1 Introduction

Procedural justice research has been a crucial part of the psychological study of dispute resolution over the past three decades. In an effort to understand when and how people are satisfied with the results of dispute resolution processes, researchers have focused on learning how an individual’s perception of the fairness of a decision-making process affects perception of the decision itself. Procedural justice research has suggested that people care deeply about the fairness of procedures used to resolve their disputes and that the fairness of these procedures has significant effects on people’s satisfaction with and adherence to their outcomes (Lind and Tyler 1988).

This research has largely taken place in the context of dispute resolution settings, such as litigation (Tyler 1984), criminal trials (MacCoun and Tyler 1988) mediation (Pruitt et al. 1993), or encounters with police (Tyler and Folger 1980), that involve third-party authorities, but recent work has suggested that fair procedures play a significant role in evaluation of dispute resolution in the negotiation context, even in the absence of a third-party authority (Hollander-Blumoff and Tyler 2008; Kass 2008; Wagner and Druckman 2012). Hollander-Blumoff and Tyler (2008) suggested that procedural justice concerns influence the way that individuals assess their satisfaction with a negotiated agreement, finding that higher levels of procedural justice were linked with more enthusiasm about a negotiated agreement and higher assessments about the likelihood of adherence to the agreement. Higher levels of reported procedural justice also led to participants reporting a more positive affect during the negotiation and a higher rating for the process of the negotiation being collaborative. In a largely zero-sum negotiation, procedural justice had no effect on the bottom-line outcome of the parties; in negotiation with greater integrative potential, procedural justice was related to a greater level of information exchange, a larger joint outcome, and a more even distribution of outcomes between the two members of the dyad.

If, indeed, individuals care about the fairness of a negotiated process, and the fairness of a negotiated process has important effects on the degree to which a negotiation is integrative or distributive, it is critical to know what fairness in the negotiation setting looks like. Past literature has identified factors that individuals use to determine whether or not they have been treated fairly in a dispute resolution process. Those factors, including trust, neutrality, opportunity for voice, and courtesy and respect, have primarily been previously studied in the context of disputes whose resolution involves a third-party authority or facilitator. That is, individuals look at the behavior of the third-party facilitator to determine how trustworthy, neutral, courteous, and respectful the facilitator is, and how much the facilitator allows the individual to present her side of the dispute.

Although past procedural justice research has explored these antecedents of procedural justice judgments in third-party settings, the present research considers in depth a largely unexplored area in the procedural justice and negotiation arenas: the antecedents of perceptions of fairness of process in the negotiation setting. The research presented here considers what factors contribute to an understanding that a dyadic negotiation process is fair, including in the analysis the assessments of both parties to the negotiation as well as those of a third-party external coder. Additionally, the research explores the implications of these assessments with respect to the psychology of interpersonal perceptions.

2 The Antecedents of Procedural Justice

The question of how people make assessments about whether a process is fair is not a new one; procedural justice research began over three decades ago with a focus on what procedures and practices individuals would find most fair in a dispute resolution setting. Over time, however, the question of what leads individuals to conclude that a process is fair has become somewhat settled, and procedural justice research has widened to investigate the wide range of settings in which procedural justice and fairness matter to individuals. Recent research has focused less on how people conclude that a process is fair and more on potential theoretical frameworks that may explain the “procedural justice effect”—that is, why so many people care deeply about the fairness of procedures that affect them, separate and distinct from the fairness or the favorability of the outcome that they receive (Tyler and Lind 1992; Van den Bos and Lind 2002).

The large literature on how individuals make assessments about fairness of the process has helped to clarify our understanding of what matters to people when they decide whether a process is fair. But the antecedents of procedural justice are subjectively rated in most research. This research relies almost exclusively on self-report: individuals are asked to rate the fairness of a process they experienced, and/or to rate various factors of a process that they experienced (Lind et al. 1997, 1990). So, for example, while the experience of having a voice and being heard may correlate strongly with an assessment of procedural fairness, both of these assessments are subjective: that is, when an individual feels that she had a voice and was heard, then she is also likely to feel that she was treated with procedural fairness. This research leaves us with an understanding of how people make assessments of procedural fairness based on those individuals’ own assessments of other factors present in the process, but little sense of the relationship between the features of a decision-making process, subjective assessments of the antecedents of procedural justice, and subjective assessments of procedural justice more generally.

A large literature has explored the potential bases of individuals’ judgments about fairness of process. Thibaut and Walker (1975) were the first to develop a framework for the importance of procedural justice judgments, and their model for procedural justice evaluations relied on two important aspects that individuals might use in assessing fairness of decision-making procedures. First, individuals might exert some degree of control over the process, and second, they might exert some degree of control over the decision itself. In many settings in which procedural justice has been explored, decision control is fixed because a third-party authority is making binding decisions. In other settings, such as mediation, there is some degree of decision control. The degree of process control also varies across settings. The Thibaut and Walker model has had strong influence on subsequent procedural justice theory and research. In particular, many studies have found that the level of input or participation that procedures permit in a particular decision-making process has important consequences for individuals’ fairness judgments about the decisions and for the individuals’ acceptance of the decisions.

In another important work on procedural justice, Leventhal (1980) identified a different set of criteria that are used in making procedural justice assessments. These are consistency, ability to suppress bias, decision quality or accuracy, correctability, representation, and ethicality. As Tyler (1990) notes, the only overlap that occurs within the Leventhal and Thibaut and Walker criteria is that of representation.

More recently, Tyler (2000) and Tyler and Blader (2004) identified four primary factors that individuals use in making decisions about the quality of procedural justice they have received: participation (voice), neutrality, courtesy and respect, and trust. First, it is important to allow parties opportunities to state their arguments, and to make clear that the decision maker is listening to those arguments. Second, people value having an unbiased and factual decision-making process in which rules are applied in a consistent manner. Third, they want to be treated with dignity and courtesy, and to have their rights acknowledged. Finally, people want to deal with people whose motives they trust. That is, they value people who act in good faith.

Although these factors appear relatively settled across procedural justice research, there may be important differences in the way in which procedural justice assessments are made in the third-party authority context versus in the dyadic negotiation context, especially because context and situation can affect procedural justice judgments (Tyler 1990). For example, Lind et al. (1997) hypothesized that relational considerations of neutrality, trust, and status recognition would play an important role in forming procedural justice judgments in the dyadic dispute resolution setting just as they do in third-party authority settings. However, Lind et al. suggested that because dyadic disputing procedures have no assurance of neutrality, parties to such a process would be more concerned about neutrality in the dyadic setting. Lind and his colleagues also hypothesized that status recognition might be a more important variable in procedural justice assessments in the dyadic context than in the third-party authority context, and that trust might be less important in assessing procedural justice in the dyadic setting than in the third-party setting. Their results were generally in line with these hypotheses. The researchers also found that the degree to which individuals felt that they had a voice in the dispute resolution process had an effect on procedural justice in the dyadic setting, although this effect was largely mediated by the relational variables they studied.

In the Lind et al. study, as well as in typical studies of procedural justice, participants are asked to provide a self-report about their experience of a dispute, including how they felt about the fairness of the process used to resolve the dispute and about potential antecedent factors leading to their fairness judgments. However, the researcher is not able to independently assess either (1) the fairness of the dispute resolution process or (2) the degree to which the potential antecedent factors are actually present in the dispute. So, for example, an individual assessing the fairness of a mediation or court case would rate the degree of procedural justice afforded, as well as rating the degree to which she had a voice, was treated with courtesy and respect, felt that she could trust the decision-maker, and believed that the decision-making process was neutral. However, researchers would not independently assess the degree to which the individual did have a voice, was treated courteously and respectfully, dealt with a trustworthy decision-maker, and experienced a neutral decision-making process. Neither would researchers assess what observable factors were present that led to the individual’s judgments with respect to those issues.

Because procedural justice is an influential factor in people’s satisfaction with a decision and because procedural justice is an important factor in individuals’ assessments of whether institutions are legitimate, it can be critical to the success of dispute resolution mechanisms to ensure that individuals experience them as procedurally fair. In this effort, however, subjective assessments may not provide enough guidance. It would be useful to know what, if any, factors observable by a disinterested third party are likely to lead reliably to those assessments. Especially in a negotiation setting, where few rules dictate the manner of individuals’ behavior, an understanding of what factors may lead to a sense of procedural justice is critical to any meaningful discussion of the role of procedural justice in negotiation.

3 Interpersonal Perception and Procedural Justice

As noted above, prior literature on procedural justice has relied on individuals’ own reports of the degree of fairness, and underlying antecedent features, that they have experienced. Yet individuals see the world, and interpret others’ behavior, very differently (Srivastava et al. 2010; Wood et al. 2010). Individual differences may cause differential interpretation of behavior. Therefore, it may be that even if any, some, or all of the antecedent factors of voice, neutrality, courtesy/respect, and trust, help to foster individuals’ perception of fair treatment in negotiation, not all negotiators would agree on whether or not the negotiation provided those factors.

In the context of third-party decision makers, rules may help to ameliorate this concern. For example, legal rules clearly regulate who has the opportunity to be heard by a judge, and when. Deviation from these rules would suggest deprivation of voice, and adherence to these rules would suggest that voice was provided. Similarly, rules about neutrality, lack of bias, and respectful behavior may help to guide individuals’ assessments about whether these factors were present during litigation, arbitration, and mediation.

In contrast, negotiation is largely rule-free. Parties in a typical negotiation are not bound by any rules for behavior. Even legal dispute resolution negotiation is governed by only a handful of rules, which are generally interpreted loosely. The only rules about negotiation process are those that deal with misrepresentation or fraud and threats and duress during negotiation (ABA Model Rules of Professional Conduct, Model Rule 4.1). Legal rules prohibit lawyers from making false statements of material fact or law, and further prohibit lawyers from failing to disclose material facts to another person when the disclosure is necessary to avoid assisting a criminal or fraudulent act by a client. However, the rules also make clear that “[u]nder generally accepted conventions in negotiation, certain types of statements ordinarily are not taken as statements of material fact.” (Comment to Model Rule 4.1) In particular, estimates about price or value and statements about a party’s reservation price in negotiation are exempt from classification as material facts. The intense debate in the legal community over what the scope of the rule is makes clear that the rules provide slim guidance for assessments of procedural justice.

The literature on interpersonal perception has suggested that parties use a variety of factors to form judgments about the behavior of others (Funder 2001; Mehl et al. 2006). Sometimes these judgments are accurate and sometimes they are not. There is a vast body of research on the correspondence bias, in which individuals ignore the situational causes of behavior in favor of ascribing behavior to dispositional qualities (Ross 1977). In all cases, however, the observer must first form some judgment about the behavior itself. The social relations model (Kenny and Albright 1987) focuses on ways to assess the accuracy and the general nature of interpersonal perception, typically in the context of group or dyadic interactions. Questions of particular interest in interpersonal perception include whether one party’s view of another is correct (target accuracy); whether one party sees the other party as she sees herself (assumed similarity); and whether a partner sees a party as the party sees herself (self-other agreement).

In the procedural justice context, this perspective leads to several important questions. First, because most findings about the significance of procedural justice are based on subjective perceptions, what behaviors by a negotiation partner will actually produce these subjective perceptions? Second, how much behavior that is observable by a disinterested third party predicts assessments of procedural fairness by the parties? Third, to what extent are dyadic partners forming judgments based on accurate information, and to what extent might they form judgments based on assumed similarity? The present research uses observations by external third-party coders to explore questions about how individuals form their judgments about procedural justice and about behavior that may relate to assessments about fairness.

4 Method

4.1 Participants

The participants in this study were first-year law students in a required law school class where they participated in a negotiation simulation exercise between two attorneys. In this exercise, each student was randomly assigned a role as a lawyer for either a homeowner or a contractor in a dispute over a building contract. A brief description of the problem is included as “Appendix.”

A total of 414 first-year law students participated in the negotiation exercise. All consenting individuals were included in individual-level analyses; for dyadic analysis, only those students in a dyad of two consenting members were included. For individual analysis, this led to 330 participants; for dyadic analysis, this led to 296 included individual participants, who comprised 148 dyads.

4.2 Procedure

Each side of the case was described in a set of material provided to students representing that side. Participants also met with a confederate trained to act as the “client,” who presented a story of his or her side of the case based upon a standardized set of material.

The participants then engaged in a bilateral negotiation with the other party designed to mirror real-world pretrial negotiations. Participants were told to do their best to represent their client by seeking a desirable outcome. As in real world negotiations, the parties were not obligated to come to an agreement and could reject any agreement and indicate a preference for moving to arbitration, which was provided as the dispute resolution mechanism in the underlying contract between the parties. Participants were given a common estimate of arbitration costs and were required to perform independent legal research about the likelihood of potential outcomes of arbitration under relevant case law. Participants were given no outside incentive for their financial or other achievements during the negotiation, but negotiation performance was analyzed and discussed by both classmates and a professor, providing independent motivation to negotiate seriously and well.

4.3 Video Coding

During the negotiation exercise, each negotiation was videotaped in webstream format for later classroom use. Students were aware of the videotaping process. Of the 148 dyads that jointly gave consent to be included in research, 107 randomly selected videos were coded by between two and eight independent coders who were randomly assigned, sometimes in pairs, to subsets of videos. In this two-stage coding process, each coder watched each coded video at least twice, initially watching the video once through in its entirety without coding behaviors but rating the video for its global fairness, both overall and with respect to each individual negotiator. Subsequently, each coder watched the video, sometimes more than once, and coded for specific behaviors, without respect to any global fairness measure.

4.3.1 Video Coding Protocol

There are currently several schemes that have been developed for content analysis of negotiation behavior. However, coding schemes that have previously been used in negotiation research do not adequately measure relevant fairness-related variables. For this reason, I developed a unique coding scheme with categories that are theoretically consistent with a model of procedural justice. These categories include behaviors related to voice, trust, courtesy/respect, and neutrality. The coding was extensively pilot-tested in order to measure appropriate variables and improve inter-rater reliability.

The coding analysis was unit-based: each individual utterance that is surrounded by speaking from the other party, regardless of the utterance’s length, comprised one unit. Coding categories were not mutually exclusive and units could also receive no code if no code was relevant. Codes were further categorized by the individual actor, such that each individual’s behavior was coded separately. The coding manual is available from the author.

4.3.2 Video Coding Reliability

Reliability for video-coded variables was measured in two distinct ways, first with the average measures reliability statistic produced by a one-way random effects intraclass correlation for two ratings of each video. For these two ratings, I used all data for every video that was coded by only two coders, and selected for all other videos two sets of coding by the two most experienced coders assigned to that video. In appropriate cases, data was ipsatized within coder (that is, transformed into ratings relative to the individual coder’s average rating) before aggregation. Second, reliability is analyzed using Krippendorff’s alpha (K alpha), a specialized and conservative measurement tool for reliability of content coding when multiple coders are present and any given data have been coded by some subset of coders, for the data in its raw numeric form. Both reliabilities are in parentheses after each variable is presented.

All analyses were run using mean video ratings for all coders; because reliability typically increases with additional coders, the reliability presented is a figure that may slightly underestimate the reliability of the mean rating of all coders. It is worth noting here that coding actual behavioral is difficult (Baumeister et al. 2007) and results based on coding of behavior may be published even when inter-rater reliability falls well below textbook standards for other types of reliability (Vazire 2010). Additionally, in exploratory studies such as those presented here, stringent standards for reliability “may be relaxed considerably” (Krippendorff 1980, p. 147).

4.4 Independent Variables

In prior procedural justice research, analyses of the antecedents of procedural justice assessments are typically based on self-report measures. Here, this research first uses this technique to explore what factors played a role in individuals’ assessments of fair process in the dyadic negotiation setting, considering how the self-report measures of input/voice, neutrality, respect/courtesy, and trust relate to the self-report measures of global fairness in the negotiation process. However, this research goes beyond the traditional paradigm by looking at the substance and action of the dispute resolution process itself through coding the negotiation videotapes for specific behavior that relates to each of the antecedents of procedural justice. Additionally, external third-party coders assess the degree to which individuals engage in fairness-related behavior, as well as assessing the general fairness of the negotiation process. Therefore, this analysis includes self-rating of behavior, other-rating of behavior, self-rating of procedural justice, external third-party coder rating of behavior for both members of the dyad, and third-party coder rating of procedural justice (Fig. 1).

Fig. 1
figure 1

Graphic representation of the types of variables collected

Full size image

4.4.1 Self-Report Input/Voice Behavior by Other (\(\alpha = .77\))

The items were: “My negotiation partner listened when I expressed my views” and “My negotiation partner gave me an opportunity to express my views.”

4.4.2 Self-Report Trust Behavior by Other (\(\alpha = .48\))

The items were: “How trustworthy was the other attorney?”, “My partner shared important information about his/her case with me,” and “My negotiation partner used deceptive tactics in dealing with me” (reverse-coded).

4.4.3 Self-Report Courtesy/Respect by Other (\(\alpha = .69\))

The items were: “My negotiation partner treated me with courtesy,” “My negotiation partner respected my client’s rights,” and “My negotiation partner cared about my client’s satisfaction.”

4.4.4 Self-Report Neutrality

The item was: “The outcome we reached was based on facts rather than personal biases.”

4.4.5 Self-Report Voice Behavior by Self (\(\alpha = .86\))

The items were: “I listened when my negotiation partner expressed his or her views” and “I gave my negotiation partner an opportunity to express his or her views.”

4.4.6 Self-Report Trust Behavior by Self (\(\alpha = .28\))

The items were: “I shared important information about my case with my negotiation partner,” and “I used deceptive tactics in dealing with my negotiation partner” (reverse-coded).

4.4.7 Self-Report Courtesy/Respect Behavior by Self (\(\alpha = .55\))

The items were: “I treated my negotiation partner with courtesy,” “I respected the other party’s rights,” and “I cared about my negotiation partner’s satisfaction.”

4.4.8 Video Coded Input/Voice (ICC (1, k) \(=\) .71, ipsatized; K alpha  \(=\) .13)

This variable is represented by a frequency count of each participant’s interrupting behavior, reverse-coded.

4.4.9 Video-Coded Courtesy/Respect (ICC (1, k) \(=\) .74; K alpha \(=\) .37)

This variable is derived from a frequency count of each participant’s behavior that is identified by coders as acting in a discourteous and disrespectful manner towards one’s partner. Specifically, contentious behavior is counted and reverse-coded.

4.4.10 Video-Coded Trust

This variable is derived from a frequency count of each participant’s behavior that is identified by coders as either sharing important information about one’s own case with the other party or as using deception during the negotiation process (ICC (1, k) \(=\).64; K alpha \(=\) .23). The important information has been specifically identified (e.g., one party was in a car accident (ICC (1, k) \(=\) .88; K alpha \(=\).83)). Other factors that are explored in this category include absolving the other party from blame (ICC (1, k) \(=\) .79; K alpha \(=\) .56); absolving oneself from blame (ICC (1, k) \(=\) .82; K alpha \(=\) .59); and expression of intent to reach agreement (ICC (1, k) \(=\) .43; K alpha \(=\) .25).

4.4.11 Video-Coded Neutrality

This variable is derived from a frequency count of each participant’s behavior that is identified by coders as identifying an objective standard by which to make decisions during the negotiation, such as fair market value of a particular piece of equipment or hours of labor (ICC (1, k) \(=\) .43; K alpha \(=\) .33), explicit contract language (ICC (1, k) \(=\) .61, ipsatized; K alpha \(=\) .16), governing industry standards (ICC (1, k)  \(=\) .50; K alpha \(=\) .29), bids made by a competitor (ICC (1, k) \(=\) .58; K alpha \(=\) .33), the legal standard governing the dispute (ICC (1, k)  \(=\) .85; K alpha \(=\) .68), or the likely outcome of the dispute in an arbitration setting (ICC (1, k) \(=\) .82; K alpha \(=\) .70).

4.5 Dependent Variable

Self-Report Global Procedural Justice (\(\alpha = .73\)). The items were: “With respect to fairness, how would you rate the conduct of your negotiation,” and “How would you rate the overall fairness of the negotiation process?”

4.6 Analysis

This study explores the question of what factors individuals use to guide their assessments of procedural justice in bilateral negotiation through a two part analysis. First, the study examines what behavioral factors relate strongly to individuals’ global fairness assessments through an analysis of the self-report surveys; that is, it will analyze the correlation between participants’ assessments of the degree of voice, trust, courtesy/respect, and neutrality experienced during the negotiation and the degree of procedural justice that they report in the negotiation. In this part of the study, I consider participants’ assessments of the voice, trust, and courtesy/respect present in the negotiation in two different ways: first, by looking at the level of voice, trust, and courtesy/respect that they report having provided to the other party, and second, by looking at the level of voice, trust, and courtesy/respect that they report that the other party provided to them. Neutrality will also be analyzed, but it is not measured in this bilateral way.

In the second part of the study, I consider what third-party coders observe with respect to potential behavioral antecedents of procedural justice. Specifically, I examine the correlation between an individual’s assessment of procedural justice experienced during the negotiation and observed behavior in the negotiation connected to voice, courtesy/respect, trust, and neutrality. Again, I consider these factors from two perspectives, looking first at the relationship between a third party’s coding of the behavior of an individual actor in the negotiation and that individual’s own perceptions of procedural justice, and second, at the relationship between a third party’s coding of the behavior of an individual actor in the negotiation and that individual actor’s partner’s perceptions of procedural justice.

Beyond the question of whether specific behavior leads to assessments of procedural justice by negotiating parties, the gathering of data about fairness-related behavior provides the opportunity to explore several additional questions of interest. In particular, do individuals see their own fairness behavior as others see it? How accurate are we at knowing when we are providing others with procedural justice, or at knowing when we are the recipients of what others believe to be fair treatment? This study explores this question through several distinct analyses. First, the study examines the relationship between an individual’s self-ratings for fairness-related behavior and the individual’s negotiation partner’s ratings of that individual’s fairness-related behavior. Second, the study explores the relationship between an individual’s self-ratings for fairness-related behavior and the third-party coder’s ratings of the individual’s fairness-related behavior. Third, the study compares the ratings of an individual’s fairness-related behavior by the third-party coder and the negotiation partner.

5 Results

5.1 Self-Coding Results

The first part of this study examined the relationship between perceptions of general procedural justice by a negotiating party and that party’s perceptions of procedural justice antecedent behavior by herself and by the other negotiator.

Perceptions about one’s own behavior with respect to giving the other party voice correlated with procedural justice assessments (r  \(=\) .3, p < .01), and perceptions about providing the other party with courteous and respectful behavior correlated with procedural justice assessments (r \(=\) .49, p < .01). Behaving in a trustworthy manner towards the other party did correlate significantly but not strongly with procedural justice assessments (r \(=\) .13, p < .05) (Table 1).

Table 1 Factors shaping negotiator procedural justice judgments
Full size table

Assessments about the other party’s behavior with respect to potential procedural justice antecedents of voice, courtesy/respect, and trust all were significantly correlated with assessments of procedural justice. The other party allowing for voice correlated significantly (\(r = .35\), p < .01); the other party treating one with courtesy and respect also correlated significantly (r \(=\) .46, p < .01). Trustworthy behavior by the other party also bore a significant connection with procedural justice assessments (r \(=\) .35, p < .01) (Table 1).

Additionally, a perception that the negotiation was conducted in a neutral manner had a significant effect on procedural justice perceptions (r \(=\) .3, p < .01). As noted above, this factor was not assessed for both parties but rather for the negotiation as a whole (Table 1).

In regression analysis, regressing the dependent variable of general procedural justice assessments on the independent variables of voice, courtesy and respect, and trust ratings for self-behavior, and neutrality, courtesy and respect ratings played a significant role (\(\beta = .42\), p < .001), along with neutrality (\(\beta = .112\), p < .05), accounting for .25 of the variance. In a similar regression analysis with voice, courtesy and respect, and trust ratings for the other party’s behavior, along with neutrality, voice (\(\beta = .141\), p < .01), courtesy and respect (\(\beta = .293\), p < .001) and trust (\(\beta = .18\), p < .001) played significant roles, accounting for .27 of the variance (Table 2).

Table 2 Factors shaping negotiator procedural justice judgments—regressions
Full size table

In a regression analysis that regressed procedural justice assessments on independent variables including an individual’s voice, courtesy and respect, and trust ratings for both herself and the other party, as well as neutrality, the following factors made a significant contribution: the voice afforded to one by the other party (\(\beta = .123\), p < .05); one’s own courtesy and respect behavior (\(\beta = .27\), p < .001); and the trust behavior of the other party (\(\beta = .17\), p < .01). All told, the factors included in this regression accounted for .30 of the variance (Table 2).

Finally, in a regression looking only at ratings of courtesy and respect behavior for both self and other and their effect on procedural justice judgments, both are significant, with slightly greater weight given to one’s own behavior (\(\beta \,=\,.3,\) p < .01) than one’s assessment of the other party’s behavior (\(\beta = .23\), p \(=\) .01), with .23 of the variance accounted for.

5.2 Video Coding Results

The latter half of the study looked at the connection between individual negotiators’ judgments about procedural justice and video-coded behaviors relating to voice, courtesy and respect, trust, and neutrality. Considering, first, an individual’s own behavior and her own assessments of the negotiation’s procedural justice (Table 3), this analysis finds a significant correlation between an individual’s reported level of procedural justice and a third-party coder’s rating of voice (r \(=\) .24, p < .01) and courtesy/respect behavior (r \(=\) .35, p < .01). For antecedent behavior that was targeted to relate to trust, all variables related to information sharing had no significant relationship to assessments of procedural justice, as did absolving the other party from blame and intent to reach agreement. However, there was a significant negative relationship between an individual’s reported level of procedural justice and a third-party coder’s rating of absolving oneself from blame (r \(=\) −.14, p < .05), and video-coded use of deception closely approaches a significant relationship with ratings of procedural justice (r \(=\) −.13, p \(=\) .056). For variables that were targeted to offer a potential relationship with neutrality, there was no significance for factors related to the use of fair market value standards or competing bids, contract language, and industry standards. However, there was a significant negative relationship between assessments of the fairness of the negotiation process and discussion of likely arbitration outcomes (r \(=\) −.26, p<.01).

Table 3 Video-coded behavior and negotiator procedural justice judgments
Full size table

Examining the relationship between video coding of one party’s behavior and perceptions on the part of the other party to the negotiation (Table 4), there was a significant correlation between the other party’s procedural justice assessments and the third-party coding of an individual’s behavior with respect to voice (r \(=\) .24, p  < .01) and courtesy and respect (r \(=\) .18, p < .01). For variables targeted to identify trust behavior, video coding of sharing important information and using deception, as well as absolving the other party from blame and expressing intent to reach agreement, did not relate significantly to the other party’s assessments of procedural justice, while absolving oneself from blame (r \(=\) −.21, p < .01) was negatively significantly correlated with assessments about the fairness of the negotiation process. For variables that were targeted to potentially suggest neutrality, there was, as above, no significance for factors related to the use of fair market value standards or competing bids, contract language, and industry standards. However, there was a significant negative relationship between assessments of the fairness of the negotiation process and discussion of likely arbitration outcomes (r \(=\) −.20, p < .01).

Table 4 Video-coded behavior and other-negotiator procedural justice judgments
Full size table
Table 5 Agreement about negotiator procedural justice behavior between negotiation partners
Full size table

In terms of the relationship between one party’s perceptions of her own behavior and her negotiation partner’s perception of her behavior (Table 5), there is a low but significant relationship between one party’s assessment of how much voice and opportunity to be heard that she afforded her partner and that partner’s perception of how much voice and opportunity to be heard that she experienced (r \(=\) .116, p < .05). However, there is a higher and significant correlation between self-report of how courteous and respectful one was (r \(=\) .34, p < .01) and how trustworthy one was in a negotiation and how the other party perceived that behavior (r \(=\) .24, p < .01). There was no agreement with respect to the degree of neutrality of the negotiation (p \(=\) .09 ns). However, global assessments of one’s own fairness behavior toward the other party did significantly correlate with global assessments of fairness behavior perceived by the other party (r \(=\) .25, p < .01).

Table 6 Agreement about procedural justice behavior between negotiator and observer
Full size table

Assessments of one’s own behavior and one’s assessments of the other party’s behavior demonstrate stronger correlations. All three of the antecedents that were measured in a “what I did/what you did” manner demonstrated these effects (voice, r \(=\) .2, p < .01; courtesy and respect, r \(=\) .69, p < .001; trust, r \(=\) .33, p < .01), as did general assessments about fair treatment (r \(=\) .58, p < .001).

An analysis of the relationship between video-coded behavior and individual judgments about procedural justice antecedents (Table 6) shows a significant relationship between courtesy/respect-related behavior and individuals’ judgments about their own courtesy and respect (r \(=\) .27, \(p < .01\)). Similarly, there is also a significant relationship between voice-related behavior and individuals’ assessments about their voice behavior during the negotiation (r \(=\) .15, p < .05). Although most of the coding variables potentially related to trust and neutrality were not significantly related to individual judgments about trust and neutrality, there were several exceptions. For one party, revealing that the client had been injured in a car accident and that the client was not able to make a profit on the transaction had a significant positive effect on trust behavior ratings (r \(=\) .27, p < .01; r \(=\) .19, p < .05). Additionally, use of deception was negatively and significantly related to assessments of trust behavior (r \(=\) −.21, p < .01). Also, discussion of industry standard was negatively related to assessments of neutrality (r \(=\) −.14, p < .05) and discussion of possible arbitration outcomes approached significance in relationship to assessments about neutrality (r \(=\) −.13, \(p = .051\)).

Additionally, the study compared the ratings of an individual’s fairness-related behavior by the third-party coder and the negotiation partner (Table 7). In this analysis, video-coders’ ratings of courtesy/respect-related behavior significantly related to what the negotiation partner experienced with respect to voice (r \( =\) .18, p < .01) and courtesy and respect during the negotiation (r  \(=\) .25, p < .01), but trust factors coded in the video had no significant relationship with the experience rating by the partner. However, discussion of potential arbitration outcome had a negative significant relationship to a negotiation partner’s assessment of neutrality (r \(=\) −.15, p < .05).

Table 7 Agreement about procedural justice behavior between negotiation partner and observer
Full size table

6 Discussion

This section discusses, first, the findings with respect to how individuals form their judgments about procedural justice in negotiation, and then considers the results from the perspective of research on interpersonal relations. The section concludes with implications for the psychology of procedural justice and for legal negotiation as well as suggestions for future research.

6.1 Antecedents of Procedural Justice

What emerges from these results is that for negotiators themselves, all four of the antecedent variables for procedural justice appear to play a role in forming assessments of fairness in negotiation. It is interesting to note, however, that individuals do not assess these factors in the same manner for themselves and others. The effects of an individual’s own behavior with respect to courtesy and respect, for example, largely mimics the effects of the other party’s courtesy and respect behavior when considered in simple correlation with general procedural justice assessments, but in regression analysis, it is clear that one’s own courtesy and respect behavior, rather than that of the other party, is a more significant driver of judgments about fair process. Similarly, voice effects for self and other both have significant effects in correlation with assessments of procedural justice. However, voice effects carry lower significance in regression analysis that includes courtesy and respect behavior. In contrast, one’s perception of one’s own behavior with respect to trust factors is not important in one’s assessment of how fair the negotiation process is, but one’s perception of the other party’s trust behavior is significantly related to global fairness judgments.

Neutrality plays a small but significant role when considering it in connection with one’s own behavior, but when considering it in connection with other behavioral factors of the other party, the factor loses importance. These survey results highlight the importance of courtesy and respect behavior, and to a lesser degree, voice behavior, in forming procedural justice judgments in bilateral negotiation.

From the perspective of objectively observed behaviors during the negotiation, behavior relating to voice and to courtesy and respect factors were the most important, with one’s own courtesy and respect behavior being the most strongly related to one’s assessments of procedural justice. When an individual acts in a discourteous, hostile, or disrespectful manner, that individual is more likely to conclude that the negotiation was not conducted in a fair manner. These results suggest that assessments of fairness of process in negotiation are largely driven by interpersonal relational factors rather than more substantive behavior such as trust-building, through information-sharing or telling the truth, or neutrality, by the use of invoked external standards. Indeed, one potential external standard, the likely outcome at arbitration should the negotiation fail to produce a successful outcome, had a negative impact on parties’ assessments of procedural justice regardless of which party raised the issue. This study suggests that procedural justice in negotiation may be more tied to evaluations of interpersonal interactions than to evaluations about substantive merits.

6.2 Interpersonal Perception and Agreement About Procedural Justice

Overall, parties did not agree very much on whether any individual actor in the negotiation had provided voice to the other party and on whether the negotiation had been conducted in a neutral manner (low self-other agreement), but parties did agree to some degree about whether one of them had been courteous to the other and whether or not they were trustworthy (some self-other agreement). In contrast, assumed similarity was significant for all procedural justice variables measured in this dual manner, and in particular was quite high for courtesy/respect and global fairness judgments.

Results also suggest that individuals were by and large more accurate with respect to assessments of their own behavior than with respect to the behavior of their partners. With respect to agreement between a third-party coder and a speaker, there was some degree of agreement between, respectively, behavior related to voice and perceptions of behavior related to voice, and behavior related to courtesy and respect and perceptions of courtesy and respect; for a third-party coder and the negotiation partner, there was similarly some agreement, although less, for courtesy and respect and voice. Speakers’ perceptions and coders’ ratings of behavior agreed somewhat for, respectively, trust and whether information sharing and deception occurred, but partners’ perceptions and coders’ ratings did not share this connection. Finally, and somewhat counterintuitively, both speakers’ and, to a lesser degree, partners’ perceptions of neutrality in the negotiation suffered when coders identified more discussion related to potential standards by which to assess the negotiation results, such as industry standards or the likely outcome from arbitration.

The results presented here suggest a fairly self-focused negotiator. While individuals did not agree all that much about either party’s behavior with respect to voice or neutrality, negotiators’ ratings of their own and the other party’s behavior for voice and neutrality were significantly correlated. Individuals did agree somewhat with respect to a speaker’s neutrality and courtesy and respect behavior, but the correlations between individuals’ neutrality and courtesy and respect judgments of themselves and others were even higher. This suggests that individuals make their judgments about what others are doing in negotiation, at least in part, based on what they themselves are doing.

In addition, because perceptions by a speaker are more significantly correlated with what a third-party coder sees than they are with what a partner sees, this suggests that individual negotiators are fairly good at knowing what they themselves are doing, but relatively less good at accurately perceiving the behavior of others. These results are consistent with the findings that an individual’s judgments about her own courtesy and respect behavior were the most significant factor in guiding assessments about procedural justice. Taking a step back and considering these findings together, the picture of procedural justice that emerges is one that is largely guided by one’s perceptions of one’s own behavior, which, in turn, are more accurate than one’s perceptions of the other party’s behavior. One’s own behavior has a direct relationship with one’s perceptions of procedural justice experience. Even as the other party’s behavior is meaningful to those assessments, it appears to be largely reliant in turn on one’s perceptions of one’s own behavior in the first instance. The addition of third-party coding here helps to demonstrate that an individual may be somewhat accurate in assessing her own behavior but less accurate in assessing the behavior of the other party with respect to fairness.

Interestingly, courtesy and respect judgments appear to be the least affected by this discrepancy. Both parties’ perceptions of one speaker’s courtesy and respect behavior are significantly correlated, as are how one individual believed that both she and her counterpart acted, as are the third-party coded behavior and what rating is given by both the speaker and her counterpart. This poses a question about why courtesy and respect is the least affected by problems of interpersonal perception among all of the procedural justice antecedent factors, and why it, as well, appears to be the strongest driver of procedural justice judgments in these results. Are people particularly good at observing and accurately judging courtesy and respect behavior because it is so important in forming judgments about fairness in negotiation? Or are procedural justice assessments based on perceptions of courtesy and respect because it is one of the few procedural justice factors that individuals are competent to judge? These questions suggest that further research on the role of courtesy and respect in procedural justice assessments is needed.

6.3 Implications

6.3.1 The Psychology of Procedural Justice

Procedural justice theorists have not reached agreement on the motivations and mechanisms that drive people to care about procedural justice separate and apart from distributive justice (that is, the fairness of outcomes) and outcome favorability. Three theories have been developed to account for procedural justice effects. First, Thibaut and Walker (1975) took an instrumentalist view, arguing that individuals preferred fairer processes because they were likely to produce fairer outcomes. Subsequently, Tyler and Lind (1992) developed the group value model, suggesting that fairer processes were valued in and of themselves, unrelated to their effects on outcome, because they conveyed important messages to individuals about their status in society that in turn affected individuals’ self-esteem. Finally, Van den Bos et al. (1997) have suggested “fairness heuristic theory” as an explanation for the effects of procedural justice: fairness judgments are important because they help to reduce uncertainty, and individuals rely on procedural justice cues to make assessments of satisfaction when there are no available cues about distributive justice or outcome favorability.

The large role played by courtesy and respect in forming judgments about procedural justice and the relative accuracy with which individuals perceive courtesy and respect treatment suggest that individuals are most interested and attuned to interpersonal treatment in negotiation, supporting the relational, group value model of procedural justice. Of course, these findings do not rule out the possibility that individuals believe that disrespectful and rude interpersonal treatment is likely to come from someone who is also committed to providing an unfavorable or unfair outcome (the instrumental model), or that individuals use disrespectful and discourteous behavior as a proxy for evaluation of the outcome (the fairness heuristic model). However, the strong focus on courtesy and respect as opposed to other factors, such as neutrality or trust, that might more directly relate to the nature of the outcome, suggest that people are particularly concerned with relationships and interpersonal treatment even in resource allocation decisions.

Of particular note here is not simply that the courtesy and respect variable was the strongest antecedent of procedural justice in the negotiation context but also that the subjective ratings of courtesy and respect behavior by negotiators had a significant correlation with objective indicia of courtesy and respect behaviors by third-party coders. In contrast, there was a lesser connection for voice behavior and no observable behavior that related to the development of trust that in turn related to parties’ perceptions of their negotiation partner’s trustworthiness. Although negotiators’ own observable trust-related behavior correlated with some third-party assessments, this is less useful, because parties’ ratings of their own trust-related behavior were not particularly important in their formation of procedural justice judgments, and were far less important than their assessments of their partners’ trustworthiness. This suggests that individuals are better at assessing whether or not they have been treated with courtesy and respect than they are at knowing when they ought to trust the other party, a finding that is certainly consistent with research on lie detection (Depaulo 1992).

Because trust behavior was not particularly observable by third-party coders, this leaves open the question of how individuals form their judgments about the trustworthiness of their negotiation partner. Individuals may use reputation or other behavior that this research did not explore as the basis for making inferences about trustworthiness. Additionally, the findings here suggest that a party’s own trust-related behavior is a significant predictor of her assessment of the other party’s trust-related behavior, which implies, especially in the absence of effects for third-party ratings of partner trust-behavior, that decisions about trustworthiness may be at least in part projections of self-behavior and motive onto others—what interpersonal perception theorists refer to as assumed similarity (Kenny 1994). Finally, this study considered components of trust that included information sharing and deception, and it may be preferable in future studies to isolate these items from one another and conduct separate analyses on these potential dimensions of trust judgments.

6.3.2 Procedural Justice in Legal Negotiation

Although they are not labeled as “procedural justice” per se, many elements relating to the fairness of process are contained in discussions of effective negotiation strategies. For example, the seminal negotiation book Getting to Yes (Fisher et al. 1981) touts the importance of “separating the people from the problem” (p. 17, pp. 32–35), suggesting active listening and general courtesy, corresponding to voice and courtesy/respect above, as important tactics for an effective negotiator. Similarly, in Beyond Winning, Mnookin and colleagues (2000, p. 47) suggest the importance of “demonstrating an understanding of the other side’s needs, interests, and perspective, without necessarily agreeing,” which shares some underlying connection with the perception of voice. Getting to Yes also stresses the importance of independent and objective information in negotiation, consistent with the neutrality factor in procedural justice.

This research provides an empirical exploration of the importance of these factors touted by legal scholars. The role of courtesy and respect is particularly unequivocal in these findings, suggesting that treating another party respectfully—dovetailing with the principled negotiation technique of separating the people from the problem—may be particularly important in fostering perceptions of procedural justice. The research presented here also suggests a significant role for voice, echoing the suggestions of interest-based negotiation strategy. This research is more ambiguous with respect to the role of trust, and neutrality. What is of particular note here is (1) that one’s own behavior is meaningful in creating perceptions of procedural justice, and (2) that factors identified as potential “objective criteria,” such as reference to likely outcomes at arbitration or other relevant standards, rather than bolstering perceptions of neutrality and thus procedural justice, were linked negatively to perceptions of neutrality and to perceptions of procedural justice.

With respect to the use of objective criteria, these findings pose a challenge to the “bargaining in the shadow of the law” paradigm proferred by Mnookin and Kornhauser (1979). In that model, endowments provided by existing legal precedent cast a powerful shadow of influence on negotiation outcomes. Yet to the extent that parties rely explicitly on appeals to external criteria that are embodied in the law, this appears problematic from the procedural justice perspective. That is, procedural justice suffers when the law casts its shadow on the negotiation. This may be because negotiators felt that arguing about the law was outside the scope of the norms for negotiation. Certainly, one would be surprised to find a negative effect on procedural justice for use of legal standards and rules in a court proceeding, but perhaps because negotiation is expressly an extra-legal process, the use of these standards here was viewed negatively. Additionally, lawyers are trained from the earliest moments of law school to craft arguments and support them with case law; this is expressly a partisan and biased process. To resort to “neutral” criteria such as a legal standard may simply be an oxymoron: no law is truly neutral and there is almost no legal principle that cannot be argued from two different sides. Thus, a deployment of legal standards in the negotiation context may signal partisanship and a lack of neutrality that is particularly harmful to perceptions about fairness of the negotiation process.

Another potential explanation for the negative effects of discussion of potential arbitration outcome is that it could signal to all parties that the speaker is not fully committed to a negotiated outcome, thereby undermining trust in the other party during negotiation. That is, mentioning the possible outcomes at arbitration was a sign that a party did not necessarily intend to reach agreement in the negotiation setting and was more likely to be wasting the other party’s time because of a willingness to jettison the negotiation process in favor of arbitration. However, the findings presented here did not demonstrate a relationship between express statements about intent to reach agreement during negotiation and perceptions of procedural justice, providing no additional support for this hypothesis.

6.4 Limitations and Future Research

The population considered here were first year law students placed in the role of attorneys, rather than either (1) practicing lawyers or (2) laypeople engaged in a dispute. Law schools work hard in the first year to socialize students to “think like lawyers,” but law students may differ in meaningful ways from practicing attorneys. For that reason, a field study to support these results would be highly desirable.

The study presented here is limited by issues related to reliability: some variables were difficult to code with consistency among raters. In particular, variables such as interrupting can be subjective because some interrupting is perceived as helpful (adding a word, completing the other party’s sentence in a friendly manner) and some interrupting is perceived as negative (cutting the other party off, changing the topic), and it is sometimes ambiguous to coders which type of interrupting is occurring. Additionally, because this research was exploratory, 27 variables were included for coding; reliability of coding might improve with a smaller number of more targeted variables on which coders could focus their attention. Future research could provide even more stringent standards for coding behavior, and could drastically reduce the number of variables. Also, because each video was approximately 45 minutes, it was challenging to keep track of so many variables; studying a shorter time period might improve reliability, although it would be likely to decrease external validity because negotiation exchanges do typically occur over longer time periods.

Another shortcoming of this study is that both voice and courtesy and respect variables were reverse-coded. That is, coders searched for incidences of negative voice behavior (interrupting) and negative courtesy and respect behavior (contentiousness and hostility). Efforts to code for more positive behavior, such as active listening and expressions of courteous and respectful deference to the other party, were not pursued after pilot testing showed them to be highly unreliable. For that reason, this study analyzes what are basically extreme examples of lack of voice and discourtesy and disrespect. This means that this research is unable to answer the question of, first, what ordinary variation in provision of voice and treatment with courtesy and respect occurs in negotiation, and secondly, whether differences in such levels of voice and courtesy and respect that do not rise to the level of obvious stifling of voice, discourtesy, and disrespect are meaningful in relation to assessments of procedural justice. In other words, the present research leaves unanswered the question of whether “shades of gray” in positive provision of voice and courtesy and respect are meaningful, or whether, instead, there is some basic level of expectation about voice and courtesy and respect that, when violated in an obvious manner, leads to lower perceptions of procedural justice.

7 Conclusion

Prior research suggested that procedural justice in legal negotiation has meaningful effects on satisfaction with the negotiated agreement and potential future adherence to the terms of the agreement, as well as on the potential for integrative bargaining. The present research explores more deeply what procedural justice in the negotiation context actually means. Courtesy and respect appear to be the dominant factors driving assessments about procedural justice, and the perceptions of an individual about whether or not courtesy and respect were present in the negotiation are borne out by the assessments of third-party observers of the negotiation. Negotiators are largely self-focused with respect to how they perceive their dyadic partner, assuming a degree of similarity in the behavior of others with respect to the potential antecedents of procedural justice as well as procedural justice assessments themselves. However, negotiators are also somewhat accurate in their perceptions of other parties in relation particularly, to courtesy and respect and voice behavior. For psychology, this research taken as a whole suggests the broad scope of procedural justice effects, even beyond the context of third-party decision making, and provides further support for the relational model of procedural justice. For negotiating attorneys, this research suggests the ongoing importance of civil and respectful behavior in dispute resolution, even in the sometimes difficult context of our adversarial legal system.