Science, 30, 202-216.
 Gaski, J.F. and Nevin, J.R. (1985) The Differential Effects of Exercised and Unexercised Power Sources in a Marketing Channel. Journal of Marketing Research, 22, 130-142.
 Gassenheimer, J.B., Houston, F.S. and Davis, J.C. (1998) The Role of Economic Value, Social Value, and Perceptions of Fairness in Interorganizational Relationship retention Decisions. Journal of the Academy of Marketing Science, 26, 322-337.