The Paradox of Accepting One's Share of Responsibility in Mediation

A saying popular among mediators goes like this: “How many mediators does it take to resolve a conflict? Just one, but the parties must want to cooperate.” This adage illustrates a fundamental mediation problem: it is the parties in conflict themselves who ultimately must work together to reach a common solution. Mediation is, in effect, a voluntary process, and the mediator holds no power to impose a solution. If the parties refuse to cooperate, then the mediation is doomed.

Several variables can explain the emergence of a cooperative spirit in mediation. The wish to reconcile, acceptance of responsibility, hostility, and indifference to the other party's concerns have been recently defined as factors that can either encourage or discourage the emergence of cooperation (Poitras 2005). In this article, I will focus on the key role that accepting one's share of responsibility can play in establishing a cooperative relationship between the parties.

Responsibility is a tricky subject in mediation. Many mediators believe that directly addressing the issue of parties' responsibility encourages blaming, confrontation, and/or impasse. After all, mediation often focuses on the future and not the past. Consequently, mediators generally avoid the issue of individual responsibility for a conflict.

This article takes a three‐step approach to exploring the role that accepting responsibility plays in the mediation process. First, I provide a brief summary of various approaches to mediation, highlighting the means they offer for encouraging cooperation between parties. Second, I describe the methodology and results of an empirical study designed to look at the importance of accepting responsibility for the mediation process. Finally, I discuss the practical implications of the study results and make recommendations for mediators who want to incorporate into their mediation processes a method for encouraging parties to accept responsibility.

Motivating parties to cooperate to resolve their conflict is one of the mediator's fundamental tasks, but there is nothing like consensus among mediators on the best way to achieve that goal. The various mediation approaches (also called “models” or “styles”) do not favor the same points of intervention for encouraging cooperation. In this section, I examine how two well‐known mediation approaches, the problem‐solving model and the transformational model, deal with cooperation. Lastly, I examine the role that the parties' acceptance of their responsibilities plays in the emergence of cooperation.

The premise of problem‐solving approaches to mediation is that mediators must encourage parties to focus on finding a solution to the conflict by limiting negative communications between them (hostility, blame, etc.) in favor of stimulating the exchange of more productive information. This approach assumes that a primary problem is the parties' inability to maintain a problem‐solving orientation in their discussions.

Problem‐solving approaches to mediation include both the interest‐based model and the evaluative model. Interest‐based models (see Fisher and Ury 1981; Lax and Sebenius 1986) suggest that mediators help parties move beyond their initial positions by uncovering and then focusing on their underlying interests. Once this is achieved, according to this model, they are better able to find a solution that benefits everyone. The evaluative approach, on the other hand, suggests that mediators help parties evaluate the pros and cons of their situation and their positions — which can include the likelihood of achieving success through litigation — then exert pressure on them to accept a compromise in line with the results of that analysis (Riskin 1996).

At the core of the problem‐solving approaches lies the parties' willingness to settle the conflict. Fundamentally, the parties agree to negotiate a solution because they realize that the conflict situation cannot go on. Their interests are not served by the status quo, and they realize that they are unlikely to be able to impose their point of view on the other party or that to do so would be too expensive, too time‐consuming, too stressful (Saunders 1985). They must therefore agree to negotiate a compromise.

Consequently, the mediator seeks to “help” the parties recognize the impasse they have reached. To accomplish this, the mediator tries to convince the parties that maintaining the conflict is at odds with their own interests. He may also encourage the parties to consider the feasibility of their alternatives and the consequences of prolonging the conflict.

While this approach is considered effective for resolving conflicts in which the relational intensity is low, some scholars argue that it is of more limited effectiveness when the conflict is more emotionally charged. In a recent study, Sarrazin et al. (2005) demonstrated that, in divorce mediations with little emotional involvement, directive mediators (i.e., mediators who focused on problem solving and evaluating) are as efficient as mediators who adopt a more facilitative approach. The authors found, however, that the facilitative approach is much more effective than the directive approach in cases where emotions run high. Some authors have even gone so far as to suggest that a problem‐solving approach can intensify a conflict situation that is dominated by emotions and values (Burton 1996; Rothman 1997). Consequently, many mediation theorists argue that working on the parties' relationship can be an important early step in mediation.

Mediation approaches based on transforming relationships postulate that mediators should, as a first step, work to rebuild the relationship between the parties. The most popular variation on this approach is probably transformative mediation (Bush and Folger 1994). To transform the relationship between the parties, the authors suggest emphasizing recognition (concerns, interests, etc.) in communications and the parties' sense of empowerment (taking charge of the conflict, decisions, etc.) while looking for solutions.

Narrative mediation and insight mediation are two other models that seek to transform relationships. Narrative mediation (Winslade and Monk 2000) puts forward strategies that focus on helping the parties construct a shared story (narrative) to explain their conflict as a means of transforming the conflictual relationship and paving the way for a consensual solution. On the other hand, insight mediators (Picard et al. 2004; Picard and Melchin 2007) seek to develop the parties' awareness of the needs underlying the conflict as well as the relationship between their actions and their satisfaction. With this awareness, the parties can redefine their relationship and pave the way for a consensus.

The parties' mutual willingness to seek reconciliation is a critical component of approaches based on solving disputes by transforming the relationship. The parties agree to cooperate because they recognize the importance of rebuilding the relationship. Some authors have suggested that a “shadow on the future” convinces parties to cooperate (Axelrod 1984; Zubeck et al. 1992) — they do so because they know they will have to deal with each other again.

Thus, one of the mediator's goals is to make the parties aware of the importance of the relationship. By focusing on the parties' future relations with each other, he highlights the likelihood that they will need to deal with each other in the future. Because of the likelihood that the parties will be required to work together again, workplace disputes exemplify the type of conflict in which the relationship should be reestablished before a lasting solution can be achieved.

Can the desire to restore the relationship overcome the hostility and bitterness that develops in highly conflictual disputes, such as psychological harassment cases? Is the promotion of a good working relationship enough to encourage cooperation between parties? In such cases, the parties' perceptions of each other's — and their own — share of responsibility for the conflict can play a key role.

According to Cloven and Roloff (1991), people in conflict try to make sense of their dilemma by first determining who is responsible for the conflict. Naturally, most people blame the other party, which can generate a desire to seek revenge (Beugré 2005); blame is thus a critical step in conflict escalation. Conversely, it makes sense that relinquishing blame by accepting one's share of responsibility for the conflict could be an important step in the resolution of the conflict.

The acceptance of responsibility is a key element of such mediative processes as restorative justice (Umbreit 1994). Victim–offender mediation, which is a popular form of restorative justice, provides for appropriate dialogue between the victim and offender by focusing on the crime's harm rather than the broken rule, and by emphasizing redress for the victim for the effect of wrongdoing (Yarn 1999). Even in small‐claims mediation, accepting responsibility plays an important role. In a qualitative study, Marian Borg (2000) explored the role of responsibility in the concession process. According to Borg's observations, when parties have not resolved issues of blame and responsibility, they find it very difficult to make concessions. In fact, in these situations, parties often believe that their concessions will be perceived as admissions of responsibility or even as expressions of forgiveness. Under such conditions, willingness to cooperate is likely to be nonexistent.

In the mediation context, the process of accepting responsibility comprises two activities: acknowledging one's own share of the responsibility and hearing the other party acknowledge his. Each one plays a different role in the development of cooperation in mediation.

Acceptance of responsibility is related to the development of trust. In a negotiation, a party trusts the other party when he is confident the other party will act in a reciprocal manner (Axelrod 1984). But the other party's refusal to accept his share of responsibility is likely to decrease trust. If the other party is not ready to admit what he has done wrong, why should I believe he wants to correct the situation? Consequently, even if both parties desire to find a solution and acknowledge the importance of reconciliation, without recognition of responsibility by the other party, the level of trust may simply be too low for a party to be willing to take the risk of opening up.

In addition, accepting one's own share of the responsibility may be just as important. If someone is convinced she is in no way to blame for the emergence of the conflict, for maintaining it, or even for the failure of prior attempts to resolve it, it is likely that she will not be particularly disposed to cooperate with the other party. Instead, she will focus on blame, retribution, and compensation in exchange for “forgiveness,” and not on the cooperation needed to find a solution (Peachy 1989). Once a party accepts some responsibility, however, she might be more willing to help find a compromise acceptable to all parties.

Analysis of popular mediation approaches supports the idea that the parties' acceptance of their responsibilities for the conflict plays an important role in mediation. Following this, I set out to analyze this roleempirically, especially in relation to such factors as the will to resolve the issues and the desire to reconcile with the other party.

The questionnaire developed for this study (see the  Appendix) was pretested. It measured one phenomenon (willingness to cooperate, a dependent variable), six factors that potentially affect the initiation of cooperation (independent variables), and one control variable. Table One summarizes the phenomenon, the factors, and the control variables included in the study specifications.

Respondents rated their own willingness to cooperate with the other party during the mediation session according to a six‐point scale: (1) very weak, (2) weak, (3) somewhat weak, (4) somewhat strong, (5) strong, and (6) very strong. Each respondent's initial willingness to resolve the conflict, willingness to reconcile with the other party, and level of acceptance of partial responsibility were rated. I also measured their perception of the other party's initial willingness to resolve the conflict, willingness to reconcile with the other party, and level of acceptance of partial responsibility. Measuring both the respondent's attitude and perception of the other party's attitude enables us to see which of the two has the biggest influence on the emergence of cooperation (i.e., one's own attitude or one's perception of the other's attitude).

In addition, a control variable was measured, that is, the the other party's willingness to cooperate as demonstrated during the mediation session. For this variable, the same six‐point scale was used. It is important to control for this variable because I wanted to go beyond measuring each respondent's reaction to the other party's willingness to cooperate. Including this control variable in our study enabled us to measure the effect of the six factors on the respondent's willingness to cooperate independently of the other party's willingness to cooperate.

Participants were recruited through mediators working in three conflict management bureaus with the Canadian Forces (Department of National Defence, Canada). All were employees involved in workplace disputes. A total of twelve mediators collaborated in the study. To ensure that participants' experience with the mediation process was sufficiently similar, I only worked with mediators who had taken a similar basic mediation course and had at least two years of experience in mediation. Their role was limited to suggesting participation in the study to any new party in mediation and distributing the research questionnaire to willing parties following the first mediation session. In order to avoid bias in selection on the part of the mediators, they systematically offered the chance to participate to all parties involved in a new mediation case. The parties that agreed to participate responded individually to the questionnaire on site, then sealed it into the prepaid envelope provided to preserve the anonymity of their responses.

The response rate of parties was 45 percent. I received seventy‐eight questionnaires and retained seventy‐four for statistical analysis. One questionnaire was rejected because it had been completed by an observer rather than by a party to mediation, while I rejected two questionnaires because more than 20 percent of the answers were incomplete (Hair et al. 2006). I also rejected one questionnaire because the respondent gave the same answer to all questions and the answers thus presented no variance.

The sample's characteristics were: 63 percent men/37 percent women, 40 percent military personnel/60 percent civilians, and 68 percent French speaking/32 percent English speaking. Note that the sample's characteristics do not differ substantially from those of the larger population of all participants in mediation cases occurring during the questionnaire collection period in the three Canadian Forces offices (St Jean sur Richelieu, Valcartier, and Ottawa).

The quantitative analysis of the questionnaires was performed in two steps. First, I used a hierarchical linear regression analysis to evaluate the influence of six factors on the respondent's willingness to cooperate with the other party while controlling for the influence of the other party's cooperative attitude. With the hierarchical regression method, the control variable was input into the equation first (using the “enter” method), while the factors were input into the equation later (using the “forward” method). The factors' influences were thus established while considering the control variable. In other words, this statistical procedure lets us measure how the factors act on the respondent's willingness to cooperate, independent of the influence the other party's attitude had on this willingness.

Following analysis, only those factors that I determined had a significant effect on the phenomenon were included in the regression equation, and the relative effect of each of those variables is indicated by its weight in the equation. Therefore, I can determine both which factors had a significant bearing on willingness to cooperate and their relative importance. Factors not entered into the equation are therefore considered to have no meaningful influence on the willingness to cooperate.

Second, a factorial ANOVA (analysis of variance between groups) design was performed to compare the ramifications of combinations of significant factors on the respondent's willingness to cooperate with the other party. This statistical procedure enables us to easily compare the effect of the combined presence of factors on the degree of willingness to cooperate with the other party during mediation. Such a design is useful for highlighting potential moderating effects.

To perform the factorial design, some questionnaire answers were transformed into binary data (i.e., 0 or 1). So, in each case, the values of the variables were compared to their median value. I used the median because this measure of the central trend is not sensitive to extreme scores and it allows groups that are equivalent as to number to be formed. When the value was higher than or equal to the median, the factor was considered to be present (i.e., 1) while when it was equal to or lower than the median, the phenomenon or the factor was considered to be absent (i.e., 0).

The results of the analyses are presented in two stages. First, those factors having a significant bearing on willingness to cooperate are identified using a hierarchical regression. Second, the interaction effects of key factors are highlighted with a factorial ANOVA design.

Using hierarchical regression analysis enabled us to define four factors that significantly predict willingness to cooperate. In order of importance, according to their weight in the regression equation (β) as well as their probability of not being statistically significant (P), the four factors are:

• willingness to reconcile (β = 0.388, P < 0.01),

• willingness to resolve (β = 0.295, P < 0.01),

• respondent's acceptance of partial responsibility (β = 0.231, P = 0.01), and

• the other party's acceptance of partial responsibility (β = −0.172, P < 0.05). (Note that here there was a negative correlation.)

These factors thus contribute to the emergence of cooperation during mediation. Statistically, these four factors enable us to explain 67 percent of the variance in willingness to cooperate (adjusted coefficient of determination R² = 0.666). On the basis of this percentage, I can conclude that the model has practical significance.

Two other factors in the study were not included in the regression equation as their contribution was not meaningful: manifestation by the other party of a willingness to reconcile with the respondent and manifestation of one's own willingness to reconcile. These findings suggest that willingness to cooperate is mostly centered on respondents' desires more so than the other party's manifestations of intent. Table Two reports the principal results of the hierarchical linear regression.

Among the significant factors, three factors in our study demonstrated a positive influence on the emergence of the desire to cooperate: the willingness to come to a resolution, to reconcile with the other, and to accept one's share of responsibility (see Figure One). On the other hand, one factor has a negative impact. Interestingly — as I stated earlier — the other party's admission of a share in the responsibility seems to hinder respondents' willingness to collaborate (regression weight β = −0.172; probability of not being significant P > 0.05).

At first glance, this result may seem strange and counterintuitive. The other party's acknowledgment of some responsibility should, in theory, encourage willingness to cooperate. This result points to the potential role of other party's admission as a moderator between respondents' acceptance of responsibility and their willingness to cooperate.

To evaluate the interaction effect between parties' acceptance of responsibility, I ran a factorial ANOVA design. In the model, respondents' willingness to cooperate was entered as a dependent variable, willingness to resolve conflict and willingness to reconcile with the other party were entered as covariates, and respondents' acceptance of responsibility and the other's acceptance of responsibility were entered as fixed factors. The corrected model is statistically significant (value for significance F = 3.21; probability of not being significant P < 0.01) as well as practically significant (adjusted coefficient of determination R² = 0.598).

As predicted, a respondent's acceptance of responsibility had a significant effect on his willingness to cooperate (value for significance F = 10.639; probability of not being significant P < 0.01). Thus, I conclude that one's willingness to acknowledge one's own share of responsibility is an important factor in one's decision to cooperate. But the impact of the other party's acceptance of responsibility was not significant (value for significance F = 0.001; probability of not being significant P > 0.05), which means that the other's acceptance does not play a direct role in a respondent's willingness to cooperate. Nevertheless, there is a significant interaction effect between one's acceptance of responsibility and the other's recognition of his responsibility (value for significance F = 4.934; probability of not being significant P < 0.03), confirming a moderating effect.Figure Two highlights the interaction effect.

The figure illustrates two particularly interesting conditions. The first occurs when the other party recognizes her share of responsibility but the respondent does not. In this case, the respondent's willingness to cooperate is dramatically lower than when no one accepts a share of the responsibility. Put simply, it seems that if the other party acknowledges her share of the responsibility but I do not acknowledge my own share, then I will then harden my position and become less inclined to cooperate. The logic seems to work like this: it's her fault, so she pays!

The second interesting condition illustrated in Figure Two occurs when both parties recognize their share of responsibility. In such cases, the respondent's willingness to cooperate is significantly higher than when he is the only one to acknowledge responsibility. Clearly, having both parties accept their share of responsibility has a dramatic impact on willingness to cooperate. I could illustrate the logic as follows: we are both a part of the problem, so we must both take part in finding a solution.

Clearly, recognition of responsibility matters to both parties. In this study, I analyzed the data from the perspective of the respondent. But the same logic with regard to recognition of responsibility can be expected to work for the other party. When the respondent accepts his share of responsibility but the other does not, I hypothesize that the other will be less inclined to cooperate. Joint recognition appears to avoid this trap.

Thus, I suggest these general rules:

• if only one party accepts his share of the responsibility, then the situation is worse than if no one accepts his share, and

• when both parties accept their share of the responsibility, then the situation is much better than if no one accepts his share.

This leads us to the paradox involved in accepting responsibility in mediation: unilateral recognition makes things worse, joint recognition make things better.

This paradox underscores one of mediators' great fears: beginning a dialogue on responsibility entails a risk of crystallizing the conflict. The instinct of many mediators is therefore to avoid talking about this delicate issue in mediation, and this may be a good strategy when only one party is willing to accept a share of the responsibility. But the findings of this study indicate that this strategy deprives mediators of an important lever for collaboration: joint recognition of responsibility. Therefore, the solution is not to avoid the subject of responsibility, but to foster a joint recognition by the parties of their respective share of responsibility for the conflict.

The results of this study suggest two intervention principles for mediators. First, if I accept that three factors play a role in the emergence of cooperation (willingness to resolve the conflict, desire to reconcile, and acceptance of responsibility) then it makes sense that the mediator should focus on all three. In fact, the full potential that accepting one's share of responsibility holds for a successful mediation depends on the parties' willingness to resolve the conflict and to reconcile with each other. Consequently, prior to broaching the subject of responsibility, the mediator may first want to ensure that the parties are actually interested in both resolution and reconciliation.

The second principle is to encourage joint acceptance of responsibility. When only one party acknowledges his errors, the other might become antagonistic. Where both parties accept their errors, however, the probability that collaboration will emerge is significantly higher. Because recognition of one's responsibility is a double‐edged sword, mediators need to avoid one‐sided acceptance of responsibility and need to encourage the parties to mutually recognize their errors.

The mediator who seeks to encourage both parties to acknowledge their responsibilities should first consider two questions. First, what type of acceptance should the mediator aim for? A party can acknowledge responsibility to various degrees. The first degree includes accepting how one's own thoughts, attitudes, statements, actions, and lack of relationship, communication, and behavioral skills could have contributed to the conflict (Cloke and Goldsmith 2000). The second degree of acceptance involves actually acknowledging responsibility and agreeing to apologize under some circumstances (Borisoff and Victor 1989). While the second is better, the first is probably a more realistic goal in the context of mediation. Furthermore, one recent study has shown that acknowledging one's intentional faults is just as effective as recognizing one's unintentional faults (Bottom et al. 2002), which suggests that simply acknowledging one's role in the escalation of the conflict — even if one did not intend this escalation — can be helpful.

The second issue for the mediator is technical: should acknowledgments of responsibility be handled in separate caucus with each party or in a joint meeting? The advantage of proceeding by caucus is that it avoids the risk that one party will acknowledge responsibility publicly and unilaterally. The caucus also provides a more favorable framework for exploring one's role in a conflict because it allows the parties to avoid losing face in front of the other. But because these study results indicate that mutual acceptance of responsibility is more more likely to foster a climate of cooperation, an eventual joint discussion might also be desirable. Ultimately, to be effective, the mediator should consider combining these strategies.

To foster the acknowledgment of responsibility, I propose a three‐step intervention strategy. The first step involves validating the parties' willingness to resolve the conflict and to improve their relationship; this can initially occur in caucus. With both parties' consent, the mediator can then announce these joint goals in a joint session. The parties should then confirm that this is an accurate summary of the situation — an important step, because fostering parties' acceptance of their share of responsibility succeeds as a mediation technique when these two other desires are already present.

In the second step, the mediator should explore with each of the parties individually their respective roles in creating the conflict. The mediator should, of course, not seek to blame the parties by trying to impute motives. Rather, she should explore with them their errors in managing the conflict, some of which may have been inadvertent and even well intentioned. A useful question to ask at this stage is: “If you had to do it again, knowing how the conflict escalated, would you handle things differently to resolve the dispute?” Acknowledging that one could have acted differently to avoid escalating the conflict is a form of acceptance of responsibility. Clearly, in some cases, one party played a bigger role in the conflict than did the other, and it could be important for that party to recognize and accept that. Nonetheless, the mediator's goal here is to make sure the two parties each acknowledge at least some responsibility for escalating the conflict.

The third step involves having the mediator summarize, in a joint session and with both parties' consent, the acknowledgment of their individual shares of responsibility in accordance with the conclusions reached in caucus. The mediator then asks the parties to confirm whether she has summarized the situation accurately.

Using this three‐step method should help facilitate the process of joint recognition of responsibility. There are some situations, however, for which using the proposed mediation strategy would be clearly inappropriate. When the bahavior of one party was abusive, discriminatory, rule breaking, or even illegal, it would be immoral for the mediator to reframe the conflict as a joint responsibility. But for many cases, the method can be used to foster cooperation among parties.

For most mediators, the assignment of responsibility for a conflict is a diffcult issue. They fear that the process of determining and acknowledging responsibility will encourage confrontation and mutual accusations and lead directly to impasse. But the results of this study highlight the important role within mediation that accepting one's share of responsibility can play. Specifically, when parties jointly acknowledge their errors, collaboration is more likely to emerge. To facilitate this process, I proposed a three‐step strategy for mediators.

This particular phenomenon was studied in the context of workplace conflict in which the parties can be expected to have an ongoing relationship and a desire to reconcile. It is unlikely that acceptance of responsibility plays as important a role in all mediation situations. For example, when parties will not work together again and have no wish or need to reconcile, a problem‐solving approach may be enough. And it would certainly be worthwhile to test the significance of accepting one's share of responsibility in other mediation contexts, such as family mediation.

Notably, this study also did not take role into consideration. Testing for the impact of role (complainant or defendant, supervisor or supervisee, etc.) would certainly also be worthwhile. Nevertheless, what this study's results indicate is that trying to foster the parties' joint recognition of their responsibilities for the conflict is a mediator strategy worth serious consideration, particularly in those mediations in which one goal is to transform the relationship.

One of this study's most interesting findings is paradoxical, that accepting one's share of responsibility is a double‐edged sword. When only one party acknowledges his errors, collaboration can actually be impeded. When both parties recognize their errors, it is facilitated. The difficulty that then emerges from this analysis is this: how does the mediator create a climate of mutual acceptance of responsibility? Who begins? Will the other party also recognize his responsibility?

I suggest that a mediator can effectively act as a bridge between parties to foster mutual recognition of their roles in creating the conflict. Further research could test the proposed three‐step process and examine more precisely the mediator's role in facilitating this recognition, which strategies work, and which do not.

Phenomenon

• –

  How strong was your desire to cooperate with the other party during the mediation session?

Control variable

• –

  How strong was the other party's cooperative attitude during the mediation session?

Respondent attitude

• –

  How strong was your initial desire to find a solution to this conflict?

• –

  How strong was your desire to maintain, reestablish or improve your relationship with the other party?

• –

  How strong was your willingness to acknowledge and assume your share of responsibility in this conflict?

Respondent perception of the other's attitude

• –

  How strong was the other party's initial desire to find a solution to the conflict?

• –

  How strong was the other party's desire to maintain, reestablish or improve his/her relationship with you?

• –

  How strong was the other party's willingness to acknowledge and assume his/her share of responsibility in this conflict?