Facilitating Conflict Transformation: Mediator Strategies for Eliciting Emotional Communication in a Workplace Conflict

Unlike many other alternative dispute resolution (ADR) processes, mediation “has the capacity to transform the quality of conflict interaction itself, so that conflicts can actually strengthen both the parties themselves and the society they are a part of” (Bush and Folger 2005: 19). While there is no universally accepted model of mediation practice, we agree with Robert Baruch Bush and Joseph Folger's (2005) claim that mediation's transformational capacity is often sacrificed at the expense of reaching agreement. This privileging of achieving a settlement over improved understanding leads to a short‐term focus that ignores the value of addressing underlying needs for long‐term effectiveness. Although there are disputes in which the goal of transformation is largely irrelevant and even at odds with the parties' wishes (Mayer 2004), previous research has shown that conflict interaction can affect future interactions (Gayle and Preiss 1998) as well as workplace climates (Friedman et al. 2000). Transformative approaches to conflict can therefore be important when the parties are in an ongoing relationship, such as in family or workplace conflicts.

Previous scholars have pointed out the centrality of emotion in conflict, suggesting that conflict transformation requires attention to emotion (Galtung 1996; Jones 2005). Consequently, mediators who employ emotion‐focused strategies may be more likely to surface underlying issues, increase understanding, and facilitate conflict transformation. It is at this intersection of theory and practice that this study seeks to offer guidance by identifying strategies used by mediators to help disputants call attention to their emotional experience.

The subject of emotion has received increasing attention from scholars from a variety of disciplines in the past decade. Here we examine three broad areas of research that are essential to our discussion of the topic. First, we briefly define emotion by way of reviewing, in very broad strokes, the different approaches taken to examine it. Next, we describe our view of the relationship between emotion and conflict transformation. Finally, we examine literature specifically linking emotion, transformation, and mediation.

Scholars from such fields as neurology, physiology, psychology, anthropology, and sociology have defined emotion based on their particular epistemological view of the world and their particular research purposes (see Lewis and Haviland 1993 for a good sampling). These definitional complexities are compounded by the use of other terms such as “feelings,”“mood,” and “affect” that are often used interchangeably with emotion.

Despite the academic breadth and phenomenological complexity, most social science scholars agree that emotion involves three basic components: a cognitive component, a physiological component, and a behavioral component, although different emphasis is placed on each component depending upon one's perspective (Jones 2005). The cognitive component acknowledges the role of perception and interpretation of events in emotion. The physiological component involves the somatic and “felt” nature of emotions. And, the behavioral component speaks to “action tendencies” and the subsequent expression (including verbal, nonverbal, intentional, and unintentional expression) that emotions trigger.

We use the term “emotional experience” to capture the pragmatic essence of emotion in the conflict process and to sidestep the definitional complications inherent in any single theory of emotion. Our focus is on how the experience of emotion affects conflict interaction. For example, the felt physical experience of emotion can alert someone that he or she is engaged in a conflict and predisposes him or her to act in certain ways (based on learned cultural patterns and idiosyncratic qualities). The way people express their emotions can communicate to others myriad messages, especially if the disputing parties share a history. For instance, silence may suggest either that a person is preparing to engage in a conflict or wants to avoid confrontation. For parties without a relational history, the interpretation of others' emotions is more difficult.

Because of the cognitive aspect of emotion, we believe that through facilitated discussion, parties can contemplate their own and the other's emotional experience, lending clarity to the conflict issues and their importance, and helping the parties identify possible areas of common ground. Such clarity also empowers parties by helping them better understand their own point of view, as well as the other party's perspective and by providing them the opportunity to empathize or at least engage in informed communication. Improved ability to communicate our emotional experience should improve self‐efficacy and allow all parties to feel more in control of the process.

It is important to begin with what we mean by transformation and why we contend that this is an important goal for conflict management. Kenneth Cloke and Joan Goldsmith (2000: 5) have defined transformation as “. . . all encompassing, lasting change. It is not minor, incremental, small scale, linear, or transitory. It leaves us different from the way we were before, and alters our sense of reality.”

When Terrell Northrup (1989) discussed the transformation of intractable conflict, she distinguished settlement from transformation by identifying three levels of change. A first‐level change is peripheral to the identities of the parties, such as a mandated “cease fire” that halts the conflict activity but does nothing to change the underlying issues. A second‐level change may involve new patterns of communication between parties but still does not change their core identities. A third‐level change involves changes in the core identities of the parties, how they see themselves, the other party, and the relationship between them, resulting in conflict transformation.

Johan Galtung (1996: 70) stated that all conflict generates energy, and the problem lies in how to “channel that energy constructively.” His triadic theory rests on the premise that “there is no viable alternative to creative conflict transformation” (1996: 70), which he contrasts with approaches to conflict that focus on solution, resolution, and dissolution.

Galtung's (1996) triadic model of conflict transformation, defines conflict as (1) parties' underlying attitudes toward the conflict situation (which includes orientation to the issues and to the relationship between the parties), (2) their manifest behaviors as related to the conflict interaction with their opponent, and (3) the conflict issue itself, which he terms the “contradiction.” While he does not refer to emotion per se, attitude research emphasizes the importance of affective or emotion components (Breckler and Wiggins 1992). For Galtung, conflict transformation can only occur when all three aspects of the triad become “articulated,” or brought into awareness, understood by the parties, and then addressed in interaction aimed at resolving the conflict. Further, he contends that in conflict resolution practices, usually only one or two parts of the triad are addressed, such as attending to the contradiction and accompanying behaviors. What is most frequently ignored, Galtung argues, is open discussion of the underlying attitudes and emotions that motivate the parties' behavior. Admittedly, Galtung's model is more complex than this summary suggests; but it clearly differs from other models of conflict by defining conflict as a constellation of elements that includes the emotional experience and orientation of the parties.

Galtung's (1996) focus on the importance of the underlying emotions and the essential need to address them in interaction is supported by Cloke and Goldsmith (2000: 7), who stress the need to “embrace and acknowledge emotion” as one of the eight necessary steps for conflict transformation, as well as by Tricia Jones (2000) who has argued that emotion is the essential element of conflict. We have previously argued that conflict is more likely to be fully articulated and thus transformed when emotion is explicitly addressed (Bodtker and Jameson 2001).

Jones and Andrea Bodtker (Jones 2000; Jones and Bodtker 2001) have suggested that emotional experience indicates how a disputant frames (defines) the conflict and is also an essential component to address in mediation in order to influence (transform) the outcome. Although they did not specifically speak of transformation, their model seeks to achieve both of the objectives identified in Galtung's (1996) and Cloke and Goldsmith's (2001) models.

Jones and Bodtker (2001) drew from Richard Lazarus's (1991) appraisal theory of emotion to support their theory. Briefly, according to Lazarus, emotions arise from a two‐stage cognitive appraisal process. In stage one, individuals experience a triggering event and assess the situation in terms of its relevance to their own goal achievement (goal relevance), whether the event makes it easier or more difficult to achieve their goal (goal congruence), and its impact on their identity (ego involvement). In stage two, individuals focus on additional issues such as judgments of accountability (who is to blame), coping potential (how well they can deal with the problem), and future expectancy (whether things will get better or worse without taking action). On the basis of these appraisals, specific emotions will be experienced. For instance, people experience anger when they assess a situation as being personally relevant (in terms of goal blockage or identity), when another person can be blamed, and when the interference is seen as unfair.

This model offers a means by which to “transform” conflict. Specifically, if a mediator or third‐party adjunct can focus participants' attention on the appraisals they are making, the parties have the opportunity to “see” the conflict differently, resulting in a different emotional experience. For instance, with the use of elicitive questioning, a disputant might reassess her coping potential. Or she may learn that the other person did not intentionally interfere with her goal achievement. Changes in either of these appraisals will alter the framing of the situation, thus a different emotion. Identifying and understanding one's emotions and examining the “causes” (appraisal sets) in an open discussion can provide additional specific, useful information for all parties involved. It accomplishes Galtung's (1996) goals by facilitating articulation of all three conflict elements and thus has greater likelihood of resulting in a lasting and satisfactory resolution.

Perhaps the clearest argument for the connection between mediation and conflict transformation comes from Bush and Folger's (1994, 2005) description of transformative mediation. Transformative mediation describes a mediation style that emphasizes empowerment and recognition with the goal of helping parties see themselves and each other in a new light, thus transforming their relationship. As Bush and Folger have described it, and following the work of Galtung (1996) and Northrup (1989), when parties engage in a process of communication that fully articulates their underlying attitudes, this helps them to better understand each other's behavior and often eliminates the conflict. While Bush and Folger have not specifically discussed the role of emotion in mediation, the underlying principles of empowerment and recognition contain emotional elements related to identity, self‐esteem, and empathy. We believe that eliciting communication about emotion is one path to recognition and therefore a possible strategy for transformative mediation.

Other research on mediation has also addressed its potential to bring about more effective, transformational outcomes. Third parties can enhance the conflict management process by fostering emotional insight in individuals to help them better understand their own feelings as well as the feelings of others (Kennedy‐Moore and Watson 1999). In an early study of perceptions of procedural justice in mediation and arbitration, Debra Shapiro and Jeanne Brett (1993: 1170) suggested the important role of the third party when they argued, “Being able to express one's opinions about the dispute and believing that the third party listened to those opinions may serve both to legitimize those emotions and dissipate them.”

Some have considered how the mediator can use reframing to help parties see the conflict in a new way (Putnam and Holmer 1992; Bodtker and Jameson 1997). Jean Poitras (2007) recently found that mediators can facilitate reconciliation by helping parties acknowledge their own roles in conflict escalation. In these studies, mediators deal directly or indirectly with parties' emotions as they move from anger at another party to acceptance of their own culpability, for example, or in other ways reconcile their original conflict reactions with new emotions that emerge through the mediation process.

Is attention to emotion more likely or more appropriate to one particular style or model of mediation than another? Bush and Folger (1994, 2005) have used the terms facilitative to describe approaches that focus on helping parties improve communication and gain clarity, evaluative to describe approaches that give parties insight into the difficulties of litigation and their likely success in a courtroom and thus encourage compromise, and transformative to characterize those approaches that emphasize empowerment and recognition with the goal of changing the relationship. A recent survey of two‐hundred‐fifty mediators from a variety of settings reported that, regardless of how mediators defined their own style, most agreed about which strategies were used “often” or “sometimes.” Of the most commonly reported strategies, two were directly related to emotion: “summarize or paraphrase feelings already mentioned by the participants” and “check out possible feelings of the participants, according to what the participants have implied” (Charkoudian et al. 2009: 299).

While mediators may perceive that they attend to emotion, Lori Schreier's (2002) research on emotional intelligence and mediation training indicates that the field of mediation suffers from a lack of attention to emotion and its consequences. She found that “close to half of the respondents [mediators], including two‐thirds of those with the most experience, thought that mediation training does not sufficiently teach how to address the parties' emotional reactions” (Schreier 2002: 107). Furthermore, 67 percent of mediators who primarily used the transformative approach reported that mediation training did not sufficiently teach them how to handle disputants' emotional reactions (Schreier 2002: 106). Furthermore, at least one study suggests that the expression of emotion might affect the ability of a third party to accurately assess parties' interests and intentions (Thompson and Kim 2000).

Jones (2005) suggests three skills that mediators can use to recognize and deal with emotion in mediation: decoding the emotional experience of a disputant, helping the disputant understand his or her own emotional experiences, and facilitating the disputant's reappraisal of the emotion to remove the obstacle of the emotional experience. Drawing on Lazarus's (1991) appraisal theory and Brant Burleson's work in supportive messages (Burleson and Goldsmith 1998), Jones also suggests specific types of questions mediators might ask in order to elicit emotional communication from parties to help with decoding, appraisal, and reappraisal processes.

Our purpose here is to move beyond theoretical speculation to identify the strategies that mediators actually use to elicit emotional communication during the mediation process. We also examine the potential effects of those strategies with the goal of developing curricula to build mediator capacity for working with emotion in conflict. The research question that guides our analysis is: What communication strategies do mediators use to elicit emotion and help parties better understand their own and each others' emotions?

We derived our data from a larger research study that examined differences in the process and outcome of two groups based on the mediator's approach. In condition one, mediators were instructed to focus on eliciting the parties' emotional experience (emotion positive group). In condition two, mediators were instructed to ignore or minimize implicit or expressed emotions (emotion neutral group). We designed the simulations to test for differences between the two groups in terms of the parties' satisfaction with the process and in terms of the type of agreements that the parties reached.

Participants played the roles of coworkers in a work‐related dispute who were referred to mediation. The conflict scenario was written to be both urgent and realistic, and the disputants were interdependent. (The scenario had been designed for a previous study and had undergone a manipulation check on these variables; see Jameson 1999.)

In the simulation, Pat is a junior accountant who has become dependent on a senior accountant, Chris, for assistance in learning the new job. While Chris was initially helpful, she/he has become annoyed that Pat appears to be overly dependent and a slow learner, particularly because Chris seeks a promotion. Pat faces an upcoming three‐month review and worries about losing his/her job. The scenario describes various personal issues that affect Pat and Chris but that are unknown to each other. These details provide the participants with additional cues about the emotions they might experience in this conflict, such as fear of unemployment, frustration, a stressed relationship, and lack of appreciation. On the task side, both Pat and Chris have a very specific (shared) conflict they must address — how they can each get their respective jobs done without harm to the other. They must also develop/maintain a positive relationship because the promotion Chris hopes to earn will require her/him to more directly supervise Pat.

The third parties in the simulation were eighteen experienced (one year minimum) mediators who participated in a six‐hour workshop on the nature and role of emotion in conflict and mediation. The workshop comprised activities designed to help mediators encode and decode emotional expressions to consider how the parties' various emotional experiences can affect their framing of the conflict and how, through communication, mediators might assist parties to identify and examine their and the other party's emotions (for more on the workshop, see Jameson, Bodtker, and Jones 2006). In exchange for the free workshop, mediators agreed to mediate the two role‐play conditions for the study (emotion positive and emotion neutral). Disputant participants in the study consisted of thirty‐six undergraduate students from two Eastern universities (twenty‐five female and eleven male). Students either received extra credit or course credit for their participation. We conducted a total of eighteen mediation role‐plays but only examined eight in this study.

Students were randomly assigned to the role of “Pat” or “Chris” and prepped in separate rooms based on their roles. They were instructed to read their characters' role, imagine themselves in a similar position, and act and respond throughout the role‐play in the way they thought they would naturally. They were given time to reflect on their characters and completed a survey instrument designed as a pretest. After about fifteen minutes the students were brought to another room where they met the other party and the mediator. In some cases, the students knew each other (from classes they shared) and in some they did not. The partiesdid not speak to each other after receiving their roles and prior to the mediation simulation.

We note that the original study did not find significant differences between the two mediation conditions. We surmise that this result is related to two limitations of our initial design: the students' naïve role as participants, resulting in their satisfaction with mediation as a conflict management intervention regardless of style, and insufficient control of the mediator strategies in each condition, resulting in the possibility that there were no significant differences in the actual mediation behaviors of the mediators in the two different groups (for more discussion, see Jameson et al. forthcoming). In this study, we have examined the transcripts closely to identify the strategies mediators use to elicit emotional experience regardless of their mediation styles.

We videotaped each of the mediation simulations. For the current study, we transcribed eight role‐plays in order to conduct a textual analysis of the mediations (Geisler 2004). (Text analysis is a research technique used to determine what can be learned from verbal data about a phenomenon of interest.) In this study, we analyzed the text to examine how mediators address parties' emotions in mediation. While it may be argued that role‐played mediations do not reflect naturalistic mediation sessions, there has nonetheless been a long and robust history of negotiation and mediation research based on simulations because real conflicts are much more difficult to gain access to and manipulation of the variables in a real conflict is difficult, if not impossible. Moreover, we are most interested in identifying the strategies that experienced mediators use when they are trying to elicit emotion. In this study and for this purpose, our use of real mediators thus reduces the limitations of the role‐play method.

We used a “best‐case” sampling method to select simulations we believed most relevant to the phenomenon we sought to examine (Geisler 2004). We analyzed a total of eight role‐plays, including four that were emotion positive and four that were emotion neutral. The duration of the simulations ranged from 25 to 120 minutes, resulting in 73 single‐spaced transcript pages. Table One displays the lengths of the role‐plays included in this analysis.

We began by independently coding two transcripts that were not included in the sample. Using the constant comparative method (Strauss and Corbin 1990, 1994) we used previous research (particularly informed by the work of Jones 2005) and examination of transcripts to identify possible emotion‐eliciting strategies. Once we had a list of possible strategies, we compared our coding and discussed differences until we reached agreement on categories and coding rules. After revising the coding sheet, we coded two new transcripts and repeated the process, revising the coding categories again four times. The final set of codes included eight emotion‐eliciting strategies:

1. grant legitimacy;

2. encourage emotional identification;

3. help the person deny the emotion (e.g., to save face);

4. challenge an emotion label (the way an emotion has been defined within the mediation);

5. confront emotion avoidance;

6. paraphrase emotion;

7. encourage emotional perspective taking; and

8. probe meta‐emotions.

After the final iteration, we trained a third coder and Jessica Jameson and Tim Linker independently coded the eight role‐plays in our sample. Because each time the mediator takes a turn to speak, he or she might utter several sentences, we used an Excel spreadsheet to identify each “t‐unit”, “the smallest group of words that can make a move in language” (Geisler 2004: 31).

For example, the following is one mediator‐speaking turn: “So you're feeling very uncomfortable about what's going on and you don't quite understand what's happening with his connection to you? We are just going to have to explore what this is all about.”

We broke this down into three t‐units:

1. “So you're feeling very uncomfortable about what's going on . . .”

2. “. . . and you don't quite understand what's happening with his connection to you?”

3. “We are just going to have to explore what this is all about.”

Each t‐unit was coded with a number 1–8 to indicate either one of the eight emotion‐eliciting strategies or a 9 to indicate the t‐unit was not an emotion‐eliciting statement. The number of t‐units ranged from 95 to 343 per role‐play (the mean was 164.63). Because the beginning of each role‐play included the mediator's introduction to mediation and each ended with a recap of the final agreement, we did not code the first or last 10 percent of each transcript. A total of 979 t‐units were coded.

We computed intercoder reliability using Cohen's Kappa, a popular measure of interrater agreement for qualitative items that takes into account agreement that might occur by chance. The reliability index was 91 percent (anything over 80 is considered very high agreement). We calculated frequencies for each category of emotion‐eliciting strategy so we could answer our research question regarding the types of strategies the mediators used. We also undertook a more qualitative reading of the transcripts and the resulting written agreements to examine common themes. In the discussion section, we provide examples of particularly interesting interaction sequences that illustrate how emotion‐eliciting strategies might facilitate conflict transformation.

While we began with eight emotion‐eliciting strategies, only five surfaced in our sample. Table Two illustrates the five emotion‐eliciting strategies and their frequencies within each simulation as well as their total frequency across all the simulations we analyzed. We note that the majority of t‐units were coded as 9, “not emotion eliciting” (n = 841, 85.9 percent). The five strategies identified were: grant legitimacy (n = 7, 0.71 percent), encourage emotion identification (n = 48, 4.90 percent), confront avoidance of emotion (n = 2, 0.20 percent), paraphrase emotion (n = 66, 6.74 percent), and encourage emotional perspective taking (n = 15, 1.53 percent).

Below we provide examples of each strategy.

This strategy occurs when a mediator verbally or nonverbally acknowledges the emotion a disputant is expressing. Specific actions range from smiling and nodding (which were not coded here due to lack of consistency in video quality) to expressing empathy and positive emotions toward the disputants. We coded this strategy just seven times in our sample, but we suspect we would have found greater instances had we been able to code the nonverbal interaction. The following mediator statements were coded as granting legitimacy:

In each of these cases, the mediator acknowledges and validates the emotion one or both parties has expressed. In the first example, the validation of Chris's experience of frustration may make him more comfortable elaborating on this emotion and helps Pat become aware of how Chris frames the conflict. It also provides an opportunity for both Chris and Pat to reappraise the situation.

In the second example, “you are both expressing your needs, so that's good,” the mediator is acknowledging the value of emotional expression and encouraging each party to continue to speak openly. In the third example, the mediator again validates one party by praising him or her for “thoughtful and attentive” behavior, which may encourage additional emotional expression and/or pave the road for more conciliatory interaction between the parties.

This was a common strategy, occurring forty‐eight times, the second highest in our sample. The mediators' strategy may be to decode the disputant's emotion while helping her acknowledge her own emotions (see Jones 2005). Mediators in this sample often directly asked parties how they were feeling, such as “And how are you feeling when he says what he says?” Mediators also try to clarify the emotion they think they are hearing expressed, such as the statements: “Am I wrong that you feel isolated?” and “You are anxious I gather to please the people in the company.” These strategies often ask disputants to elaborate on emotions, presumably to help mediators decode the emotion as well as help each party better understand what he and the other party are feeling and why. An example of this is: “Other than feeling alone, how are you feeling about the situation?” This appears to be a fairly common mediator approach to identifying parties' underlying needs and concerns, which Lorig Charkoudian et al. (2009) also found.

This strategy occurs when a mediator suggests that a disputant is denying feeling a particular emotion when his behavior suggests otherwise. We found few examples of this strategy, which we believe is because mediators are more likely to use this strategy in caucus and only one of the eight role‐plays in our sample used a caucus strategy. In fact, the two instances we found did occur in caucus. The examples of statements indicating this strategy are:

In each of these examples the mediator appears to suspect the party feels something he or she is reluctant to express, either to the mediator, the other party, or both. We believe this is an important strategy for helping disputants express emotion and identify underlying concerns, particularly in the workplace where norms often lead to emotion suppression (Cloke and Goldsmith 2000; Fineman 2003). While this can be a powerful and enlightening strategy when it is successful, mediators should be cautious about a party's comfort in admitting emotions he has not identified himself. If it backfires, it could make him feel embarrassed or defensive, especially if it is an identity‐based emotion (see below) or is in conflict with what he has learned about the appropriate experience and expression of a particular emotion (i.e., a meta emotion; see Gottman 1994, for a thorough description).

This was the most commonly used emotion‐eliciting strategy and occurred sixty‐six times in our study. The following examples illustrate how this use of paraphrasing specifically focuses on emotion.

“So she was really helpful, but now you are feeling cut off.” Acknowledging that the party feels “cut off” may suggest fear of losing a valuable resource and/or a threat to the relationship between the parties. It may also signify anger toward the other person if the sever is seen as intentional.

A second example: “She was answering all your questions and now you feel self‐conscious about it.” Self‐conscious emotions are inherently identity based (Lewis 1993). Embarrassment is a “softer” form of shame in which a person blames herself for the actions or events that led her to experience a particularly distressing emotion. This can be a powerful revelation for all parties involved because it suggests that in order to “transform” this emotion (e.g., feel better) the party must reappraise the situation in terms of self‐efficacy (e.g., by recognizing her own actions are not to blame, thus reducing the embarrassment). Such self‐disclosures should be handled with sensitivity because they can make a person feel vulnerable. Such revelations can elicit empathy from the other party and bring the parties closertogether, but they can also be used as “leverage” in the power relationship between the parties.

Another statement that indicates a paraphrasing strategy was “I'm hearing that you've got some frustration, or some discomfort in the way the relationship has proceeded after the initial period.” Frustration is a mild form of anger and/or fear and discomfort suggests fear. In this case, these emotions seemed to be specifically directed at the relationship between the parties. Further exploration of this emotion may reveal whether the party blames the other party, himself, or the situation and provides the opportunity to focus on the parties' continued relationship.

In each of these examples, the mediator sought to reveal the party's emotions, rather than focusing exclusively on the factual circumstances of the conflict. This strategy can enhance understanding of how the parties have framed the conflict and identify the underlying needs that must be addressed to reach agreement. (This also supports the finding of Charkoudian et al. 2009 that paraphrasing emotion was a common strategy across mediation types.)

This was the third most common emotion‐eliciting strategy we found in our sample, although it occurred infrequently with only fifteen statements falling into this category. While each of the strategies described previously can facilitate emotional perspective taking, with these statements the mediator intentionally directed one party to consider the situation from the other's point of view. For example:

Along with “confront avoidance of emotion,” this is probably the most difficult strategy available to mediators to elicit emotional awareness and emotional understanding. Both are directive and some mediation theorists have specifically warned against these approaches (e.g., Bush and Folger 1994). It may be possible that mediators shy away from such a direct approach for fear of making a party feel coerced or manipulated. It could lead to an insincere agreement if a party has felt bullied, or it could generate distrust if, for example, one party feels that the other party did not acknowledge him of his own accord, but was forced into it by the mediator. That said, in the mediations we examined, this technique seems to have yielded positive results.

Mediation is an ADR method that has the potential to transform relationships and we have argued that understanding and addressing underlying emotions is a necessary component in achieving conflict transformation. Because previous scholars have suggested that mediators need more training in dealing with emotion as part of the mediation process (Schreier 2002; Jones 2005), our primary goal was to identify specific communication strategies mediators use to elicit the discussion and exploration of disputants' emotions during the mediation process. Our assumption is that this is a necessary first step in developing and implementing an emotion‐focused training component for new and experienced mediators. We identified five emotion‐eliciting strategies used by mediators in a role‐play scenario and suggest that these strategies — grant legitimacy, encourage emotion identification, confront avoidance of emotion, paraphrase emotion, and encourage emotional perspective taking, have practical implications for mediation training.

As we read through the transcripts, we noted that some of the most interesting outcomes of these strategies occurred when they were used in concert: the mediator began with encouraging emotion identification to get the discussion started; moved to paraphrasing, perhaps to make sure the emotion was correctly identified; and then encouraged emotional perspective taking to make sure both parties understood the emotion being described. The following excerpt from a mediation in which the mediator played particular attention to emotion provides an example of this process. We have annotated this excerpt (annotations are in brackets) to illustrate how the mediator strategies were coded and what we identify as a possible interpretation of the impact on the disputants:

This role play resulted in one of the most comprehensive written agreements, which included two specific statements of respect: “Chris and Pat both agree to understand the demands on each other's time and respect each other's needs” and “Both Pat and Chris state they respect the job the other is doing.” Their agreement included the more typical instrumental support they would provide each other, such as Chris introducing Pat to others in the office. While some of this can be attributed to the mediator's style, the two statements of respect were generated by the parties and were not typical of the final agreements in our sample. This supports our argument that an effort to elicit emotion helped transform the parties' relationship and also encouraged them to formally acknowledge each other's interests and concerns.

While the previous mediation lasted two hours, the next excerpt comes from an emotion‐neutral mediation that was completed intwenty‐five minutes. The following exchange occurred near the end of the session and illustrates the potential impact, as well the complexity of, emotional communication:

This mediation was much shorter in length than the one excerpted earlier and produced an agreement with only four items (half as many as the previous agreement). This agreement included Pat's statement of apology as well as a statement that both parties would “respect the deadlines of the other.” The other two items were specific instrumental details. While the second mediation was arguably effective and more efficient, we do not see the same level of recognition occurring in this example as we saw in the previously excerpted mediation in which the mediator spent more time exploring emotion throughout the process. We cannot, however, be certain whether the apology voiced in the first example reflects sincere recognition or was a strategic maneuver on the part of a lower‐power party to ingratiate herself with a higher–status colleague. Text analysis cannot provide us with clues as to the speaker's motive in this example, and a compelling follow‐up question would explore the long‐term impact on the conflict when the emotional expression is strategic rather than “genuine.” While this study cannot answer that question, it underscores the complexity of the role of emotion and emotional communication in mediation.

Readers may wonder whether authentic emotion is possible in a role‐played scenario, particularly because the participants in this scenario faced no real outcomes — we did not simulate the existence of a future relationship between role‐play participants or fabricate a payoff structure, which is often used in simulations to enhance participants' sense that they have a stake in the outcome. While access to actual mediation sessions would be ideal, we confronted at least two major obstacles to getting these data: because mediation is confidential, most mediation environments would not allow for audio or video recording, which was needed for the fine‐grained text analysis we wanted to do, and we wanted to access a specific type of conflict in which the parties will have a future relationship and thus are most in need of transformation. Finally, even if we had access to actual workplace mediations, it seems likely that the presence of a recording device would have had some impact on disputants' willingness to engage in the very phenomenon of interest: communication and expression of emotion. For all these reasons, the role‐play seemed the most useful method for our purposes.

We believe that participants expressed some real emotions during this role‐play. First, this role‐play scenario was based on a situation that many students are familiar with — a coworker who is perceived as not doing his or her share and as overreliant on another worker. Any student who has participated in group projects is likely to have experienced one side or the other of this scenario. Furthermore, on more than one occasion disputants were brought to tears, either because they felt ashamed in their role or because they felt relief or happiness at the end. When we debriefed the participants many of them shared the emotions they were experiencing throughout the role play, and they expressed their surprise that they felt this way even though they knew the situation was not “real.”

Aside from the limitations inherent in role‐plays, working with emotion presents its own challenges. As noted earlier, one is the difficulty in determining if an expressed emotion is authentic. In the second excerpt above, for example, it is possible that Pat's apology was strategic. In the role‐play scenario, Pat is concerned about his/her three‐month evaluation and it is clearly in his/her best interests to get Chris on his/her “good side.” While this mediation appeared to end well, we have no way of knowing how Chris interpreted Pat's apology. We caution mediators to be aware of the strategic use of emotion and its potential to undermine the effectiveness of the process and resulting agreement.

Another limitation of the text analysis method is that it did not allow us to confirm our interpretations of the mediation sessions with our participants. The best we can do here is share the emotion‐eliciting strategies used by mediators and suggest possible implications of those strategies for the mediation process.

A final comment involves individual differences in mediators' styles and the mediator's ability to effectively elicit emotion in mediation. Mediator #026 and Mediator #025 were both placed in the emotion‐positive group and were thus both specifically instructed to use emotion‐eliciting strategies. Mediator #026 used sixty‐three emotion‐eliciting strategies, which was significantly more than any of the others, while Mediator #025 used only seven. Two of the mediators in the emotion‐neutral condition, #022 with thirteen and #203 with nine, used more emotion‐eliciting strategies than did Mediator #025. This suggests that exploring emotion comes more naturally to some mediators and supports the finding of Charkoudian et al. (2009) that at least two of the emotion‐eliciting strategies we found seem to be established as common mediator strategies (encourage emotion identification and paraphrase emotion).

Clearly we would advocate for more theoretical work exploring the role of emotion in mediation and conflict transformation. While we are encouraged by the work being done on negotiation and emotion, much of this work is still focused on how emotions affect negotiation results, with practical implications for the strategic expression or suppression of various emotions (see Li, Tost, and Wade‐Benzoni 2007; Steinel, Van Kleef, and Harinck 2008 for reviews). One recent exception is the work of Roger Fisher and Daniel Shapiro (2005), who have explicitly developed the link between attending to emotion and negotiation. They have suggested that reaching mutual agreement is possible when negotiators attend to five emotionally charged “core concerns” that motivate conflict: appreciation, affiliation, autonomy, status, and role. Note that appreciation and autonomy in particular seem to parallel Bush and Folger's (1994, 2005) discussion of recognition and empowerment, although recognition might address any of these five core concerns as long as one party acknowledges its role in the conflict. Attention to these core concerns could have implications for mediators and thus represent a fruitful area for additional theoretical and practical work.

We would also like to further explore the link between conflict transformation in the workplace and improved social capital. Just as Barbara Gayle and Raymond Preiss (1998) found that recollections of previous conflict interactions influence further interaction, we contend that a focus on transformation will have a positive impact on future interactions and further relationships. This is consistent with the finding of Ray Friedman et al. (2000) that employees create their workplace climate through how they engage in conflict interaction. When employees engage in more cooperative behaviors, they feel more positively about the workplace and are more likely to be satisfied and collaborative in general.

Our example of the apology also highlights the need to examine the role of respect and trust in how disputants interpret each other's emotional communication. As Jones (2005) has suggested, mediators may need to do more than bring the emotion to the surface; they need to help parties appraise and reappraise the emotion — or see the emotion in a different way, perhaps by removing blame from the other party — in order to achieve conflict transformation. The role of the mediator in facilitating or restoring respect and/or trust is therefore another important avenue for further theoretical examination.

Mediation is often referred to as more of an art than a science, and experienced mediators will admit that much of what they do is guided as much by intuition as by training. But one of the difficulties new mediators face is knowing which questions will lead disputants to identify and discuss their underlying concerns and interests. This is particularly difficult when disputants seek to “save face” and make a good impression with other organizational members with whom they are in conflict. While we have seen the benefit of a transformative and emotion‐focused approach in our mediation practice, we have also found ourselves at a loss for words in the midst of a difficult mediation session. A main goal of this study was to identify specific questions that mediators can ask to elicit emotion and turn difficult mediations into more productive sessions, hopefully with transformative outcomes. While some of the strategies we identified are adaptations of typical mediator strategies such as legitimizing and paraphrasing, we have also identified new strategies, such as confronting emotion avoidance. A closer examination of these and other strategies could be used to develop training for mediators in how to address emotion. For now, we make the following recommendations for mediator training in emotion:

1. Discuss the centrality of emotion in parties' orientation to the conflict. Mediators must understand that emotion provides clues to underlying interests and accept that the communication of emotion is a facilitator for rather than a barrier to effective communication and conflict transformation.

2. Recognize that addressing emotion is not the same as having parties focus on the past. While the events that brought about the conflict may be in the past, the emotion parties are experiencing is in the present, and parties must reappraise the situation, or see it differently, in order to change any negative emotions they are feeling and lead to more positive future interaction.

3. Adapt existing mediator strategies, such as paraphrasing and perspective taking, to include discussion of emotions parties are experiencing as well as their perceptions of events.

4. Learn to use additional strategies that create a climate conducive to the discussion of emotion. Two that we have discovered are granting legitimacy, to validate parties' emotional experience, and encouraging emotional identification, to help one or both parties better understand the source of the conflict.

5. Use a variety of emotion‐eliciting strategies in concert. While there is no “magic bullet” that will ensure transformative outcomes, we found that some of the best examples of recognition occurred when mediators started by getting parties to identify their emotions, validated the parties by acknowledging their emotions, paraphrased each party's emotional experience to help both parties better understand it, and finally, either directly or indirectly, encouraged the parties to engage in emotional perspective taking to allow each party to hear that the other understood them and how they were feeling about the conflict.

6. Use the caucus to explore emotions. This may be particularly important when there is a power imbalance and one party is more vulnerable; one or both parties are particularly reticent about sharing their emotions; the mediator suspects one party feels shame or guilt, which could exacerbate power imbalances; and/or the mediator questions the authenticity of expressed emotion.

We have demonstrated a connection between encouraging emotional communication and identifying underlying interests, which has the potential to lead to conflict transformation. This study improves our understanding of how mediators attend to parties' emotions in the mediation session. To continue this research, we plan to examine additional transcripts to validate the strategies we found and uncover new ones. We also hope to get access to naturalistic mediations, where we can use postmediation evaluations to explore the strategies mediators use and compare mediator satisfaction with the parties' evaluations of procedural justice, satisfaction with the mediator, and perceptions of final agreements. We also hope to interview mediators about how they confront emotion and talk to them about the strategies we found and whether they use them (and/or others) in practice. With greater tools and strategies for addressing the emotions mediators inevitably confront in a mediation, we believe training in the communication of emotion can improve the chances that mediation will achieve its transformative potential.