Abstract
Applying the six-factor model of organizational justice, this study examines the relationship between disputants’ (i.e., grievants and respondents) perceptions of organizational justice and satisfaction with workplace mediation. Using secondary data, collected postmediation from participants in the (former) North Carolina Department of Correction’s (DOC) mediation process, the findings show that perceptions of organizational justice and mediation satisfaction are high for both grievants and, especially, respondents. Logistic regression results find statistically significant relationships between mediation satisfaction and three factors of organizational justice—distributive justice, procedural justice–process, and disputant–disputant interpersonal justice—as well as unexpected results for procedural justice–mediator and disputant–mediator interpersonal justice.
Keywords
Introduction
Organizational justice is concerned with fairness. It is “fundamentally about how employees construct meaning around their treatment in the workplace—whether they perceive themselves to have been treated fairly and how this perception impacts the nature of the employment relationship” (Sowa, 2016, p. 228). Employees’ perceptions about how their organization treats them have important effects on attitudes, behaviors and, in turn, organizational outcomes. Employee discipline and grievance systems represent critical arenas for shaping perceptions and ensuring the provision of organizational justice (Folger & Greenberg, 1985; Fryxell & Gordon, 1989).
In the public sector, substantive and procedural due process protections are built into grievance and discipline systems. Grounded in constitutional and administrative law (Hassan, 2013; Rubin, 2009) and reflecting an underlying bureaucratic paradigm (Moynihan & Ingraham, 2012), these systems shape collective views of organizational justice through individual cases (Colquitt, Conlon, Wesson, Porter, & Ng, 2001). Grievances are employee initiated and entail allegations that the employer failed to adhere to organizational objectives and procedures. Such allegations are often related to the denial of a work-related benefit (e.g., pay, promotion, job assignment). Discipline is employer initiated, with the employer (manager or supervisor) claiming that the employee broke a rule or did not follow appropriate procedures for achieving the organization’s objectives (Haraway, 2005; Posthuma, 2003). It entails the imposition of a penalty, which in a progressive discipline system may lead to or—depending on the severity of the offense—result in employee demotion or dismissal. The perception of fairness is particularly important given the seriousness of these disciplinary consequences.
Organizational justice is a salient issue in light of ongoing developments in the field of public human resource management. Traditional, merit-based civil service systems that promote politically neutral competence and professionalism and protect public servants from political interference and abuse are increasingly under siege (Brewer & Walker, 2013; Kearney & Coggburn, 2016; Selden, 2006). Over roughly the last 25 years, civil service reform at the local, state, and federal levels has altered the status of public employees and the public employer–employee relationship (Coggburn et al., 2010; Condrey & Battaglio, 2007; Hays & Sowa, 2006; Kearney & Coggburn, 2016; Light, 2006). Merit-based systems are often criticized as cumbersome, inflexible, inefficient, unresponsive (i.e., to political/executive control and citizens), and outdated, hence civil service reform has targeted procedural safeguards, especially those embodied in employee discipline and grievance systems.
1
As Rubin (2009, p. 129) noted, Disciplinary, appeal, and termination procedures are arguably the most rule-bound of government personnel management rules. The rules require documentation, opportunities for improvement, and opportunities for hearings. It is claimed that such requirements create a disincentive for managers to discipline an employee or begin termination proceedings.
In some settings, employees’ procedural protections have been substantially reduced or, with employment at will, eliminated (Coggburn, 2006; Coggburn et al., 2010; Hays & Sowa, 2006). This is a problematic trend given the potential harm done to employees and their innate expectations of organizational justice (Boswell & Olson-Buchanan, 2004; Daley, 2007; Rubin, 2009; Rubin & Kellough, 2012; Sowa, 2016).
A different tact to addressing concerns with grievance and discipline systems focuses on overall dispute system design (DSD). Originating with Ury, Brett, and Goldberg (1988), DSD research provides guidelines for developing grievance and disciplinary processes that are less cumbersome for managers and perceived as fair by employees. This work shows that employees’ perceptions of fairness increase when disputing parties focus on interests, employees retain a measure of control over the outcome (Blancero & Dyer, 1996; Brett, Barsness, & Goldberg, 1996; Gordon & Fryxell, 1993), and employees believe their position truly has been considered (Shapiro, 1993).
Reflecting a DSD approach, some public employers have adopted alternative dispute resolution (ADR) to manage conflict and pursue organizational justice. There are a number of ADR processes (Lipsky, Seeber, & Fincher, 2003; U.S. OPM 2001), with mediation being the most frequently adopted in the public sector (Bingham, 2004; Mareschal, 2003; McKenzie, 2015). Including mediation in the grievance process follows several of Ury et al.’s (1988) recommendations for effective DSD. First, mediation is an interests-based approach in which a neutral third party encourages disputants to negotiate underlying interests rather taking adversarial rights-based positions. Second, mediation gives grievants (employees) the opportunity to speak for themselves, express emotions, and participate in the development of the final agreement, which increases their satisfaction with the process and outcome. Finally, mediation is cost-effective, being both timelier and less expensive than a typical grievance hearing. In contrast to eliminating procedural protections, mediation holds promise for balancing employers’ interests in both managing conflict and controlling costs (e.g., temporal, pecuniary, opportunity) and frustration associated with formal grievance processes (Nabatchi & Stanger, 2013) with employees’ interests in due process and organizational justice (Sowa, 2016).
Despite its potential, there is a general paucity of empirical research on workplace mediation (Bollen & Euwema, 2013), including work focused on organizational justice in the public sector. This study seeks to extend knowledge of organizational justice in the context of workplace mediation through an analysis of the North Carolina Department of Correction’s (DOC) 2 facilitative mediation process. The key research question is as follows:
In addressing this question, the study applies Nabatchi, Bingham, and Good’s (2007) six-factor model of organizational justice (see also Nesbit, Nabatchi, & Bingham, 2012). The study begins with descriptions of organizational justice and workplace mediation and a brief review of the associated research. Next, the context of the DOC mediation program is described. This is followed by a discussion of the data and methods used to assess the six-factor model. Finally, the logistic regression results are presented along with a discussion of their implications.
Organizational Justice
Organizational justice is “ . . . concerned with the ways in which employees determine if they have been treated fairly in their jobs and the ways in which those determinations influence other work-related variables” (Moorman, 1991, p. 845). 3 In comparison with business management, organizational behavior, and organizational psychology, research on organizational justice in public administration is sparse (Cho & Sai, 2012; Choi, 2011; Rubin, 2009). Existing work tends to corroborate findings from the other disciplines 4 showing that organizational justice promotes multiple positive outcomes for both individuals and organizations. Positive perceptions of organizational justice have been linked to greater job involvement, organizational identification, expectations for career development, job satisfaction, organizational citizenship behaviors, trust, and cooperation and lower absenteeism and reported turnover intent (Cho & Sai, 2012; Choi, 2011; Hassan, 2013; Rubin, 2009).
Within these studies, public administration scholars have assessed different facets of organizational justice. This reflects organizational justice’s development as a multidimensional construct (Cohen-Charash & Spector, 2001; Greenberg, 1993; Nabatchi, Bingham, & Good, 2007). As such, researchers have found that the different forms of organizational justice have independent effects on perceptions and workplace attitudes (Choi, 2011).
The Four-Factor Model
The most widely accepted conceptualization of organizational justice is the “four-factor model” consisting of distributive, procedural, informational, and interpersonal justice (Colquitt, 2001; Greenberg, 1993; Tyler & Bies, 1990).
Drawing on equity theory (Adams, 1965), distributive justice refers to the perceived fairness of outcomes (e.g., compensation, promotions, or other rewards), where individuals’ perceptions are shaped through comparisons of their outcomes relative to others. In this way, distributive justice relates to individuals evaluating and recognizing their own self-interests (Nabatchi & Bingham, 2010). Distributive justice has been found to be among the most important justice factors in shaping perceptions of trust, job satisfaction, job involvement, and employees’ intentions to leave the organization (Choi, 2011; Hassan, 2013). In the context of mediation, distributive justice would suggest satisfaction is a product of the outcome, that is, the content of the case’s resolution or agreement (Nabatchi et al., 2007). Research shows that distributive justice has strong effects on complainant–respondent agreement regarding dispute settlement (Nesbit et al., 2012) and that respondents have higher levels of satisfaction with mediation outcomes relative to complainants (Bingham, 2004, 2012).
Procedural justice shifts focus from the fairness of outcomes to the processes (e.g., rules, procedures) through which those outcomes are decided. In other words, procedural justice perceptions reflect judgments about the degree to which individuals view decision making as fair (Rubin, 2009) thus making mediation satisfaction “a function of process” (Nabatchi et al., 2007, p. 150). Public administration researchers have given considerable attention to procedural justice, consistently finding it to have a strong influence on employee perceptions and organizational outcomes (Rubin, 2009; Rubin & Chiques, 2015; Rubin & Kellough, 2012; Rubin & Weinberg, 2016). This is especially true for those higher in the organizational hierarchy who derive value messages from how they are treated (Potipiroon & Rubin, 2018; Rubin & Chiques, 2015). And, research shows that perceptions of procedural justice are particularly salient to assessments about the effectiveness of grievance systems, including mediation (Walker & Hamilton, 2011).
The final two factors of the four-factor model focus on individuals’ behavior and interactions—that is, how they communicate, what information they provide, and their personal conduct—in the process of reaching decisions (Cho & Sai, 2012). Interpersonal justice recognizes a two-way interaction between the employee and supervisor, the quality of which influences perceptions of fairness. Specifically, an employee (grievant) experiences interpersonal justice when they perceive polite and respectful treatment by their supervisor or the respondent (Nabatchi et al., 2007). Disputants with higher organizational status (e.g., supervisors and managers) may be particularly attuned to interpersonal justice perceptions because their social identity is verified when they experience fair and respectful treatment from the other party (Potipiroon & Rubin, 2018). The salience of interpersonal justice to higher-status employees also appears to hold in the context of workplace mediation (Nesbit, Nabatchi, & Bingham, 2004; Nesbit et al., 2012).
Finally, informational justice “focuses on the enactment and explanation of decision-making procedures” (Nabatchi et al., 2007, p. 151; see also Nesbit et al., 2004). It has been found to be an important predictor of employee loyalty and cooperation (Cho & Sai, 2012). In the context of mediation, informational justice indicates that the process has been explained such that disputants understand how decisions about the case will be resolved and by whom. Previous work has found this factor of justice to be important for employees’ assessments of dispute resolution but not for supervisors, which is likely due to supervisors being more experienced with and knowledgeable of the mediation process than employees (Nesbit et al., 2012).
The Six-Factor Model
Due to the addition of a third party as the mediator, Nabatchi et al. (2007, p. 149) argued that the four-factor model of organizational justice fails to adequately account for the “multiple and divergent directions of interaction” created through the mediator’s involvement in conflict management. In other words, the four-factor model only applies in a bidirectional process, such as between a supervisor and employee, as opposed to the multidirectional process among a supervisor, employee, and mediator.
The authors propose and validate a six-factor model of organizational justice for workplace mediation (Nabatchi et al., 2007). Their approach extends the four-factor model by splitting procedural justice into two factors: procedural justice–process and procedural justice–mediator. Doing so aims to “facilitate research on organizational justice by enabling more direct comparisons of the perceived fairness of the process, as well as. . . the person who enacts the process” (Nabatchi et al., 2007, p. 153). Procedural justice–mediator recognizes mediation’s success is contingent upon both participants accepting the mediator as an “honest broker” who is able to counter the inherent power imbalance in disputes involving employees and managers, thereby ensuring that both positions receive due consideration (McKenzie, 2015).
Similarly, the six-factor model splits interpersonal justice into two factors: disputant–disputant interpersonal justice and disputant–mediator interpersonal justice. This theoretical refinement aims to “distinguish between interpersonal justice as a function of the relationship between the disputing parties and interpersonal justice as a function of the relationship between a disputant and the third party neutral [mediator] who is enacting the dispute resolution process” (Nabatchi et al., 2007, p. 153; see also Nabatchi & Bingham, 2010). The resulting six-factor model consists of distributive justice, procedural justice–process, procedural justice–mediator, informational justice, disputant–mediator interpersonal justice, and, disputant–disputant interpersonal justice.
In the only known test of the six-factor model, Nesbit et al. (2012) found that it generated mixed results in explaining corroborated resolution (i.e., where both parties agreed mediation resolved the conflict). They found that distributive, procedural justice–process, and (for employees) informational justice were significant predictors of corroborated mediation resolution. Conversely, the mediator-based factors proved less important: procedural justice–mediator failed to attain statistical significance in any of the eight tests whereas of disputant-mediator interpersonal justice attained it in just one (Nesbit et al., 2012). The unexpected findings for the mediator-based factors were attributed to the attenuated role mediators play in controlling the process and outcome of transformative mediations and accompanied by a call for additional tests of the six-factor model in different workplace mediation contexts. The following two sections describe this study’s approach to meeting that call.
Workplace Mediation
In the context of workplace grievances, mediation represents an informal procedure that has the advantage of not creating a permanent record officially blaming or declaring someone at fault. It affords grievants’ and respondents’ voice, the opportunity to provide information and explanations, and control over the outcome (Brett et al., 1996). Research has shown that workplace mediation produces a number of important organizational and individual benefits (Bollen & Euwema, 2013). These benefits include allowing conflicts to be resolved much earlier in the process than formal, adjudicative alternatives (Bingham & Novac, 2001; Nabatchi & Stanger, 2013); reducing the number of formal complaints filed (Bingham & Novac, 2001); providing an opportunity for supervisors and employees to practice and improve their conflict management and communication skills (Bingham, Hallberlin, Walker, & Chung, 2009); reducing occupational stress (Tallodi, 2015); and improving workplace climate and (Bingham et al., 2009; Bingham, Raines, Hedeen, & Napoli, 2010).
In assessing perceptions and outcomes associated with mediation, researchers have focused attention on certain aspects of DSD. For current purposes, the key aspects are the mediation style, type of neutral who mediates, and type of disputes that can be mediated.
Mediation Style
Although there is no consensus on how to classify mediation styles (Charkoudian, 2012), distinctions are typically made among facilitative and evaluative mediation, both of which are settlement-based, and transformative mediation, which emphasizes empowering participants and achieving recognition and understanding (Bush & Folger, 1994, 2005); Nabatchi et al. (2007) noted that, In facilitative mediation, the mediator structures the process for the parties and engages in problem-solving techniques, including reality testing, to move the parties toward settlement. In evaluative mediation the mediator assesses the legal and substantive merits of the claim, and gives the parties information about the strengths and weaknesses of the case. (p. 154)
In contrast, transformative mediation emphasizes improving relationships among participants by empowering them to make their own decisions while promoting recognition of the other party’s circumstances and position (Bush & Folger, 1994, 2005). It operates from the assumption that if parties can overcome differences and make a human connection (i.e., through empowerment and recognition), an agreement will either follow or be rendered unnecessary.
Mediation represents the predominant ADR approach, yet, research on workplace mediation is limited, tends to be descriptive, and offers little empirical evidence on the outcomes associated with different mediation styles (Binkert, 2016; Bollen & Euwema, 2013; Nabatchi et al., 2007). Researchers face cost, access (i.e., to actual mediation sessions), and time challenges that impede comparisons of mediation styles and their associated outcomes (Bingham, 2004; Bollen & Euwema, 2013; Wall & Kressel, 2012). The notable exception is the substantial body of work 5 focused on the U.S. Postal Service’s (USPS) Resolve Employment Disputes, Reach Equitable Solutions Swiftly (REDRESS) program (e.g., Nabatchi et al., 2007; Nesbit et al., 2012), which uses external neutrals and transformative mediation to address disputes related to alleged workplace discrimination.
As REDRESS uses transformative mediation, much of the existing empirical research focuses on that style of mediation. As Bollen and Euwema (2013) noted, “mediation research has focused on mediation designed to foster parties’ understanding and recognition (factors that prime readiness) rather than on workplace mediations that are especially designed to achieve settlement” (p. 345). Nabatchi and Bingham (2001) found that USPS specialists charged with selecting mediators for REDRESS understood and could correctly describe transformative mediation behaviors, thus ensuring transformative practices. Nabatchi, Bingham, and Moon (2010) found that mediators in the national pool for REDRESS agreed on a number of statements that reflected transformative mediation and could correctly identify mediation behaviors as being transformative or evaluative, respectively.
Other researchers have made efforts to distinguish mediation styles and assess their associated outcomes. Charkoudian (2012) found evidence of four mediator styles, ranging from highly directive and evaluative to nondirective and transformative. Bingham (1997) found high levels of employee and supervisor satisfaction with facilitative mediation, including the mediation process, role of the mediator, and outcomes. Subsequent work found satisfaction with transformative mediation mirrored that of facilitative (Bingham et al., 2010; Moon & Bingham, 2007), with employees and supervisors being satisfied with the process and mediators and employees being slightly less satisfied than supervisors with outcomes. Kressel, Henderson, Reich, and Cohen (2012) found that mediators labeled as transformative or evaluative were rated as less effective than those classified as facilitative. And, in a laboratory comparison of mediators who used either a more emotion-focused or strict problem-solving style in role-played mediations, high satisfaction with the mediation process and outcomes were found regardless of mediation condition (Jameson, Bodtker, Porch, & Jordan, 2009). Such findings have led to the suggestion that mediation’s stylistic differences may be overstated (Pruitt, 2012). Regardless, knowledge of mediation in workplace grievance focuses heavily on transformative mediation, thus leaving ample opportunity for researchers to consider other mediation styles.
Type of Third-Party Neutral
A second design aspect of mediation is its use of a neutral third party who has no relationship with the grievant or respondent nor any vested interest in the mediation outcome. Although managers may sometimes act like mediators in trying to facilitate an agreement among employees (or supervisors and employees), they are not truly neutral because of their status, relationship to disputants, and investment in the conflict outcome (Bingham, Chesmore, Moon, & Napoli, 2000; Karambayya & Brett, 1989).
Workplace mediation programs concerned with organizational justice, therefore, use either internal (inside), external (outside), or shared neutrals. Internal neutrals are employees of the agency or organization. In addition to cost savings, an advantage of internal neutrals is that they are familiar with the organizational context, giving them insights that someone unfamiliar with the organization would not have. There are, however, disadvantages of internal neutrals in that they threaten the anonymity of disputants and, as they work for the organization, may find it difficult to be truly neutral (Bingham et al., 2000) or viewed as such by employees. External neutrals cost more to the organization but they have no relationship with employees or management and, therefore, may be more trusted by employees. Bingham et al. (2000) conducted the only known empirical investigation that compared organizational justice perceptions of internal versus external mediators. They found that the type of mediator had bearing on reported satisfaction with mediation, including the mediation process, mediator, and outcome: In each instance, external mediators were rated significantly higher than internal mediators. Furthermore, mediations were significantly more likely to come to a complete agreement when externals were used (see also Bingham & Pitts, 2002).
Finally, shared neutrals are employees from one agency who are used as a mediator in another agency. Shared neutrals have been used at the federal level (Bingham et al., 2000) but other levels of government can also train employees across agencies to create a shared pool of mediators so that a mediator is never a member of the same agency as the disputants. The use of shared neutrals has a number of benefits. For example, the State of Indiana (2014) noted their use improves access to mediation due to reduced costs, enhances perceptions of mediator neutrality, increases the willingness of parties to participate in mediation, and increases opportunities for state employees to use and practice conflict management skills. Shared neutrals, therefore, may represent the best of both worlds, as mediators are internal in that they have some contextual understanding of government rules and procedures, yet they are external to the specific agency and not directly connected to the disputants. To date, research on shared neutrals is limited (but see Waller, 2000), and there are no known studies assessing organizational justice in a mediation model using shared neutrals.
Type of Dispute Being Mediated
The types of disputes eligible for mediation is an important DSD consideration because too many issues could overwhelm an agency and other issues could have broader (agency- or government-wide) implications if included (Sowa, 2016). Yet, due to the limitations of access to employment conflict data, it is often difficult to ascertain the specific types of disputes that are included in studies of mediation and perceptions of organizational justice. Bollen, Ittner, and Euwema (2012), for example, compared supervisors’ and employees’ perceptions of procedural justice in hierarchical labor conflicts, but offered no more specific information about conflict type.
Most mediation studies examine mediation in the context of labor relations and collective bargaining agreements, such as Ury et al.’s (1988) work on the coal mining industry (see also Bingham et al., 2000). These studies have consistently shown that grievants are highly satisfied with the mediation process (Brett & Goldberg, 1983) and that perceptions of procedural justice are higher in mediated than arbitrated cases (Shapiro & Brett, 1993). While also in the context of management–labor disputes, the USPS REDRESS mediation program specifically addresses EEOC (Equal Employment Opportunity Commission) complaints of discrimination in employment based on race, sex, national origin, ethnicity, religion, age, disability, and other protected categories (Nesbit et al., 2012). The series of studies on REDRESS mediation program found high levels of satisfaction with process and outcome in these cases, especially when there were high perceptions of interpersonal justice and the grievant received an apology. This is logical given the sensitive nature of discrimination claims (Nesbit et al., 2012). Overall, however, there is a need for more clarity in mediation research about the specific issues being mediated and work that considers issues beyond workplace discrimination and harassment.
In sum, the literature on organizational justice in the context of workplace mediation focuses primarily on transformative mediation, using external neutrals, and addressing issues of workplace harassment and discrimination. Thus, there are opportunities for additional work that considers different types of mediation, mediators, and issues.
Facilitative Mediation at the NC DOC
Given the importance of context to understanding workplace mediation structure and functioning (Bingham, 2004; Bollen & Euwema, 2013; Nabatchi & Bingham, 2010), this section describes the North Carolina DOC and its mediation process. Prior to its 2013 consolidation into the Department of Public Safety (DPS), DOC was one of the largest state agencies, with 19,706 employees and a budget of approximately US$1.4 billion (State of North Carolina, DPS, 2012). DOC “[was] responsible for the custody, supervision, and rehabilitation of adult offenders sentenced to community/intermediate punishments or prison” (State of North Carolina, DOC, 2011, p. 6). DOC operated 67 prisons, housing over 40,000 inmates, and supervised another 108,000 offenders in the community through probation, parole, and postrelease responsibilities. Most DOC employees worked directly with offenders, which required strict adherence to law, policy, and rules for behavior. Employees facing disciplinary action for performance or policy violations were accorded due process through access to an employee grievance system.
Grievance and Mediation Process
The State of North Carolina mandates due process protections for its employees and requires “just cause” for disciplinary actions toward permanent (nonprobationary) state employees. 6 The state’s Office of State Human Resources (OSHR) spells out these procedural safeguards in the State Human Resources Manual. Beginning in 2005, this included two separate policies, the original Employee Appeals and Grievance Policy, which did not include mediation, and the optional Employee Grievance and Mediation Process, which did. 7 The manual listed three objectives for the mediation process:
Assure employees have access to grievance procedures to address their grievable issues rapidly, fairly, and without fear of reprisal,
Contain costs to process grievances, and
Resolve workplace issues and problems efficiently and effectively.
In 2013, the two grievance processes were consolidated into the current uniform Employee Grievance Policy, which includes mediation as a required first step. Prior to policy consolidation, state agencies could choose to adopt the mediation and grievance policy or operate under the original appeals and grievance policy. DOC adopted the mediation process, which called for employees and supervisors to first meet informally to resolve disputes. Failing success at informal resolution, the grievance shifted to a formal process that included three steps: (a) mediation, (b) internal agency hearing, and (c) external appeal. Mediation could result in either an agreement (ending the grievance) or impasse. In cases of an impasse, the grievant could pursue an internal hearing where an officer or panel would render a recommendation to the agency head for a final agency decision. For items defined in the State Human Resources Act as appealable, the grievant could appeal the final agency decision to the state’s Office of Administrative Hearings (OAH) where an administrative law judge would render a final recommendation to the State Human Resource Commission. 8 These steps are consistent with recommendations from the DSD literature in that they attempt to use an Interest-based process (mediation) before moving to the more costly and time-consuming rights-based process (Brett, Goldberg, & Ury, 1990; Ury et al., 1988).
At the time of DOC’s policy adoption, agencies were given flexibility to decide, which grievable issues under the State Human Resources Act were eligible for mediation. The DOC limited those issues to disciplinary cases—written warnings, demotions, and dismissals. More complicated and uncertain questions arising from work rule interpretations were excluded, as were those issues (e.g., workplace discrimination and harassment) governed by separate state policies.
DOC employed a full-time coordinator to organize the mediation process. The vast majority of mediators were shared neutrals drawn from other state agencies, though they were occasionally augmented with qualified nonstate employee mediators. The shared neutrals received no compensation (other than travel reimbursement) for facilitating mediations. Prior to leading a mediation, a candidate mediator received 40 hr of training, which OSHR explicitly described as facilitative mediation, and was required to serve in an apprentice role (i.e., as comediator) in at least four mediations under the tutelage of an experienced mediator. The aim was to develop the knowledge of mediation and practical skill needed to help ensure successful mediation (see Goldman, Shapiro, & Pearsall, 2016; Raines, 2000). All mediators were required to abide by the OSHR’s “Mediator Code of Conduct,” which include tenets ensuring their impartiality, respectful conduct, and confidentiality and, consistent with facilitative mediation, prohibiting their issuance of legal or professional advice to either party.
Finally, information about the grievance and mediation policy was made available through DOC and OSHR, as was mediation training for respondents. This information clarified the purpose of mediation, roles of participants, and details about the mediation process as well as offered tips for mediation success. 9
Given this context, DOC’s mediation program offers the opportunity to extend previous research by assessing organizational justice under a different style of workplace mediation (facilitative), using a different pool of mediators (shared neutrals), and addressing a different set of issues (disciplinary). The next section describes the data and methods used to examine how DOC mediation participants’ perceptions of organizational justice relate to their satisfaction with mediation.
Model, Data, and Methods
This study applies the six-factor model of organizational justice to DOC’s facilitative mediation program (Nabatchi et al., 2007; Nesbit et al., 2012). Results are first estimated for a combined test of the model, controlling for participants’ role (i.e., grievant or respondent). Next, given previous findings showing that supervisor and subordinate perceptions of organizational justice differ (Bollen et al., 2012; Rubin & Chiques, 2015; Rubin & Weinberg, 2016) and following Nesbit et al.’s (2012) analytic approach, separate tests of the six-factor model are estimated for grievants and respondents, respectively.
Following most workplace mediation research (Bollen & Euwema, 2013), this study uses postmediation survey data obtained directly from participants. At the conclusion of each DOC mediation, participants received a brief (i.e., 14 item) evaluation survey addressing levels of agreement or disagreement (on a 5-point Likert-type scale, ranging from strongly agree to strongly disagree) with statements related to the interactions between disputants and disputants and mediators, mediation process, and mediation outcome. It is important to note that DOC developed the survey for its own internal administrative and evaluative purposes. As such, it did not include the same number or wording of items used in the seminal work operationalizing the six-factor model (Nabatchi et al., 2007; Nesbit et al., 2012). Given that limitation, this study follows Rubin and Weinberg’s (2016) strategy of selecting survey items according to their congruence with established theory and measures previously used in the organizational justice and mediation literature. It is fortuitous that the DOC survey included 10 items that allow for a reasonable approximation of the six-factor model of organizational justice 10 and a meaningful mediation outcome measure. The internal reliability of the three factors measured with multiple survey items will be assessed with Cronbach’s alphas. Given the noted data limitations (e.g., the ratio of items to factors), 11 however, a confirmatory factor analysis (CFA) is not suitable (Brown & Onsman, 2010; Comrey & Lee, 1992; Marsh, Hau, Balla, & Grayson, 1998) and will not be reported. Such an approach emphasizes the scholarly contribution of testing theory in a new context with reasonable yet, admittedly, imperfect measures over a more sophisticated yet, given the current data limitations, inappropriate methodological approach (Davis & Stazyk, 2017).
An additional limitation is the exclusion of demographic data (e.g., age, years of service, gender, race, and ethnicity) from the agency’s survey. 12 The total number of survey responses was 766, of which 664 (including 335 grievants, 329 respondents) are used in the analysis due to missing cases.
Dependent Variable: Mediation Satisfaction
Research on mediation outcomes typically focuses on short-term outcomes, including measures of disputant perceptions, settlement rates, and cost effectiveness (Bollen & Euwema, 2013). Likewise, the current analysis uses a single survey item—I felt positive about the overall mediation process—as an indicator of disputants’ overall satisfaction with DOC’s workplace mediation. As depicted in Herrman’s (2006) mediation model, satisfaction with mediation is a disputant-based perception representing a short-term outcome and the most common metric for assessing resolution (Bingham, 2012). Respondents who agreed or strongly agreed with the statement are coded “1” and all others are coded “0.” Given the dichotomous nature of the dependent variable, the tests of the six-factor model use logistic regression.
Explanatory Variables: The Six-Factor Model of Organizational Justice
Organizational justice is assessed with a set of variables that approximate the six-factor model. In contrast to Nabatchi et al. (2007), where multiple indicators were developed for each of the six factors, the number of survey items included in DOC’s was limited. However, the DOC survey did contain a sufficient number of items having face validity as indicators of organizational justice. In the three instances, where multiple item indicators are used, the measures hold in respect to Cronbach’s alpha reliability measures (Table 1).
Mediation Satisfaction and Organizational Justice.
Note. Cell entries are the percentages of indicating agreement or strong agreement with each item.
DV = dependent variable.
Distributive justice is measured with a single item—Mediation resolved the dispute to my satisfaction—that captures a participant’s holistic view of the mediation’s outcome. Based on previous research (Bingham, 2004; Bingham et al., 2009), a positive relationship between distributive justice and mediation satisfaction is expected.
Procedural justice–process focuses on the mechanics of the mediation itself. It balances the ability of participants to express their views with the timeliness of the process. Parties to a dispute must be able to tell their side of the story and to do so in a positive environment. It is measured with a two-item index (Cronbach’s α = .69): Mediation provided a positive opportunity to discuss this grievance,” and “I had sufficient time to tell my side of the story. Procedural justice–process is expected to increase the odds of participant satisfaction with mediation.
Procedural justice–mediator addresses the extent to which the mediator and comediator are viewed as being fair and impartial. Given that mediators may play a somewhat more directive role in facilitative relative to transformational mediation (Nesbit et al., 2012), their influence could be more pronounced in shaping satisfaction with mediation. This aspect of organizational justice is measured with a two-item index (Cronbach’s α = .86): The mediator was fair and impartial, and The comediator was fair and impartial. Procedural justice–mediator is expected to increase the odds of mediation satisfaction.
Informational justice notes the degree to which the mediation process is clearly explained to and understood by participants. It is measured with a single item: The mediator(s) explained the mediation process to my satisfaction. It is expected that perceptions of informational justice will increase the odds of reported satisfaction with mediation.
Disputant—mediator interpersonal justice assesses the extent to which mediators accord each participant respect and treat them with civility. This is an underpinning if mediators are to maintain their role as neutral third parties. This factor is measured with a two-item index (Cronbach’s α = .88): The mediator treated me with respect, and The comediator treated me with respect. The expectation is that disputant–mediator interpersonal justice will increase the odds of satisfaction with mediation.
Disputant–disputant interpersonal justice examines grievants’ and respondents’ perceptions of having been treated with respect by the other party. It is measured here with a single item: The agency respondent/grievant treated me with respect. Perceptions of disputant–disputant interpersonal justice should increase the odds of reported satisfaction with mediation.
Results
Table 1 reports survey responses to the mediation satisfaction and organizational justice items contained in the DOC postmediation survey. In general, the univariate analysis shows that a strong majority (78%) report overall satisfaction with mediation, with grievants reporting less satisfaction (63% satisfied) than respondents (92% satisfied). Similarly, results show that solid majorities of grievants and respondents agree that they experienced most factors of organizational justice during mediation, including procedural justice (process and mediator), informational justice, and interpersonal justice (disputant–disputant and disputant–mediator). As with overall satisfaction with mediation, results for the organizational justice items show levels of agreement (i.e., those who “strongly agree” or “agree”) that are consistently higher for respondents relative to grievants. 13 The starkest contrast between grievant and respondent perception is for distributive justice, where a majority (70%) of respondents report satisfaction with the outcome relative to a sizable minority (40 percent) of grievants.
Logistic Regression Results
Table 2 reports logistic regression results for three tests of the six-factor model of organizational justice and its effects on the odds of mediation satisfaction. The table reports adjusted odds ratios, exp(B), for the six factors organizational justice. The odds ratios are interpreted as the increase (odds ratio > 1.0) or decrease (odds ratio < 1.0) in the odds of the dependent variable occurring, which means here the increase or decrease in the odds of reporting satisfaction with mediation associated with a specific factor of organizational justice, holding the other factors constant. The first column of data reports results for an initial test of the six-factor model that includes all participants (the “combined test”), controlling for their role (grievant or respondent) in the mediation. The second and third columns report separate test results for grievants and respondents, respectively. Overall, each test of the six-factor model is statistically significant and, as indicated by Nagelkerke R2 statistics, accounts for a sizable portion (i.e., 50% for the combined, 46% for the respondents only, and 44% for the grievants only) of the variance in reported satisfaction with the mediation process.
Estimated variable coefficients are reported as adjusted odds ratios; exp(B) and levels of significance = *p < .05. **p < .01.
The dependent variable, mediation satisfaction, is a dichotomous measure with “satisfied” as the target category and “not satisfied” as the reference category.
Combined test results
Looking first at the combined test of the model (i.e., for grievants and respondents), the first data column shows three factors of organizational justice—distributive justice, procedural justice–process, and disputant–disputant interpersonal justice—have a statistically significant relationship to mediation satisfaction. The odds ratio for distributive justice (2.12, p < .01) indicates that a unit increase in reported satisfaction with the mediation’s outcome roughly doubles the odds of reporting satisfaction with the overall mediation process. Likewise, the odds ratio for procedural justice–process (3.03, p < .01) indicates that the perceived fairness of the mediation process has a significant influence on mediation satisfaction, which represents a threefold increase in the odds of reported satisfaction with the mediation process, the largest effect reported for the combined test. The odds ratio for disputant–disputant interpersonal justice (1.40, p < .05) indicates that those experiencing respect from the other party during mediation are significantly more likely to report satisfaction with mediation. Finally, and consistent with previous research noting differences based upon disputants’ hierarchical status, the variable controlling for participants’ role type is significant. Given the coding (0 = respondents, 1 = grievant), the odds ratio (.38, p < .01) suggests that being a grievant reduces the odds of reporting satisfaction with the mediation process by about 62% as compared with a respondent, a finding that corroborates results reported above in the univariate analysis.
Separate test results for grievants and respondents
The results for the grievants- (second column) and respondents-only (third column) tests of the six-factor model offer additional insights. Looking first at the results for grievants, the estimated effect of distributive justice is very similar to those from the combined test: Grievants’ positive perception of distributive justice more than doubles the odds of satisfaction with mediation, increasing it by a factor of 2.26 (p < .01). The odds ratio for procedural justice-process (1.94, p < .05) is also significant but suggests a smaller effect for grievants than observed in the combined participants’ test. Still, these factors of organizational justice provide substantial increases in the odds of mediation satisfaction among grievants, a finding that seems particularly important given the severity of disciplinary issues (i.e., written warnings, demotion, and dismissal) being mediated.
As for the respondents-only test of the six-factor model, the effect of distributive justice (odds ratio = 1.95, p < .01) on the odds of mediation satisfaction is consistent with the combined and grievants only tests. What is noticeably different is the very large odds ratio for procedural justice-process (12.83, p < .01), which means that higher scores on this index have a strong, positive relationship with the odds of respondents’ reporting satisfaction with mediation. Also of note is the result for the disputant–disputant interpersonal justice variable, which was significant in the combined and the respondents-only tests. (For grievants, reporting the experience of disputant–disputant interpersonal justice had a positive but attenuated effect relative to respondents.) For respondents, experiencing disputant–disputant interpersonal justice about doubles (2.18, p < .01) the odds of mediation satisfaction.
The results for the remaining factors of organizational justice—informational justice, procedural justice–mediator and disputant–mediator interpersonal justice—are not statistically significant in any of the three tests of the six-factor model. For informational justice, the results suggest that it decreases the odds of mediation satisfaction (odds ratio < 1.0) in the combined and respondents-only test but increases them (odds ratio > 1.0) in the grievants-only test. In the case of procedural justice–mediator, the odds ratios were below 1.0 in all three tests, which would indicate, counterintuitively, a negative effect of mediator fairness on mediation satisfaction. Similar results were found for disputant–mediator interpersonal justice, though the results were positive in the combined (odds ratio = 1.30) and grievant-only (odds ration = 1.10) models.
Another look at the mediator-based factors
Given the insignificant statistical test results for the mediator-based factors and their integrality to the six-factor model, additional tests were performed using a model that combined the two mediator-based factors (i.e., procedural–justice mediator and disputant—mediator interpersonal justice) into one. 14 Logistic regression results for the model with five factors were consistent with those of the six-factor model: The distributive and procedural justice factors increased the odds of satisfaction with mediation in all three tests, and disputant–disputant interpersonal justice increased the odds of mediation satisfaction for respondents. The tests also revealed statistically significant results for the single mediator-based justice factor but they were contrary to expectations: Perceptions of mediator-based justice reduced the odds of reported satisfaction with the mediation process. The effects were statistically significant in the combined respondent-grievant (odds ratio = .40, p < .01) and respondents-only (odds ratio = .16, p < .05) models and trended toward significance in the grievants-only model (odds ratio = .55, p < .10).
Discussion
The test results of the six-factor model for DOC’s mediation program corroborate findings from prior research (Bingham, 1997; Bollen & Euwema, 2013; Bollen et al., 2012). This includes the high percentages of grievants and respondents reporting that they experienced factors of organizational justice in mediation and were satisfied with the mediation process. These perceptions of justice and satisfaction were consistently higher for respondents than grievants, a finding that is also well documented in the literature. The positive effects of distributive, procedural, and interpersonal justice on mediation satisfaction were also generally consistent with previous work. Given the severity of issues being mediated (i.e., written warnings, demotions, and dismissals) at DOC, these are important findings that lend credence to the value of mediation in promoting organizational justice. The insignificant results in the logistic regressions for procedural justice–mediator and disputant–mediator interpersonal justice on mediation satisfaction are consistent with previous findings (Nesbit et al., 2012) yet at odds with theoretical expectations. Finally, the high levels of satisfaction with shared neutrals and facilitative mediation support previous thinking about the importance of aligning DSD with organizational context. The main finding that is at odds with prior research is the insignificant effect of informational justice on mediation satisfaction across all three tests of the six-factor model. This section discusses the implications of these findings around the three aspects of DSD presented in the literature review: the style of mediation, type of third-party neutral, and type of issues mediated.
Facilitative Mediation Style
Given the study’s aim of assessing the six-factor model of organizational justice, it is interesting that, in all three tests, neither procedural justice–mediator or disputant–mediator interpersonal justice had a significant effect on the odds of reported mediation satisfaction. This finding is consistent with Nesbit et al.’s (2012) study of transformative mediation. It is possible that the suppressed effects of the mediator in transformative mediation are also true for the facilitative style of mediation at DOC. 15 Although facilitative and evaluative mediation are both settlement-based approaches, mediators using an evaluative style may take a more active role in the process by weighing the strengths and weaknesses of the parties’ arguments, making suggestions, and more directly steering parties toward a solution (Jameson, Berry-James, Daley, & Coggburn, 2017; Raines, 2000), thus making the mediator-based factors of organizational justice more salient than in facilitative mediation. In this way, facilitative mediation may have more in common with transformative mediation than its settlement-based counterpart (i.e., evaluative mediation) when it comes to the effects of the mediator-based factors of organizational justice on mediation outcomes.
The significant negative relationship between the combined mediator-based factor in the secondary analysis (i.e., with five factors) points to another possibility. It is plausible that these results relate to the role and behavior disputants’ expect mediators to play versus what is experienced. For example, if disputants erroneously expect the mediator to adjudicate their dispute but instead experience mediator behavior that is consistent with facilitative mediation, then it makes sense that disputants would report experiencing mediator-based justice yet, in wanting the mediator to decide in their favor, be less likely to report satisfaction with the mediation process. Haynes, Haynes, and Fong (2004) suggested just such a possibility, noting that . . . no matter how much information the mediator gives to the clients before they come to mediation and repeats in the opening few minutes, the clients do not see a mediator sitting in front of them, but a judge. They each want the mediator to judge them well and to judge the other badly. (p. 5)
A related point stems from Nesbit et al.’s (2012) conclusions about the role of the mediator in regard to interpersonal justice within the context of transformative mediation. The authors note that a key role of the mediator is facilitating two-way communication between the disputants, something that would elevate perceived effects of disputant–disputant interpersonal justice over mediator–disputant interpersonal justice in mediation. Given research that shows respondents are especially sensitive to interpersonal justice (Nesbit et al., 2004; Potipiroon & Rubin, 2018), the current findings that disputant–disputant interpersonal justice has significant effects on the odds of respondents’ (but not grievants’) reported satisfaction with mediation make sense. DOC’s facilitative mediation process appears to have succeeded in placing more control over resolving disputes in the hands of the disputants and, in so doing, heightened respondents’ perceptions of disputant–disputant interpersonal justice.
For mediators, the results suggest that, even when doing everything “right” in terms of dealing with disputants (i.e., providing organizational justice, facilitating two-way communication) in a facilitative mediation, disputants’ satisfaction with mediation is not guaranteed. These possibilities suggest the need for additional work assessing organizational justice in the context of evaluative mediation or, ideally, contrasting transformative, facilitative, and evaluative styles.
A final point regarding facilitative mediation relates to the findings for informational justice. As reported, this factor did not have statistically significant effects on the odds of reported mediation satisfaction in any tests of the six-factor model at DOC: the effect was slightly positive for grievants but negative for respondents and in the combined test. This differs from previous REDRESS-based research on transformative mediation where informational justice was consistently important for employees, though not supervisors (Nesbit et al., 2012). This may have to do with the straightforward goal—settlement—of DOC’s facilitative mediation relative to the broader goals—empowerment and recognition—of the REDRESS transformative mediation program. Even though the effect is not statistically significant, it points to a potentially important substantive difference between grievants and respondents on this aspect of organizational justice: informational justice may be more important to grievants, who are initiating mediation, than respondents who likely possess greater knowledge of and experience in mediation (Nesbit et al., 2012). It is possible, too, that the negative effect for respondents reflects their feelings of being second guessed or saddled with additional informational and procedural burdens to their decision-making authority, something that might be likely given the already stressful disciplinary issues being addressed (Daley, 2008).
Shared Neutrals
This study’s findings also shed light on the potential value of shared neutrals as mediators in DSD. Previous research has found important differences in mediation outcomes related to a mediator being from inside or outside of the organization, with more positive assessments being tied to outside neutrals (Bingham et al., 2000; see Bingham & Pitts, 2002). That there was such high—indeed, almost uniform—agreement about the fairness and impartiality of mediators and comediators and the respect they gave to disputants speaks to the attractiveness of shared neutrals as mediators (Raines, 2000). Indeed, from a DSD perspective, shared neutrals drawn from other government agencies represent an attractive, less costly option for governments that produces relatively high satisfaction and justice perceptions. Raines (2000) noted that shared neutrals are perceived to be more objective (i.e., similar to external/outside neutrals) than inside neutrals who can create perceptions of structural bias in the dispute system (Bingham et al., 2000). The current findings seem to support such perceived objectivity for shared neutrals. What is more, as government employees, shared neutrals have the advantage of understanding the governmental context. Although more research is needed, the current findings point to the attractiveness—from cost, perceived objectivity, contextual knowledge, and performance perspectives—of shared neutrals for public sector workplace mediation.
Issues of Employee Discipline in Workplace Mediation
The positive effects of procedural justice-process and disputant–disputant interpersonal justice on the odds of mediation satisfaction suggest that the most important value of mediation may be the process itself: The creation of a space in which grievant and respondent can interact in a respectful manner. This may be especially relevant given the types of disputes at play in DOC’s mediation process, which were highly face-threatening disciplinary actions. This is an important point, as in many employment grievances the agency-controlled process affords the grievant little or no power, making them feel unsafe discussing their fears and concerns with their supervisor (respondent). In fact, it is not uncommon for grievants to want mediators to take an intermediary role so that they do not need to face their perceived adversary. In other grievance procedures, face-to-face interaction is either avoided (e.g., when a case is read and decided by a hearing officer) or mediated by attorneys (as in an arbitration or legal hearing). For respondents, the results show that a fair process and respectful treatment send strong value messages, thus affirming their status and underscoring their own sensitivity to experiences of organizational justice. 16
Differences in perceptions of justice among grievants and respondents seem relevant to this issue of dispute type. This study supports previous findings about differences in justice perceptions among grievants (employees) and respondents (supervisors or management representatives). The three significant variables in our study included distributive justice and procedural justice-process variables, whereas disputant–disputant interpersonal justice was significant only in the respondents-only test. It is no surprise that distributive justice, or satisfaction with the outcome, more than doubles the odds of grievants’ satisfaction with mediation. In the cases mediated within DOC, grievants are seeking to restore terminated positions (or, perhaps, be allowed to resign in lieu of termination) or overturn written warnings. These are cases for which there is rarely a middle ground or compromise because they are fairly clear situations where either a rule was broken or it was not. Given that respondents typically cannot overturn the previous decision, the best outcome a grievant can hope for is to leave with their reputation intact and perhaps get a recommendation for a position at another agency. Given this type of conflict, it is not surprising that the overall perceptions of distributive justice were lower for grievants than respondents.
The significantly higher relationship between procedural justice-process and disputant–disputant interpersonal justice and mediation satisfaction for respondents is also not surprising. This result is consistent with Rubin and Chiques (2015) who found procedural justice perceptions were significantly higher for supervisors than employees and strongly related to satisfaction. Our findings also corroborate Potipiroon and Rubin’s (2018) conclusion that higher status employees (supervisors) are more sensitive to perceptions of procedural and interpersonal justice and Nesbit et al.’s (2004) finding that supervisors are more apt than employees to perceive interpersonal justice. Given the differences in the types of disputes mediated in this study as compared with the EEOC-related disputes in REDRESS, the results suggest these relationships hold regardless of the dispute context.
Conclusion
Organizational justice in the context of workplace mediation is complex and multifaceted, encompassing outcomes, procedural factors by which they are determined, and interpersonal relationships. As called for by Nabatchi et al. (2007), the purpose of this study was to assess the six-factor model of organizational justice in a different (non-REDRESS) workplace mediation context. Specifically, it assessed organizational justice within the context of workplace mediation in a state agency, utilizing shared neutrals and facilitative mediation, and addressing a set of the most contentious issues in an effort to answer the key research question:
The results of the analyses point to an answer that both corroborates and, in some instances, contradicts previous studies of organizational justice in mediation and, in so doing, suggests the need for future research.
Given the importance of organizational justice to public administration, there is a need for additional research, including that which assesses justice in the context of workplace mediation (Nabatchi et al., 2007). As Bollen and Euwema (2013) noted, there is a need for research that identifies the antecedents to mediation’s adoption, explains the conditions under which it (in its various forms) succeeds, and assesses its short- and long-term outcomes. Most organizational justice and workplace mediation studies use existing data from postmediation surveys, which may result in imprecise measurement of important constructs. Hence there is an opportunity to assess current measures and develop new ones. A logical starting point would be attempts to validate Nabatchi et al.’s (2007) organizational justice measures in different workplace mediation contexts. Such work holds promise for overcoming the methodological limitations of using existing postmediation data and producing better understanding of the six-factor model’s theoretical usefulness through more precise tests. Similarly, mixed-method approaches (e.g., surveys and structured interviews of disputants and mediators, content analysis of mediation agreements) also hold promise for developing understanding of justice in mediation and, potentially, identifying additional factors that affect success. Current mediation research also tends to focus on short-term outcomes such as settlement and satisfaction, so there is a need for work looking at longer-term outcomes (Herrman, 2006), including postsettlement outcomes such as improved relationships, performance, and status of disputants (e.g., evaluations, salary, job status), and how the development of “anticipatory justice” (Goldman et al., 2016) might help attract and retain public servants.
In the broader context, this study suggests that mediation offers a sensible middle ground between the procedural labyrinth of some traditional grievance and discipline systems and draconian efforts to eliminate essential safeguards that have long provided a sense of justice to public servants. Although the DOC grievants were not always satisfied with the outcome (i.e., distributive justice) of mediation, majorities still reported experiencing other factors of organizational justice and were satisfied with the overall mediation process. This suggests that the process does indeed matter. Providing employees with procedural justice may help with the acceptance of less-than-favorable outcomes and has a clear positive effect on perceptions of mediation’s worth. Perceiving a process as fair and being satisfied with it represent welcome developments in an era of eroding protection and status for public servants.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.