When Clients Throw Punches and Chairs: How Mediators Respond to Violence

Eruptions of physical violence during mediation or immediately after are uncommon, but when violence occurs it presents practical, ethical, and logistical challenges for mediators. Our survey asked mediators to describe their experiences with physical violence and threats of violence, including hitting, slapping, punching, pushing, throwing objects, and choking (Wynne et al. 1993). We did not ask about the use of aggressive or hateful language in order to keep the focus on the use of (or explicit threats to use) physical violence in mediation.

Any violence directed specifically toward mediators—instead of or in addition to violence between disputants—would meet the definition of “workplace violence.” Violence against judges and attorneys has been documented by the American Bar Association and other organizations (Tebo 2005), but we found no similar data examining violence against mediators or between parties during mediation sessions. Although some mediations may span multiple sessions, we did not specifically look at violence that might occur between sessions, because we focus here on the reactions of mediators to those incidents that occur in their presence.

Occasionally, work‐related violence against judges makes national news (King 1982; McFadden 1988; Associated Press 1989; Roh 2005; Kiefer and Bland 2011). Although violence against judges and attorneys is rare, numerous mechanisms are in place to help prevent it and to intervene when it occurs. Courts typically have metal detectors, law enforcement personnel, and even panic buttons in place designed to help protect both court personnel and the public. Continuing Legal Education seminars (in most states attorneys are required to take such courses to maintain their licenses) sometimes focus on screening cases for violence and the ways in which attorneys can de‐escalate difficult situations. Similarly, several articles have been published in the field of social work on how practitioners can prevent violence and deal with it effectively when it occurs (Newhill 1995; Rey 1996; Mayhew 2000; Criss 2010; Pollack 2010; Padyab et al. 2012).

Although mediation is an ancient practice, in its modern Western iterations it is newer than the practice of law or adjudication. Informal mediation is practiced every day by parents, teachers, managers, and government officials, among others. But in this article, we focus specifically on the formal mediation that occurs in courthouses, attorney's offices, and community mediation centers. For our purposes here, mediation can be described as a process in which a neutral third party facilitates a problem‐solving discussion between the parties designed to resolve a legal or non‐legal dispute. The mediators in our study include paid and volunteer mediators, but all have been trained and meet the requirements of mediation programs in the U.S. jurisdictions in which they practice.

Mediation scholars have not systematically examined how often violence erupts in mediation nor how to respond in ways that reduce the chances of injuries to parties, attorneys, and mediators. With this study, we seek to begin to fill that gap. We asked 112 mediators of varying levels of experience who work on a broad array of cases to describe their practical experiences dealing with violence in mediation. Our goal is to learn from their experiences and build a body of research about the best practices for preventing violence in mediation and for enhancing the safety and well‐being of practitioners and the people they serve.

Until recently, violence in mediation has not gained much attention from alternative dispute resolution (ADR) scholars. The scant research on violence in mediation has focused primarily on two issues: whether criminal cases of domestic violence should be mediated and whether family law cases in which violence has occurred were appropriate for mediation (Cobb 1997; Krieger 2001). Several articles have focused on how to screen domestic violence cases for possible mediation (Zylstra 2001; Ellis and Stuckless 2006; Ballard et al. 2011; Michigan Courts 2014). We found no studies that examined the occurrence of violence against mediators and/or participants during or following mediation sessions nor how mediators respond to such violence.

Across mediation programs, jurisdictions, and case types, the policies for preventing and responding to violence differ widely. There has been no systematic research to indicate which strategies are most effective, how much different strategies cost, nor which strategies work best in specific scenarios. Likewise, the procedures for pre‐screening cases for past violence vary widely within and between mediation programs, with no nationwide standard widely accepted or used in the United States (Beck and Frost 2007). At best, some states have instituted best practices or requirements regarding case screening and intake processes, but these do not apply to other states or even across case types (Clemants and Gross 2007).

Therefore, in this study, we seek to address that deficit as well as to answer two other related questions that have been comparatively ignored in previous research: how common is the outbreak of violence during mediation sessions, and how do mediators react once violence occurs? We hope that this study contributes to a deeper understanding about what violence in mediation looks like as well as common mediator strategies to address violence when it occurs.

For our purposes, “mediation violence” includes physical violence, or threats of such violence, against people or property that occurs during a mediation session or immediately afterward (i.e., as parties leave the mediation). Because tempers are often high during mediation, we do not include general verbal aggression such as name calling or incivility in our definition of violence unless such outbursts can reasonably be construed as specific threats of physical harm.

Although violence in mediation is a relatively new area of study, violence has been studied in the related fields of law and jurisprudence. Violence against attorneys and judges is, sadly, not uncommon (Wax 1994; Hansen 1998; Kelson 1999; Tebo 2005; Brown and MacAlister 2006). In one survey, nearly 60 percent of the members of the American Bar Association (ABA) section of Family Law reported experiencing threats by opposing parties, and 17 percent reported receiving threats from their own clients (Hansen 1998: 93). In another study, nearly 60 percent of Canadian lawyers surveyed reported receiving “varying number and degrees of violence and/or threat” (Brown and MacAlister 2006: 556). In addition, several studies have identified a trend toward increased violence against judges and attorneys (Kelson 1999; Mayhew 2000; Kelson 2000; Jayaratne, Croxton, and Mattison 2004; Pollack 2010; Hyde 2012).

Mediators and their clients have at least occasionally been the targets of violence. In 2013, a plaintiff shot and killed both a mediator and the defendant immediately following mediation, and later killed himself (Kotev 2013). Also, in 2008, a man beat his wife to death after a court‐mandated mediation session and then killed himself (Johnson 2008).

Mediation is common in the types of cases that often give rise to strong emotions, such as divorce and custody cases, as well as probate matters, neighborhood conflicts, and even roommate disputes (Zylstra 2001; Ellis and Stuckless 2006; Ballard et al. 2011; Michigan Courts 2014; Association for Conflict Resolution [ACR] 2014), and the presence of strong emotions presumably increases the risk of violence. (We would also expect greater risks of violence for mediators of violent international conflicts, civil wars, or those who work with notably violent populations, such as prisoners. Our sample is more fully described below, but we acknowledge here that we did not examine mediation in these particular contexts.)

The mediation community has become more aware of the potential for violence in mediation (Mayhew 2000; Poitras and Raines 2013; Association for Conflict Resolution 2014). Mediation trainings across the United States typically focus on the de‐escalation of destructive emotions in order to prevent negative behaviors as well as impasse (Hedeen, Raines, and Barton 2010). Claire Mayhew (2000) discussed how to prevent client‐initiated violence in a series examining workplace violence across industries and sectors, including mediation, but her recommendations are brief (only two pages long) and very general (be prepared, be flexible, use positive language).

Although Jean Poitras and Susan Raines (2013) did not specifically address violence in mediation, they did devote a chapter of their book to how mediators can manage emotions in mediation because strong emotions can be a particular source of impasse and conflict escalation. Poitras and Raines (2013) surveyed 206 expert mediators and identified the following strategies for dealing with intense emotion in mediation:

• making space for emotional expression by using caucuses to continue mediating with the parties in separate rooms, acknowledging and normalizing the existence of strong emotions, taking breaks, and reminding parties they are in control of the outcome;

• helping parties understand each other's emotional concerns;

• identifying the underlying needs and interests that give rise to strong emotion;

• coaching parties to manage and express emotions constructively; and

• helping parties to focus on a specific goal(s) in mediation such as creating a better relationship between the parties going forward.

A Taskforce on Safety in ADR created by the Association for Conflict Resolution (ACR) seeks to promote the safety of practitioners and participants. The taskforce guidelines (2014) are the only professional guidelines published in the United States designed to prevent and address violence in and around mediations. They include step‐by‐step protocols that ADR practitioners can follow before, during, and after violence in mediation occurs. They suggest general safety protocols and the careful selection of site and meeting locations with safety in mind. They list screening procedures to assess the potential for violence. Also, they include step‐by‐step protocols that ADR practitioners and organizations can follow after violence occurs in mediation. The guidelines can be found on the ACR website (acrnet.org).

According to these guidelines, the safety of mediators and participants can be secured from the organizational level to the individual mediator level. To ensure mediators' safety, the guidelines recommend that both organizations and mediators must establish safety plans and review them periodically (ACR 2014). Organizations and courts can improve the safety of mediators by enhancing overall building security, creating a “no weapons” policy, and developing notification procedures with others for immediate assistance (ACR 2014). On the other hand, the guidelines assert, mediators need to identify exit plans and ask for support from security personnel in case they meet with potentially violent individuals (ACR 2014).

Many mediations occur outside the security of the courthouse—in offices, community centers, library meeting rooms, schools, and elsewhere. When mediators meet with parties in their personal office space or an attorney's office, they can remove such potential weapons as “pens, pencils, letter openers, staplers, picture frames, corkboards, lamps, light furniture, and chairs and any other office equipment from view” (ACR 2014: 6). Additionally, for their own protection, mediators should remove from the office anything that reveals personal information, such as home addresses and personal photographs (ACR 2014).

In such related fields as health and social services, practitioners commonly profile clients to ensure their safety, particularly in regard to risk factors such as addiction or criminal records (Newhill 1992; Rey 1996; Lowe 2011). Such practices raise difficult ethical questions for mediators, however. No single trait can predict violence will occur, but several factors are known to make it more likely (ACR 2014). For example, there are higher risks associated with mental illness such as paranoid delusions, command hallucinations, and syndromes. Other characteristics associated with a greater potential for violence include histories of arrest, violent, or impulsive behavior; experience of abuse; witnessing violence; substance abuse; and use, possession, or fascination with weapons (Todahl and Walters 2009; ACR 2014; Michigan Courts 2014). In cases involving domestic violence, questionnaires and interviews are sometimes used to screen for these characteristics prior to mediation (VerSteegh 2003; Landrum 2011; Michigan Courts 2014).

The literature on domestic violence includes published guidelines with recommendations about how to conduct mediation to keep participants safe (ACR 2014; Michigan Courts 2014). When participants display signs of danger, mediators should favor the safety of all involved over resolving the dispute (ACR 2014). According to ACR guidelines (2014), when violence occurs mediators must decide whether they can stop the violence and, if they cannot, they should avoid physical confrontation with a violent participant, exit the room, ask for assistance, and call the police and emergency services as necessary (ACR 2014). Also, mediators must “gain the attention of the participants,” ensure the safety of the parties by escorting the threatened participants out, “inform security personnel or colleagues present of possible violent participants, terminate the session, and make sure that the participants stay separated during or even after mediation” (ACR 2014: 11).

Although these guidelines make an effort to address violence in mediation, they cannot possibly be suitable for every situation because each mediation practice and program is unique (ACR 2014). Also, these protocols are much less fully developed than those in other related fields, which prescribe, for example, ongoing screening methods, de‐escalation strategies, more specific practitioner responses to violence, and training strategies. Therefore, in this article, we also draw on the literature from related fields.

Research from other fields has provided extensive insights into how violence against service providers in general (i.e., health care providers, social workers, attorneys, etc.) arises as well as into its significant consequences. According to a handbook on preventing client‐initiated violence prepared by the Australian Institute of Criminology, typical client‐initiated violence “escalates through inadequate design of the environment and poor service provision,” and “[S]ome arises through anxiety, inadequate coping skills, or illness in the client, and some through pure malicious intent” (Mayhew 2000: 5).

Violence toward practitioners manifests itself not only through face‐to‐face communication, but can also include stalking and threats issued via telephone calls, e‐mail, and letters (Mayhew 2000). Literature from a variety of professions describes the consequences that this kind of violence has had on practitioners. Mojgan Padyab and her colleagues (2012: 144) found that witnessing violence can have such negative health impacts as “physical symptoms, anxiety and sleep disorders, and social dysfunction” and diminished productivity. For organizations, violence can lead to difficulties in recruiting valued staff (Mayhew 2000; Tebo 2005; Padyab et al. 2012). Multiple studies of workplace violence against social workers (Rey 1996; Criss 2010; Padyab et al. 2012; Padyab et al. 2013) and judges (Mayhew 2000; Tebo 2005) have linked fear of violence to job burnout and stress. One study reported that “Some judges, concerned for their families’ safety, may opt for early retirement,” and “well‐qualified lawyers may think twice about accepting a federal judicial appointment” (Tebo 2005: 44).

The research on client violence toward such professionals as attorneys reveals little about which measures are likely to be the most successful at reducing the risk of harm. As “Human beings are fundamentally unpredictable, [nobody] can [fully] predict, assess, and sort out the high risk from the low risk” (Klaw 2013: 39). Although legal scholars and practitioners acknowledge the unpredictability of the risks from clients, several have argued on behalf of preventive measures that can make both judges and attorneys safer (Wax 1994; Kelson 1999, 2000; Mayhew 2000; Jayaratne, Croxton, and Mattison 2004; Tebo 2005; Pollack 2010; Hyde 2012).

Like the ACR guidelines, which were drawn in two levels, organizational and individual, the literature on violence prevention in related fields generally identifies “two avenues of prevention: one deals with organizational factors and the other [deals] with worker training” (Jayaratne, Croxton, and Mattison 2004: 452).

As noted above, guidelines for domestic violence mediation caution that mediators need to recognize the possibility of violence by continuously screening participants (Michigan Courts 2014). Although ACR guidelines urge ADR practitioners to focus on safety when they observe signs of danger, these guidelines do not identify the warning signs of potential violence. Literature from related fields, however, does specifically identify such signs. In Facing Physical Violence, British social psychologist Glynis Breakwell (1989) recommended that therapists be alert for such warning signs as hypersensitivity, rapid mood swings, signs of atypical excitement or arousal, passivity, and unusual patterns of nonverbal communication. Experts have recommended that nurses be alert to such signs of potential violence as stamping, staring, anxiety, mumbling, and pacing, and that they be particularly alert to such signs of anxiety as flushed appearance, hyperventilation, rapid speech, dilated pupils, physical indicators of pain, repetition of the same or similar questions or requests, and slurring or incoherent speech (Luck, Jackson, and Usher 2007).

Advice to practitioners in other fields has also focused on how to prevent violence as well as how to recognize its warning signs. Stephen Kelson (1999) argued that a healing or “client‐centered” approach can prevent violence against lawyers. He wrote that when lawyers “involve the client in the process and in exploring new solutions; encourage the client to make substantial decisions; provide advice based on a client's values; acknowledge the client's feelings; recognize their importance; [and] repeatedly convey a desire to help,” it enables clients to deal with their own problems and helps prevent violence against lawyers (207–208).

Existing studies on social work (Griffin 1993; Rey 1996; Pollack 2010) and legal practice (Kelson 1999) have also discussed strategies that practitioners can use when confronted with violence. These include expressing empathy for participants’ feeling of pain and rage (Rey 1996); “remaining confident; continuing to talk in a normal tone of voice; using nonverbal communication to calm the situation; using diversionary tactics; feigning submission; redirecting attention; maintaining distance from violent participants; removing potential weapons from view if possible; removing individuals inciting the violent participants; asking for assistance from bystanders” (Kelson 1999: 219); and “using gentle persuasion” (Pollack 2010: 278).

Perceptions about what constitutes an unacceptable level of verbal or physical violence differ according to culture and even personality. What one mediator finds threatening, another mediator might not. We did not, in this study, specifically investigate varying interpretations of what constitutes violence. Instead, we explored the various responses mediators have commonly used when they have identified what they personally believe are unacceptable levels of aggression or violence.

As mentioned above, ACR guidelines currently suggest fairly obvious measures, such as exiting the room and asking for assistance. The simplicity of the advice speaks to the nascent nature of work on this issue and the lack of specific data and research gathered from mediators to better understand which strategies work best. Although acknowledging that such a response is far from ideal, some authors (Kelson 1999; Pollack 2010) have discussed the possibility of using force for self‐protection. Christina Newhill (1995) wrote that when a person's safety is at stake, it is acceptable to use a safe physical restraint while being careful to avoid areas prone to injury or areas that could interfere with breathing such as the throat and abdomen.

An extensive amount of literature on social work (Tully, Kropf, and Price 1993; Newhill 1995; Rey 1996; Criss 2010) and legal practice (Kelson 1999; Tebo 2005) discusses violence prevention training. Some jurisdictions and law schools offer security training for new judges (Tebo 2005), and some agencies train social workers in how to avoid client violence (Criss 2010).

Comprehensive programs to train social workers about client violence cover the following topics:

• theory of causes of client violence;

• extent of the problem;

• how to assess the likelihood that a client will become violent;

• becoming aware of one's own reactions and self‐control;

• techniques for verbal de‐escalation of violence;

• physical control;

• designing safe office space, including waiting area, reception desk, and security;

• safety procedures for working with unknown or suspicious clients;

• debriefing and support of traumatized workers in order to prevent post‐traumatic stress disorder; and

• agency policy for preventing, managing, and follow‐up with workplace violence (Rey 1996).

Not surprisingly, these are also common features of violence training programs for lawyers as well.

The techniques recommended for other professionals may or may not be appropriate for mediators for a variety of reasons. We discuss these in detail in the discussion section of the paper. In our study, we sought to gain insight into the specific impact of client violence on the practice of mediation, how commonly it occurs, and common mediator responses to that violence.

In late 2012 through June 2015, we invited mediators to participate in a study on the topic of violence in mediation. We built our sample from three groups: attendees at the national conferences of the Association for Conflict Resolution held in Minneapolis in 2013 and in Cincinnati in 2014, registered court‐connected mediators in Georgia, and members of the New York mediators’ online listserv discussion group. Mediators from the following states are represented in our sample as shown in Table One.

Approximately 46 percent of the participants were male; 52 percent were female; and 3 percent declined to state their gender. We conducted the survey using the Internet survey service Surveymonkey.com and recruited participants by distributing a handout at the ACR conferences, sending e‐mail messages to the Georgia mediators, and posting a message on the New York listserv.

This is an exploratory qualitative study, and because of its relatively small sample size and the non‐random selection methods that we used, we do not seek to generalize our findings to the entire population of U.S. mediators or mediations. In addition, because all mediators in our sample live in the United States, our results may not easily be generalized to other countries and cultures. We believe that the mediator stories that we gathered, however, provide important insights about how and when violence arises in mediation, and allow us to begin to build a list of strategies mediators use to prevent, reduce, and manage violence in mediation.¹

Previous research indicates that mediator experience correlates to settlement rates and party satisfaction: mediators who have mediated fifty or more cases have higher rates of both settlement and party satisfaction than do mediators who have mediated fewer than fifty cases (Poitras and Raines 2013). Mediators who have completed one hundred or more mediations have the highest settlement and satisfaction rates compared to inexperienced mediators (those who have mediated fewer than fifty completed cases) (Poitras and Raines 2013).²

Survey participants’ experience levels varied, with more than one quarter (29 percent) of participants reporting they had completed more than five hundred cases, which is considered to be highly experienced. Thirty percent of respondents reported that they had mediated fifty or fewer cases (Table Two).

We also gathered information about how long each mediator had been practicing. Years of practice, however, may not necessarily correlate with experience levels because some mediators may only mediate occasionally, such as full‐time attorneys or counselors who mediate only a few times a year, while other mediators may have been in practice only a few years but have completed hundreds of mediations. As shown in Table Three, 64 percent of the mediators in our study indicated they have practiced mediation for six or more years.

We also asked the participants about the types of cases they mediate. Previous mediation research has indicated that the type of mediation may be correlated with the likelihood of violence: concerns about violence and safety arise more frequently in family law cases (Cobb 1997; Zylstra 2001; Krieger 2001; Ellis and Stuckless 2006; Ballard et al. 2011; Michigan Courts 2014).

As shown in Table Four, our study included mediators from a diverse range of case types. We know that some mediators practice in multiple areas of practice. We asked them, “What type of case do you mediate most often?”

As previously noted, our results may not be generalizable to mediation in general. We did, however, find that family mediators, who made up a smaller percentage of our sample than commercial/civil mediators, did report experiencing more violent cases than did mediators in other practice areas (Table Five).

We did not gather information about participants’ genders, races, or incomes. We also did not ask participants whether they mediated in secure (e.g., courthouses) versus non‐secure settings (e.g., attorney's offices, community centers, schools), or whether they mediated primarily court‐ordered cases or cases outside of the court system. Such details could provide important insights for future research.

In our survey, we asked the following questions. Possible answers were: yes, no, or unsure/can't recall.

• Have you witnessed violence during a mediation session?

• If yes, how many times have you witnessed violence in mediation?

• Did you ever fear for the safety of the parties as they left the mediation location?

• Did you ever see a party throw something at the other party?

• Did you ever see a party punch a wall or vent violently without hurting anyone?

• Have you ever mediated with parties who had a history of physical violence against one another?

• Have you ever been blocked from leaving the room by one or more parties?

• Have you ever stood between a violent party and another person?

• Have you ever mediated a case in which you were threatened as the mediator?

• Have you ever mediated between parties who brought weapons to mediation?

• Have you experienced violence from an attorney against you or a party?

• Have you been physically punched, struck, or slapped during mediation?

• Have you accompanied a threatened party to a shelter or other safe place?

• Have you called law enforcement to report a threat against a party?

• Have you reported abusive or violent behavior to the judge in the case you were mediating?

• Have you wanted to report actual or threatened violence but were unable to do so due to rules of confidentiality?

• Have you mediated with a party who disclosed s/he was bipolar or schizophrenic?

• Have you mediated between parties with an active protective order?

We also asked a series of open‐ended questions without any length limitations. These were designed to elicit mediators’ stories about the violence they have experienced in mediation:

• If you have witnessed violence in mediation, please give an example or story to illustrate an episode that occurred (being sure to avoid any identifying information that could jeopardize party confidentiality).

• If you were threatened as the mediator, please provide a story to illustrate what happened.

• Please share any stories or problems related to violence you have encountered in mediation.

• What strategies do you use to diffuse anger or deal with potentially violent situations?

• Please share any ideas for improving the safety of mediators during the mediation process.

• Please share any ideas for improving the safety of parties during the mediation process.

None of the mediators in our sample reported being the victim of physical violence as a mediator; none were struck, punched, or slapped during mediation. Some did, however, report that they took affirmative steps to prevent violence that seemed imminent (such as standing between parties during potentially violent episodes). As shown in Table Six, more than one‐third (38 mediators, 34 percent of respondents) reported witnessing violence in mediation. (Unless otherwise indicated the total number of responses was 112, i.e. n = 112.)

More than one‐third of the respondents (38, 34 percent,) also stated they had mediated cases in which the parties threatened violence against each other during the mediation. The same number of respondents (38, 34 percent, n = 111) reported that they had mediated a case in which they feared for the parties’ safety as they left mediation.

As shown in Table Six, the most common types of violence that respondents have witnessed in mediation have been parties punching a wall, throwing a chair, or otherwise venting violently without actually hurting anyone; twenty‐seven mediators (24 percent, n = 111) reported this. Nineteen respondents (17 percent, n = 110) reported that they had stood between fighting parties.

Our respondents were less likely to experience threats to themselves than they were to witness parties threaten each other, but a small number (4, 4 percent, n = 110) did receive oral or written threats against themselves and family members, and an even smaller number (2, 2 percent, n = 110) reported that they had been blocked by a party from leaving the mediation room.

Mediators reported that whether to report violent outbursts and threats can present a difficult choice because they must balance concerns about safety with the need to protect client confidentiality. Rules on reporting vary between states and sometimes within them. Ten participants (9 percent, n = 110) responded that they have called the authorities to report actual or threatened violence in mediation: eight (7 percent, n = 111) of them reported a threat to child protective services when they perceived that a child was at risk, and two mediators (2 percent, n = 108) reported a threat to themselves. Interestingly, five (5 percent, n = 108) respondents indicated they wanted to report threats or violence to the judge or authorities, but felt unable to do so because of confidentiality or ethics rules.

We also asked our participants about whether they had mediated cases that featured characteristics generally considered to pose the greatest risk of violence (e.g., parties with active restraining orders, histories of physical abuse or mental illness, etc.). The responses appear in Table Seven. Half the respondents (55, 50 percent, n = 111) reported that they had mediated between parties with a history of violence, although clearly all those mediations did not actually become violent. A third of respondents (37, 33 percent, n = 110) reported that they had mediated cases in which at least one of the parties disclosed that she or he suffered from schizophrenia or bipolar disorder. In addition, 32 respondents (29 percent, n = 110) had mediated between parties with an active protection or restraining order in place. In most instances, such an order does not preclude court‐ordered mediation, although some safety precautions may be required such as the use of separate caucus sessions. In cases with a history of violence and/or with a protective order in place, courts often require that mediation take place in secure settings such as courthouses, where there are usually metal detectors and security personnel.

Finally, nine respondents (8 percent, n = 111) reported that they have worked with parties who brought weapons to mediation. (Some of these mediators mentioned the importance of mediating in the courthouse, where metal detectors can screen for weapons.) Interestingly, one mediator stated it is his standard practice to ask parties if they came armed. If they answer affirmatively, he asks them to take their weapons out to their cars and then come back into mediation. With gun laws varying so much by state, it is likely that a significant percentage of parties carry weapons on their persons or in their cars in some jurisdictions. This need not indicate intent to do harm, only the availability of weapons.

The sources of violence in mediation are more diverse than one might predict. The literature indicates cases with familial relationships are prone to higher levels of conflict and violence, such as divorce and disputes over inheritance (ACR 2014). In this study, however, several mediators described situations that arose in non–family law cases, including workplace, landlord‐tenant, real estate, and other kinds of disputes. We asked mediators to describe the types of cases in which they experienced or feared violence. Some brief examples include the following:

“There was a civil case over property damage in which both parties were seeing the same man. Parties made threats of violence against each other and the deputy had to be called,” one mediator reported. Another wrote, “It was a neighborhood dispute and one woman reached over the table and pulled hair (a weave) out of the other woman.” Still another told us, “Plaintiff and defendant entered the room and you could tell there was tension. After a few minutes into the mediation, they both stood up and came across the table at each other. It broke out into a fist fight. It was during the lunch hour and the start of a holiday weekend. One of the assistants ran out to get help and the only person she could find was a judge. He had them both arrested to spend the entire weekend in jail.”

One mediator wrote of “several instances involving an injured worker threatening his employer in general terms for causing serious injury and not wanting to pay.” Another wrote that “The State Board of Workers Comp had to install a metal detector to prevent parties from taking weapons into mediation or hearings.”

One mediator wrote that a “white landlord started using racial slurs towards his black tenant. [The] landlord then proceeded to lean across [the] table and told [the] tenant to punch him (the landlord) in the face. I ended the mediation before the situation escalated.”

In one case a “plaintiff [was] suing for alleged non‐payment of pay checks,” the mediator wrote, and his “wife attempted to strike [the] former employer.”

A mediator described the following incident: “Attorneys for [the] parties [were] advised of serious threats by each side towards the other. I asked parties if they were in possession of any weapons in mediation. I asked them to place the weapons in the trunks of their cars and return. All parties (six) got up and left the mediation room and went to their cars, opened their trunks and placed something into the trunks (I could not see what) and returned without incident. Four and a half hours later after much venting (in caucus), they were able to begin resolving their dispute. In my experience, probate cases are often very difficult as the parties revert to issues they had as children with each other or with their parents and forgiveness is not easily reached.”

“There was a case where the husband kept referring to a football game they attended a few years previous,” a mediator wrote. “Every time he mentioned it she became quiet and introverted. In caucus it came out that she was cheering the opposing team to the husband's favorite. Each time the other team scored she teased him. He finally punched her in in the face when she teased him one too many times and broke her jaw. Just by recalling the situation, he could control her.”

Another mediator wrote, “I was told in caucus that defendant had [a] gun in [the] car and would use it if there were problems in mediation.” Another mediator described a case in which one party called the other on the phone and recorded himself as he killed the family dogs. He recorded this on voicemail and it was played during the mediation.

The study's respondents shared a number of stories that help us better understand what kind of violence they have seen and the ways in which they intervene or respond when violence occurs. For this study, we asked mediators explicitly to describe only those incidents in which they witnessed violence or heard specific threats of violence. (It is possible that parties sometimes use coded language or even body language to convey threats of harm or coerce each other. These incidents could easily be missed by mediators unaware of the history between the parties.)

Mediators described various experience of violence in mediation. These included witnessing a party slap her own attorney in the face, mediating among gang members who had verbally threatened and menaced each other on the way into mediation, being prevented from leaving the mediation room when a party blocked the door, as well as being threatened orally and in writing. One mediator wrote, “I got a dirty phone call from a very controlling party after conducting his mediation. I immediately recognized his voice. It was threatening and I called the police.”

A few mediators also mentioned that they needed to keep parties physically separated because of histories of physical or sexual abuse. In cases such as this, two mediators noted that the alleged perpetrator pressed to see or speak with the alleged victim, attempting to use the occasion of mediation to gain access.

As shown in Table Eight, respondents shared stories about how they prevented violence by defusing tension and how they responded once violence erupted during mediation.

A few mediators indicated, not surprisingly, that with hindsight they would have handled the situation differently, but did the best they could to react appropriately in the moment. In part, the goal of this research is to provide mediators with some ideas to consider before violence occurs so they may be better prepared. Table Eight lists the most frequently mentioned strategies and the percentage of mediators mentioning them. (Some mediators mentioned multiple strategies so the numbers will not add up to 100 percent.)

As shown in Table Eight, 45 percent (24) of the respondents who answered this question (n = 53) cited reflective listening, validating emotions, and getting to the root causes of anger and fear as means of preventing violence. Active listening can help ensure that angry and/or potentially violent parties believe they are being listened to, normalizing emotional expression can help dissipate anger, and getting root causes can help de‐escalate the conflict (De Mayo 1996).

The second most frequently cited strategy for preventing violence (often used in combination with reflective listening) is the practice of separating the parties when tempers flare and/or mediators perceive a risk of violence (43 percent, 23 mediators). Mediators wrote that this strategy can de‐escalate tensions because it gives parties chances to vent to the mediator (and possibly their own attorneys) without angering or offending the other party. Mediators mentioned that in caucus, they sometimes engage in one‐on‐one coaching with the parties to help them communicate more constructively, in case they are brought back into joint session. The mediator may also show more empathy or convey concern for the party's emotions in caucus without the other side interpreting this as a sign of bias. One respondent noted, “If the mediator can get a handle on defensiveness and create a ‘safe zone’ other emotions become less intense.”

Depending on their preferred style and the case type, some mediators may use a caucus‐only model of mediation for certain cases, which might reduce the risks of some kinds of violence (e.g., one party against the other) but not all types of violence (e.g., violence against the mediator, the attorney, or in the parking lot when arriving or departing the mediation session). A risk of this practice is that the party in caucus might begin to believe the mediator is on his/her side, so it is incumbent that the mediator remind the party of his/her neutrality as necessary (Poitras and Raines 2013). One respondent specifically cited the use of “coaching” in caucus prior to the commencement of mediation, while others cited communication strategies in caucus that implied the use of coaching without explicitly using that term. (See Hoffman 2011 for a comprehensive discussion of the benefits and risks of caucusing.)

The results of our study suggest that violence is an ever‐present risk in mediation. Although no mediators in our sample were physically hit nor the targets of direct violence, many have been threatened and seen violence erupt in mediation. With this article, we seek to help mediators (and possibly attorneys) prepare themselves individually for those cases that may become violent and to stimulate analysis and discussion about the utility of various strategies to prevent and manage violence in mediation more generally; these range from broader case screening to additional training for mediators and attorneys.

In this section, we discuss the implications of our survey results in light of the research and practice advice from related fields as well as the sparse existing literature on violence in mediation specifically. First, we examine the tension between the ethical rules governing mediation and the need for safety, such as confidentiality and limits on reporting bad behavior. Then we share a variety of strategies for mediators to consider as options when and if they encounter violence or threats of violence in mediation, based on the findings of this study.

In this study, respondents told us—and we have found this to be true in our own experience as mediators—that the ethical rules governing mediation sometimes prevent mediators from reporting the bad behavior of the parties during mediation, which can undermine the safety of all concerned. Such bad behavior sometimes includes violence in mediation but more often includes threats of violence or other harm to the parties.

Rules governing confidentiality in mediation vary between U.S. states and sometimes even county by county within states. The Model Standards of Conduct for Mediators (American Arbitration Association, American Bar Association, and Association for Conflict Resolution 2005: 6) clearly state that “a mediator shall maintain the confidentiality of all information obtained by the mediator in mediation, unless otherwise agreed to by the parties or required by applicable law.” This includes no explicit exemptions related to violence. But the document (2005: 8) also includes this passage: “If a mediator is made aware of domestic abuse or violence among the parties, the mediator shall take appropriate steps including, if necessary, postponing, withdrawing from or terminating the mediation.” Many states allow mediators to breach confidentiality to report threats of harm to others or threats of self‐harm, such as suicide. Although parties are usually allowed to report threats of child abuse to the appropriate authorities, in many states or counties ethics rules prohibit directly reporting these to the judge in the case (if the case is court‐referred). One respondent wrote that he/she had wanted to report to the judge so that appropriate safety precautions could be taken when the parties appeared before the judge, but rules barring any communication with the judge made this impossible.

The Model Standards of Practice for Family and Divorce Mediation (Association of Family and Conciliation Courts 2000) seek to address concerns about safety and confidentiality by stating the exceptions to confidentiality. Standard VII of that document (2006: 6) states that “The mediator shall disclose a participant's threat of suicide or violence against any person to the threatened person and the appropriate authorities if the mediator believes such threat is likely to be acted upon as permitted by law.” Standard VIII exempts concerns about child abuse from confidentiality provisions.

Likewise, according to the Uniform Mediation Act (UMA) (National Conference of Commissioners on Uniform State Laws 2001) that was “approved and recommended for enactment in all the states,” exceptions to confidentiality are stated specifically when there is “a threat or statement of a plan to inflict bodily injury or commit a crime of violence” or there is intent to “plan a crime, attempt to commit a crime, or to conceal an ongoing crime or ongoing criminal activity” (40). (The UMA was drafted jointly by ADR professional organizations including the American Arbitration Association and the Association for Conflict Resolution. Regulators in each state or court jurisdiction may choose to adopt its rules or not.)

None of these standards codes, however, detail which authorities the mediators should report to and what exact protocols they should follow when the mediator or the participant is in danger. A particular case involving a California psychologist is critical to understanding some of the flaws in current confidentiality laws that prohibit the mediator from reporting threats made in caucus session to the other party, but allow the mediator to report them to the police. In the Tarasoff case, California psychologist Lawrence Moore learned that a patient of his named Mr. Poddar posed a threat to Ms. Tarasoff's life. Although he reported that to police, he did not report it to Ms. Tarasoff, and two months later Mr. Poddar killed her. To close this gap, the Supreme Court of California held “the psychologist who treats a patient who utters violent threats must report them to the ‘victim,’ a duty‐to‐warn doctrine” (Siegel 1979: 249). Just like the practice of clinical psychology before the Tarasoff case, however, the practice of mediation often lacks statutory or regulatory protocols to provide clear guidance to mediators.

Establishing statutory or regulatory safety protocols presents clear ethical dilemmas. Noted mediator Robert A. Baruch Bush wrote (1994: 15), “Mediators are sometimes legally required, sometimes requested, and sometimes tempted to disclose or report information from or about a mediation session to outside parties or agencies. In some instances, it is easy to see that confidentiality should prevail, but in others mediators do not find it so clear.” Bush presented several examples of ethical dilemmas posed to mediators, such as “allegations of past or threatened violence or crime,” “mediators are tempted to communicate to a court, either upon request or on their own,” and “a party has a severe need for professional help of some kind, beyond the matter in mediation” (Bush 1994: 15–16). These exemplify situations in which it is unclear whether absolute confidentiality should exist. Establishing protocols for mediators to follow would likely simplify their decision making when they confront such dilemmas. But, on the other hand, every case has unique aspects that might make rigid rules impractical. Some of the mediators in our study noted frustration with perceived limits on the extent to which they could report threats or acts of violence, and certainly more work needs to be done in this area.

Our literature review and survey results indicate that the field of mediation has less developed research and fewer guidelines for violence prevention and intervention than other related fields. But we also acknowledge that implementing rigid protocols and mandatory screening and increasing the presence of security personnel or equipment (metal detectors, for example) could make mediation more expensive, more administratively burdensome, and less accessible to parties and organizations. More work is needed to determine where our attention and resources can be best spent (e.g., training, screening, security personnel).

In addition to balancing concerns about safety and confidentiality, mediators may also be concerned with the impact of their strategic responses on their reputation, future business, and standing with program administrators who assign cases. While these are important concerns, none of the mediators in our sample mentioned them. Future research may wish to investigate the professional implications of mediator responses to violence.

Based on the themes that emerged from the survey, and as the literature predicts, parties who perceive that the mediator and/or others in the session don't fully understand their concerns are more likely to continue to convey a high degree of emotion (Poitras and Raines 2013). By explicitly acknowledging the emotional message has been received, mediators can allow parties to move forward and explain the reasons behind their emotions in the service of dispute resolution.

One‐on‐one meetings can be used to convey that the mediator has heard and understood clients’ concerns if the mediator believes this would allow a cooling‐off period or help the venting party to regain her composure. Instead of trying to silence the parties, the best advice that emerges from this study is to ensure that the parties feel heard, using caucus if needed, to ensure that venting does not have a negative impact on the overall goal of dispute resolution.

When parties sense they have little or no control over their situations, they are more likely to be angry and act out. About 26 percent of respondents (see Table Eight) indicated they remind parties about the limits of the mediation process and that the parties themselves control the outcome, assuring them that no agreement will be reached without their consent. Empowering clients in this way may reduce their perception of threat, which will likely reduce the risk of violent behaviors. If they seem to feel trapped, mediators can specifically remind parties that they have the power to end mediation at any time, but that they might want to stay to share their concerns.

In a related vein, mediators mentioned the use of distraction or redirection through well‐chosen questions designed to refocus the party's attention to issues with fewer emotional triggers or back onto issues over which they have more control.

A significant percentage of respondents (13 percent) reported that they seek to increase the presence of security personnel once it becomes clear the parties may be acting aggressively (see Table Eight). Having a security guard or sheriff's deputy stand outside the door may help dissuade violent parties from acting out, assuming they have the capacity for self‐control. This is most common in court settings or in other venues in which security personnel may be found. For many mediators and in many settings, however, this is simply not an option. In these cases, screening cases in advance will be even more critical.

The strategic use of the mediator's voice, presence, and body language can have a calming influence on a distraught party. More than one quarter (26 percent) of mediators cited the importance of using their own calming or authoritative presence as needed. We think it likely, however, that the number is greater because most mediators may not explicitly recognize this as strategic behavior; observational research, we believe, would find greater use of this strategy.

A soft voice is often more effective at calming an angry party, although some mediators reported the use of assertive, stronger tones as well. One respondent stated: “I'm a Brooklyn girl—I cuss them out or sometimes I try to calm them down and talk softly. Whatever it takes.” Vocal tone as well as the use of silence can be used to send signals to the parties about the (in)appropriateness of their behaviors or to issue specific directions such as, “Please sit down” or “Avoid using swear words, please.” Mediators can also invoke and reiterate basic ground rules, using their own authority and/or an invitation to engage in constructive behaviors together. By the use of silence, mediators can convey that any solution is up to the parties or that some behaviors don't justify the mediator's attention or response.

When none of the strategies listed above seem workable, some mediators (9 percent) reported that they chose to declare impasse and end the mediation.

One mediator suggested that mediators should be trained to predict and respond to violence before it happens, while another reported that she/he tries to create a calming environment through the use of lighting, music, food, and the like.

We note that only three of these strategies come with any financial burden, and they were the least often cited. Increasing the presence of security personnel, declaring impasse, and ensuring a calming physical environment (e.g., lighting, paint color, etc.) include costs to the mediation program or the parties, but they were the least likely to be cited by the mediators in our study. The good news is that the most commonly used strategies would incur minimal expense.

This study is an exploratory qualitative study of the violence that occurs within or immediately after the mediation process. Although there have been no studies linking mediator styles to increased or decreased violence, we note that some mediators separate the parties at the beginning of mediation, shuttling between them in a caucus‐only format. This is more common among mediators who practice evaluative/directive styles of mediation as opposed to those who practice according to facilitative or transformative models. (For more information on mediation styles, see Charkoudian et al. 2009.)

Mediation style may indeed affect the likelihood that parties become violent in mediation. We did not ask the mediators in our study to label the mediation style or process steps they use, leaving that as a fertile area for future research. Previous research, however, indicates that the majority of mediators may state a preference for one style of mediation, but once the mediation starts, they tend to vary their approach based on the needs of the parties and the case (Charkoudian et al. 2009; Poitras and Raines 2013). Thus any study that sought to establish a correlation between mediation style and responses to violence would require observations of mediation sessions.

Also, future studies should seek a random sample that better represents a full cross‐section of U.S. mediators. A larger, more randomized sample would allow for a more definitive measurement of the incidence of various types of violence within specific case types. It would also allow for a deeper analysis of the techniques used by mediators and make more definitive recommendations about training and regulation.

The ultimate goal of research on violence in mediation would be to develop best practices for the prevention of and response to violence before, during, and after mediation. This would require further study into the costs, benefits, and efficacy of various techniques.

Research should specifically examine the trade‐offs between the costs of screening and the benefits of the information gained through screening. In theory, if unlimited time and funds were available, parties could be thoroughly screened for violent histories, screened for mental health concerns, and then mediation could proceed only in those cases deemed “safe.” This also raises an ethical question: if someone has been violent in the past should he or she lose the right to try mediation?

It remains unclear whether the presence of attorneys or other support people has an impact on violence in mediation or not. Our study shows that attorneys and mediators can be the targets of violence, although they are less often targeted than opposing parties. But it seems reasonable to ask whether attorneys, union representatives, or others may also reduce (or increase) the problems of violence in mediation.

We found little comparative research on the issue of violence in mediation across jurisdictions, be they state or national. Does the incidence of violence itself vary from region to jurisdiction? We know that the training of mediators varies across mediation programs and jurisdictions as does the type of screening and preparation required. How do these differences in training and practice across regions and jurisdictions affect practice? A comparative study would likely allow us to learn from our fellow mediators elsewhere.

Making violence prevention a more integral element of mediation training programs is one obvious means of improving mediator competence in these situations. Some may resist the idea of additional training: violent episodes are uncommon enough that we could be training mediators for problems some of them will never encounter. (Noting the reflexive ubiquity of the idea that improved training will address all practice problems, one anonymous reviewer of this article referred skeptically to the “training epidemic.”) We believe, however, that mediators should be prepared to handle these difficult and potentially serious situations should they indeed arise.

But before effective training programs could be developed to teach mediators specific behavioral responses, more research would need to be undertaken to reach a consensus on which are the most effective. Without that data, it would be premature to develop in‐depth training on violence prevention and mitigation. We have attempted to begin that process with this study. (See additional discussion of training issues below.)

In many states, mediator training includes material on screening for domestic violence, dealing with escalating tensions, and safety planning to help parties leave and remain safe (Hedeen, Raines, and Barton 2010). However, comprehensive training on violence prevention and containment in mediation is often omitted in the field of mediation. This research attempts to import common training elements from other fields.

We do not seek to create the impression that bad behavior is expected or common in mediation. But while violence in mediation occurs in a minority of cases, all mediators would be wise to develop the skills needed to prevent it when possible, and handle it safely when necessary. Our study indicates that violence in mediation can occur across case types, even in commercial mediations where it has long been thought to be unlikely. The limitations of our sampling method make it impossible to know for certain which types of cases are most likely to lead to violence in mediation, and this may actually vary by jurisdiction or regional subcultures.

We anticipated that family and workplace cases would be the source of most violent outbursts, and that other case types would have relatively few violent outbursts. Instead, we found that mediators from across the case‐type spectrum had experienced violence in mediation. Notably we found a surprisingly high incidence of violence reported by commercial mediators, but no case type seems to be completely immune to these challenges.

Mediators should be trained how best to respond to de‐escalate the situation and assist the parties as they seek to overcome the underlying problems that have given rise to the outburst while attending to safety needs but, as we noted above, before techniques are taught their efficacy should be understood.

While not part of our study, previous studies indicate that family mediators often receive this type of training, but many mediators who specialize in other case types get much less training focused on violence prevention and de‐escalation (Hedeen, Raines, and Barton 2010). We failed to ask mediators what, if any, training they had received to help them prepare for and manage violent outbursts in mediation. Future studies should consider asking this question in order to determine how to improve the training of mediators on this issue.

In addition, implementing training on this issue could present logistical challenges. Training can be time‐consuming and expensive; violence prevention instruction would need to be squeezed into an already full curriculum for students who need to learn a lot in a brief time.

As we noted above, more screening might also be helpful, in particular for non‐family cases, which are less typically screened; additional research would ascertain the value of screening as a violence prevention tool.

Our findings indicate mediators use a variety of strategies to respond when violence in mediation occurs or seems likely to occur. These include reflective listening, separating the parties (caucus), reminding parties they control the outcome, adjusting their tone and demeanor, increasing the security presence, and terminating mediation without agreement. Clearly, mediators adjust their strategies, ground rules, and communication styles to meet the needs of the parties and the case at hand. Mediators appear to be dynamic in their reaction and response to violence in mediation, as needed to de‐escalate the situation and move forward safely when possible.

When violence arises during mediation, mediators stated they sometimes breach confidentiality to report the incident to law enforcement or other appropriate authorities. In many mediation programs, rules exist that allow this communication to occur. Some mediators, however, perceive ethics rules as obstacles that have prevented them from making the disclosures they felt necessary to help ensure the safety of the parties, the mediators, or the attorneys.

It seems obvious there is a tension between the need to keep parties and mediators safe, but to avoid overly burdensome training, screening, and reporting protocols. In the end, each mediation program and each mediator must decide how to prevent and manage violence in mediation based on the needs of individual cases and parties. Even when pre‐screening occurs, individual mediators should be prepared to deal with violent outbursts in mediation because no screening regime can prevent all such incidents.

In sum, we hope this research makes clear that violence can occur in many different case types and is not solely the concern of family mediators; that it also suggests some useful strategies to help mediators prevent violence when possible, and respond effectively when prevention fails; and that it stimulates deeper discussion about how best to prepare mediators and others for these difficult and potentially dangerous cases.