The EEOC Mediation Program: Mediators' Perspective on the Parties, Processes, and Outcomes

TABLE OF CONTENTS

PREFACE AND ACKNOWLEDGMENTS

I. INTRODUCTION

II. RESEARCH METHODOLOGY

1. The Subjects and Sampling Technique
2. Research Protocol
3. The Survey Instrument
4. Data Coding and Analysis

III. RESULTS

1. Characteristics of the Mediator and Mediation Session
2. Dispute Resolution Status
3. Participant Conduct that Contributes to Resolution of the Dispute
   1. Charging Party Conduct that Facilitates Resolution
   2. Charging Party Representative Conduct that Facilitates Resolution
   3. Respondent Conduct that Facilitates Resolution
   4. Respondent Representative Conduct that Facilities Resolution
   5. Mediator Conduct that Facilitates Resolution
      1. Facilitative Conduct
      2. Evaluative Conduct
      3. Personal Style
      4. Other Conduct/Tactics
4. Turning Points of the Mediation
   1. Communication
   2. Attitudes
   3. The Offer
   4. The Mediator
   5. The Representative
   6. Feelings
   7. Other
5. Non-Resolution
   1. Reasons/causes and Perceived Responsibility
   2. Conduct that Interferes with Resolution
      1. Charging Party Behavior that Interferes with Resolution
      2. Charging Party Representative Behavior that Interferes With Resolution
      3. Respondent Behavior that Interferes with Resolution
      4. Respondent Representative Behavior that Interferes with Resolution
   3. Barriers to the Resolution of Dispute
6. Hindsight: What the Mediator Would Have Done Differently
7. Evaluation of the Dispute Resolution Skills of the Parties Legal and Non-Legal Representatives
8. Mediator Suggestions for Process Improvement

IV. CONCLUSIONS

V. IMPLICATIONS

APPENDIX A

PREFACE AND ACKNOWLEDGMENTS

Dr. McDermott wrote the survey used in this study with assistance from Drs. Jose, Obar, and Dr. Mollie Bowers, a co-author on our first study. Drs. Polkinghorn and McDermott coded all data in this report. Students, under the close supervision of Drs. Obar and Jose, entered the data. Drs. Obar and Jose also performed quantitative data runs. Drs. Jose, McDermott, Polkinghorn, and Obar wrote the report (in order of editorial contribution).

We would like to thank our student assistants Susan Frank and Lubomira Spassova for their assistance in with the data entry. While so many at the EEOC have provided assistance we would like to especially thank the following people: Ida Castro, Chairwoman of the EEOC during this research project; Commissioner Paul Miller; Paul Richard, Special Assistant to Commissioner Miller; Elizabeth Thornton Director of Field Programs; Paula Choate, Director Field Coordination Programs; Irene Hill, Attorney Advisor; and Sharon Bowers, Operations and Policy Specialist. Our key EEOC contact, Steve Ichniowski, National ADR Coordinator, was of invaluable assistance. Also, this second project could not have gotten off of the ground without the timely support from Robert Jackson, of the EEOC Training Division. Tulio Diaz, Director of the Washington Field Office, and Mary O?Melveny (Special Counsel, Litigation, Communications Workers of America) were early internal advocates for the use of mediation at the EEOC and we would like to recognize their contribution to our field.

Finally, we thank Janet Dudley-Eshbach, President of Salisbury University, her administration, and the Perdue School of Business for their support for this research conducted under the auspices of the Center for Conflict Resolution at Salisbury University.

I. Introduction

This is the second of two research reports on the evaluation of the EEOC Mediation Program. The first report is entitled An Evaluation of the Equal Employment Opportunity Commission Mediation Program (EEOC Order No. 9/09/00/7632/2; www.eeoc.gov). This first report examines the program from the charging party and respondent perspective with regard to their satisfaction with the EEOC mediation process. This second evaluation report focuses exclusively on the observations, opinions, knowledge, and attitudes of the mediators.

This report presents mediator feedback on the dynamics of the mediation process, including the participant (charging party, respondent, mediator) conduct that facilitates resolution of the dispute¹, why the dispute was not resolved, mediator tactics, behavior that acts as a barrier to a resolution, the role of legal counsel and other representatives, and the turning point(s) in a successful mediation. Mediator evaluations of the skills of the parties' legal and non-legal representatives are also highlighted in the report. The report also focuses on mediator suggestions regarding process improvements. The report ends with sections on "conclusions" and "implications" which discuss the conclusions of this research endeavor and their implications.

The content of this report is the presentation of the results of 2062 mediator surveys, the overwhelming majority of which provides rich and descriptive qualitative data. It is unique because it reports the feedback from mediators, who have been given wide latitude in a largely open ended survey to discuss their opinions and thoughts on the conduct of the parties.

A conscientious effort has been made by the research team to protect the integrity of the data. Almost every mediator survey has been team coded by the same two researchers (Drs. McDermott and Polkinghorn). Such a time consuming protocol is necessary to ensure that the coders can immediately gain feedback and double check with each other on the development of codes and to discuss interesting and intriguing comments. This process also ensures that the coding is done by members of the team who are not only familiar with qualitative research, but who are also practitioners in dispute resolution (mediator and employment law litigator). The data entry and manipulation are primarily performed in tandem by the other two team members (Drs. Obar and Jose) with help from student assistants. By entering the data together the team is able to reinforce the integrity of the data. The two team members coordinate all the data runs together, firstly to rectify the data set, looking for missing or miscoded data and secondly to examine the data runs simultaneously, checking for any questionable or unexpected results. This research protocol contributes greatly to the reliability and validity of the data.

The researchers appreciate the conscientious efforts made by the mediators in answering all the questions thoroughly and thank them for patiently dedicating their time in doing so.

The data presented in this report is only the "tip of the iceberg" of the comprehensive and rich database. Further data analyses will be performed to explore many aspects of the mediation process.

II. Research Methodology

A. The Subjects and Sampling Technique

The participants in this study include all mediators who conducted a mediation session for the EEOC under the supervision of the 50 EEOC field offices, from about March 1 to July 31, 2000. The sampling technique is to capture all mediations conducted during this period so that the profile of the sample is the same as the profile of the overall population of the EEOC cases. The results from the first study verify that this sample is completely consistent with the overall profile of EEOC mediations. (See: An Evaluation of the Equal Employment Opportunity Commission Mediation Program)

The raw data set is 2062 useful mediator surveys. Only a handful of surveys have been deemed to be invalid or useless; this assures the researchers that the mediators have taken the research endeavor seriously. In many instances, a mediator conducted several mediations and conscientiously filled out the survey for each of these mediations. This reinforces the researchers' assumption that all of the mediations provide a unique glimpse into the conduct of the parties, the dynamics of the process, and the mediators' opinions on the resulting outcome. This also suggests that the mediators deem each mediation session to be unique.

B. Research Protocol

The integrity of the data is predicated on the research protocol. The research protocol requires the mediators to complete the surveys at the end of each mediation session and then place the completed survey in an envelope, seal it and forward it to their local ADR coordinator. The coordinator for each office then forwards the packages to the researcher team. Anecdotal evidence suggests that in a few cases the protocol has not been strictly followed; for example, the envelopes were not sealed or the mediators filled out the survey in front of the participants. The research team does not consider these breaches to have had a measurable impact on the overall data set.

C. The Survey Instrument

The survey is titled "Mediator-Results Survey," and is designed to measure mediators' perceptions of the parties, dynamics of the process, and the variety of outcomes that may result. The survey has 11 significant questions. Some of these main questions include subparts. The first set of questions address the characteristics of the mediation. The remaining questions seek information from the mediator regarding the conduct of the parties at mediation, the resolution status of the mediation, participant actions that contribute to or interfere with the resolution of the dispute, mediator tactics used to resolve a dispute, what the mediator would have done differently in those mediations where the dispute was not resolved, the skills of the parties' legal and non-legal representatives, barriers to the resolution of the dispute, turning points in the mediation, and process improvement suggestions. With the exception of the questions pertaining to the characteristics of the mediation session and the skills of the representatives, all the other questions are open-ended. (Please refer to Appendix A for the survey questionnaire.)

The open-ended questions are designed to maximize mediator feedback. As a result, the mediators have provided a wide range of personal opinions, insights, comments and ideas within their answers that expose numerous non-obvious patterns that would not have been discovered if the survey utilized closed ended questions.

D. Data Coding and Analysis

Each mediator is asked to complete the questionnaire for every one of his/her mediations. The data from the open-ended questions are team coded by Drs. McDermott and Polkinghorn. In order to ensure greater inter-rater or inter-coder reliability, the coders worked side-by-side throughout the seven months of the coding process. This part of the process involved hundreds of hours of coding. Each survey has to be read in its entirety and in many cases by both coders. The subjectivity of attributing cause and effect for each survey requires the coders to discuss many of the surveys individually. In some cases it requires both the coders to read a survey, write down a set of ideas or working codes, and then debate the interpretation until a consensus is reached on the most probable intent of the mediator. In the majority of cases the mediator has been extremely articulate and clear in providing answers, thus making the interpretation of the data simpler and less subjective.

For two questions the coders have also entered their opinions regarding whether the mediator is attributing the failure to resolve the mediation to either the charging party or the respondent, their representative(s), or to some other discernable reasons. The main rationale for this process is that mediators, as professional neutrals, are often not comfortable identifying a particular party as the cause of the lack of resolution of the dispute. After a thorough read of the survey the coders determine which party or parties caused the mediation to stall or fail. Thus, through the coders, an attempt is made to provide some additional perspective on which party interfered with resolution.²

The following data analyses are presented as summary statistics, such as frequencies, percentages, and averages (means). In addition, in order to provide a rich descriptive example of the data patterns that are quantitatively presented, a number of qualitative data entries are used throughout the report.

III. Results

A. Characteristics of the Mediator and Mediation Session

Mediators are asked whether this is the first mediation between the parties, the number of times the parties had met before, if at all, whether the mediation was held over consecutive days, the time spent in any prior mediations, and the time spent in that prior session or sessions. The remainder of the survey elicits information, which is solely based on the mediation that took place on the day that the survey was executed. The researchers did not try to follow any mediation through to completion. Rather, we are using a "snap shot" approach to focus on the mediation. This is justified for cost reasons and because our focus in this study is on the immediate feedback from mediators concerning the participants, the process dynamics, and the outcome of this one mediation session. Recording the mediator's perspective at the time of the mediation is therefore required.³

During the early administration period of the survey a change was made to one of the threshold questions. We added a question in the survey to identify whether the person is an EEOC staff mediator, a contract mediator or a pro bono mediator. As shown in Table I, of the 2062 cases⁴ where mediators provide feedback, 1660 cases (81%) provide information regarding the mediator type, of which the vast majority (80%) identify themselves as EEOC staff mediators. Mediators in the rest of the cases are either contract mediators (9%) or pro bono mediators (11%). Mediators in 402 cases (20%) did not provide information regarding the mediator type.⁵

Table I
Mediator Type

Mediators are then asked to identify the type of mediation session. As indicated in Table II, the vast majority of the cases (97%) involve the mediator's initial session with the parties. A few mediators (55 cases, 3%) report that they had pre-mediated this case before coming to the present session. This 3% subset of mediators are further questioned regarding the number of times they have met before, the length of the session on these prior meetings, and whether these prior meetings were held on consecutive days. When asked to identify how many times they met before, mediators in 24 cases (67%) report that they met once, while mediators in 10 cases (27%) indicate that they met two times. Mediators in 19 cases (35%) did not answer this question. (See Tables III and IV.)

Table II
Initial Or Ongoing Mediation

Table III
The Number of Prior Mediation Days

Table IV
Length of Time of the Mediation Sessions on Prior Mediation Days

As indicated in Table V, mediators in 35 of the 55 cases who conduct prior sessions with the parties provide information regarding whether the sessions are held on consecutive days. Of this group, the vast majority (mediators in 32 cases, 91%) report that the sessions are not held on consecutive days.

Table V
Were Prior Mediation Sessions Held On Consecutive Days?

Mediators are also asked to report the duration of the mediation session. The vast majority (85%) lasted five hours or less. Sessions of five hours or less in duration are measured in one hour intervals of 1-2 hours (15%), 2-3 hours (30%), 3-4 hours (24%), or 4-5 hours (13%). The average length of time for the mediations is 3.66 hours. This information is presented in Table VI.

Table VI
Length of Time of the Mediation Session on the Day of Survey

In summary, the majority of the mediators surveyed are EEOC staff mediators.⁶ The majority of the mediation sessions are the first between the parties and the average length of the session is 3.66 hours. The next section presents mediator feedback regarding the resolution status of the disputes.

B. Dispute Resolution Status

Of the 2062 cases studied, 1223 (60%) reached a resolution while 811 (40%) were not resolved by the end of the surveyed session. Information about the resolution status is not available for 28 cases (1%). Table VII presents this information.

Table VII
Resolution Status of the Dispute

C. Participant Conduct that Contributes to Resolution of the Dispute

Detailed questions elicit descriptive accounts concerning the behavior of the participants that help to facilitate the resolution of the dispute. Specifically, the survey inquires about the conduct of the charging parties, charging parties' lawyer or representative, respondents, and respondents' lawyer or representative. Mediators are also asked to identify their own conduct that they think contributes to the resolution of the dispute.

For each of the four possible parties present (charging party, charging party attorney or other representative, respondent, and respondent attorney or other representative) the mediator is asked, "If the dispute was resolved, describe the conduct you believe facilitated the resolution by (the particular party)." The mediator then identified the conduct of each of the parties present.⁷

1. Charging Party Conduct that Facilitates Resolution

For charging parties, the number one behavioral attribute that contributes to the resolution of the dispute is their flexibility and openness. This flexibility and openness involves a wide range of actions and behaviors described by the mediators as honesty, flexibility, openness, realism, straightforward dealing, good problem solving skills, taking some responsibility for the dispute and other similar types of behavior. It may be that those who either have a good substantive case or who have confidence in their abilities are able to act in a more open and flexible manner. Another interpretation may be that the relationship of the charging party to the respondent is sufficiently strong, which may aid in the charging party's ability to be open and flexible. In some of the cases it appears that the mediators think flexibility and openness can be attributed to the character and disposition of one or more parties. In other instances flexibility is manifested in the desire to be creative and crafty in approaching the challenge and constructing a solution. This particular interpretation is indicative of someone who is process savvy, knows how problem solving and decision making processes operate, or has some knowledge of how to incorporate others' ideas and needs to search for the common ground on the issues being discussed. In 56% of the cases that are resolved, the mediators report behavior that we classify under this category.

The next behavioral attribute that facilitates settlement is the charging parties' focus and preparation (12%). The observations within this category include knowing what one wanted, staying focused,being prepared, and command of the facts. Being prepared may not only help the parties consider the strengths and weaknesses of each person's case, but also help them consider the common ground where compromise may be found. The attribute of "being focused" demonstrates how serious the charging party is, which is important to note given the finding in the first study that indicates that the charging party relies heavily on the mediator for assistance in process and outcome questions. An example of a typical response to the code "focused and prepared" looks like this: "Charging party was experienced and knew what she wanted and could accept. She didn't let the fact that the respondent had an attorney scare her."

Charging parties' need for closure (7%) and their willingness to listen to their lawyer (7%) also contribute to the resolution of the dispute. The need for closure may be indicative of emotional fatigue, process fatigue and/or a controlling desire to conclude the process and move on with one's life, possibly without regard to the quality of the outcome. This need is important to recognize since it may pose a dilemma to the extent that it interferes with a party's ability to vindicate their rights under the law. An example of a typical data entry for the code "need for closure" looks something like this: "Had a desire to let go of issues after venting the concerns."

The willingness to listen to ones' lawyer suggests two conclusions. The first is the implicit understanding that mediation, in this case, is a supplement to the legal process. This implies that EEOC mediation is an arena whereby rights based disputes are being translated into interest based discussions. The coders also note that this code implies that the charging parties acknowledge their attorney as a stabilizing factor within the process.

2. Charging Party Representative Conduct that Facilitates Resolution

Charging parties' representatives (legal and non-legal) may also contribute to the resolution of the dispute by engaging in a wide range of conduct, strategies and tactics that are conducive to resolution. In many cases representatives act as a reality check (9% of the cases), are flexible and compromising (7%) and engage in conduct that is coded as expression or conduct that established commitment to resolution (2%). Reality checking is an excellent way to compare what the charging party wants and can legitimately expect to get from mediation or any other decision making process. Any good counselor realizes that providing a check on the probability of success for a given course is the main task that levels or matches the charging parties' aspirations to the most likely set of expectations. Reality checks include providing hypothetical scenarios and an honest examination of the probability of success of a given decision making route. The code "flexible and compromising" has a number of properties, the most notable being able and willing to incorporate others' wants, demands, desires or needs into the crafting of the solution. The code "expression of conduct that established commitment to resolution" is rather large and can be sub-coded to tease out subtle patterns. However, for the most part, this code houses data instances that demonstrate the willingness of the representative to overtly state the desire to work together with the respondent and mediator to work out a solution. In many instances this is indicative of an expression of coming to the table in good faith.

Charging party representatives are supportive of their client (3%), listened (2%), and engaged in creative problem solving (1%). These three codes, when combined, verify a simple observation that support of the client, effective communication, and problem solving go a long way to helping the charging party and the respondent create a solution.

The demeanor (5%) of the advocates, such as calm and professional behavior, and preparation/advocacy skills (6%) also help in the resolution of the dispute. The main properties of these codes focus on the seriousness and professional conduct of the representative. As we will see in a later section, a main reason for failure is the exact opposite set of behaviors and traits.

3. Respondent Conduct that Facilitates Resolution

The category of flexibility and openness contains the largest amount of observed behaviors contributing to resolution (27%). These behaviors are mainly descriptions of positions or conduct at the mediation, such as a willingness to negotiate and/or to explore settlement options, act in good faith, flexibility, expression of a desire to keep the charging party as an employee, and other such conduct.

Another act by the respondents that helps to resolve the case is the expression of a commitment to the resolution of the dispute (17%). This includes the need to resolve the dispute that very day, a desire to avoid an EEOC investigation, the expressed desire for the charging party to leave having saved face or leave feeling the problem has been solved, and even the private admission of liability to the mediator.

We also identify behavior that falls under the category of good mediation skills (16%). This category involves the identification of mediation specific skills, such as the ability to communicate, sincerity, patience, and the willingness to share information. A good example of this code from the data is "A willingness to involve the Company in the mediation process. Openness in discussion leading to a resolution."

We find that 13% of the cases are resolved by apology/admission/remorse,empathy, or recognizing the charging party's concerns. Expression of apology/admission/ remorse by the respondent is important for resolution. Three percent apologized, expressed remorse or engaged in apologetic conduct. This contrition is usually attributed by the mediator as an important reason for resolution of the dispute. An example from the data is: "Apologized up front."

A related respondent behavior is the expression of empathy. In 5% of the cases, showing concerns for the charging parties' feeling contributed to resolution of the dispute. Close to empathy, but coded as a separate observation, is the conduct of addressing and/or understanding the charging parties' concerns (5%). Thus, in 1 of 10 cases either expressing empathy or showing that one understands the charging party concerns results in a resolution of the dispute.

Other respondent behavior that contribute to the resolution of the dispute includes listening (10%) and direct communication avoiding taking a position and/or the facts of the case (3%). These codes are indicative of a person who can effectively communicate and understands the distinction between an interest based discussion and a positional based debate.

4. Respondent Representative Conduct that Facilitates Resolution

Mediator perceptions of the conduct of the respondents' representative that facilitate resolution include conduct that we categorize as demeanor elements (14%). This includes a number of descriptive characteristics, such as being calm, knowledgeable, persuasive or attentive. Expression of a commitment to resolution is the second largest group of behaviors by representatives that contributed to the resolution of the dispute (10%). Next we find flexibility/ willingness to compromise (7%) and the mediation skills of the representatives (7%).

The respondents' representative also engages in conduct directed at the client which we categorize as client convincing conduct (5%), that includes educating the client and reality checking. Focusing conduct (4%) is, among other things, the act of helping the client understand developments as the process continues as well as getting the client to concentrate and prioritize items, ideas or options. Also important is the expression of respect for the charging party (3%). An example from this code is "Respondent's attorney was non-adversarial and responded to CP in a very positive, encouraging manner." The information regarding participant conduct that facilitated mediation is summarized in Table VIII.

Table VIII
Charging Party, Respondent, and their Representative Conduct that Facilitates the Resolution of the Dispute

^{Note: Total number of resolved cases is 1223. Since some mediators have given multiple responses, the total number of responses differs from actual number of resolved cases. The percentage given is a function of the resolved cases and not of the responses or participants.}

5. Mediator Conduct that Facilitates Resolution

As shown in Table IX, mediators describe at length their own conduct that facilitates the resolution of the dispute. Not surprisingly, mediators offer a large amount of data to the inquiry, which might be due to the fact that they can reflect more on their own behavior than on the behavior of other parties. It might also be the only opportunity for them to express their ideas on how they have positively impacted the dynamics of the process and the resultant outcome. The survey is designed so that the mediators are asked two separate but closely related questions in an attempt to identify any conduct or tactics they use. The first question asks the mediators to identify the conduct that they engaged in to facilitate resolution. The second question asks them to identify any particular mediation tactics they use to resolve the dispute. This research design provides the potential for redundancy so that we capture the maximum mediator feedback on their conduct and tactics that promote a solution. Thus, while the mediators identify certain conduct that they engage in response to our first question, they often identify additional conduct, tactics, and strategies in response to the second question. The coders identified a few mediators who gave the same one line answer for every survey that they filled out, which brings into question how serious they may have been in completing the survey. However, the overwhelming majority of mediators took the time to separately answer, often in great detail, our questions regarding the conduct that they engaged in to bring about a resolution.

The responses can be characterized as both facilitative and evaluative behavior. Facilitative behavior focuses on process considerations. Evaluative behavior involves the behaviors used by advocates, counselors, and people who provide direct substantive input into the deliberations. This is interesting because the EEOC program trains mediators in facilitative mediation and requests that the mediators use facilitative techniques. Our results indicate that in the real world of mediation this is not what always happens. Often the mediators indicate that they used both facilitative and evaluative tactics and often appear to do so fluently.

a. Facilitative Conduct

Mediators in 53% of the cases engage in strategies and tactics that we place in the facilitative category. Listening is a prevalent tool. Within this category we include a wide range of communicative behavior. This broad listening can be further broken down into: listening, reflexive questioning, paraphrasing, restating and facilitating a catharsis or dialogue (10%), reframing/use of probing questions (this includes moving the parties from positions to interests – 4%), helping parties see different vantage points (5%), clarifying facts/areas of agreement (3%), and defusing negative emotions (3%).

In addition to listening, mediators also use other types of facilitative behavior, such as encouraging openness, honesty, and direct communication (11%), keeping parties focused (8%), encouraging resolution (5%), gaining trust/rapport (3%), and stepping back/letting parties fashion the remedy (3%).

b. Evaluative Conduct

Mediators in 46% of the cases use evaluative conduct to help resolve the dispute. Mediators report engaging in behaviors that we broadly classify as "reality checking" behavior (27%), evaluating the strengths and weaknesses of the charge (7%), providing knowledge of the law/process (6%), exploring/ proposing options (6%), and negotiating actively (1%). The specific act of "reality checking" is the single most often-cited mediator conduct that contributes to the resolution of the dispute.

Although evaluative mediation skills are not emphasized in the EEOC training, they often happen and we posit that this may be due to a number of reasons, two of which stand out. The first is that with 80% of the mediations being mediated by EEOC staff we have a pool of process professionals who have substantial knowledge of the substantive matters being discussed and they may therefore venture into substance based discussions. Another is the notion, previously discussed, that the substance of the charge is legal or rights based in nature and so the transferal to an interest based forum like mediation does not completely eliminate the psychological nature of a rights based dispute. In other words, everyone knows that if the mediation fails, the dispute will return to a rights based format.

c. Personal Style

Personal style was identified in 19% of the cases. The "demeanor" of the mediator has a significant impact on the resolution of the dispute in 6% of the cases. Other personal variables cited include empathy (4%), persistence (4%), neutrality (3%), optimism (2%), and creativity (1%).

d. Other Conduct/Tactics

Caucus is an important mediator tool. Many mediators attribute the successful resolution to their use of or manipulation of the caucus (13%). This includes use of the caucus, use of long joint caucuses, attorney only caucuses, and caucuses where the attorneys are excluded. Success is at times predicated on separating the parties, which is indicative of a number of things: the parties are too emotional, they can not agree on an issue face to face, the issues are so contentious that only the representatives can discuss them, their relationship is too strained for face to face discussions, or a potential loss of face looms that can block the movement toward agreement. Whatever the reason, caucusing effectively temporarily removes one of the basic tenants of mediation – face-to-face real time discussion.

Process control is used by mediators to bring about dispute resolution. In 4% of the cases mediators identify process control tactics that contributed to a resolution.

Other identified tactics include such things as transformative or "quasi-transformative" tactics (1%). Transformative mediation is a relationship-based approach to conflict intervention that stresses that the mediation, if facilitated properly, can become a form of empowerment for the parties.⁸ One potential explanation for this low frequency may be the nature of the issues being discussed, the need for a solution in order to avert an investigation or litigation, and the relatively low probability that these parties will be working with one another in the future. In short, the context, relationships, types of issues, and the emotional state of the parties may necessitate the need for the mediator to become more directive, controlling, and evaluative simply to keep the process on track.

Mediators also report that they control the introduction of evidence (1%) as a means of facilitating resolution. This means that the mediator can act as a filter as information is brought into the process. The way information is presented has an impact on how the other interprets it. In this instance, the mediator is assisting the communication process regarding the presentation of the fact pattern. The category also emphasizes the literal control of evidence. In some instances, the mediator will frame the discussion so that evidence is presented in a logical and timely manner. In both instances, this is a process control issue that focuses on the way information is used to discuss issues.

The use of pre-mediation calls (1%) is also another means of preparing and controlling the process. When mediators pre-arrange the presentation of issues they are effectively manipulating the process. In one instance the mediator kept the parties separated during the entire mediation and acted as a shuttle diplomat taking offers back and forth. The use of a pre-mediation conference call tipped the mediator off that the parties would likely be hostile to one another so the mediator decided to minimize this possibility by controlling contact during the process.

Table IX
Mediator Conduct that Facilitates Resolution of the Dispute

^{Note: Total number of resolved cases is 1223. Since some mediators have given multiple responses, the total number of responses differs from actual number of resolved cases. The percentage given is a function of the resolved cases and not of the responses or participants.}

In summary, the conduct that contributes the most to the resolution of the dispute is the parties' and or their representatives' flexibility, realistic attitude, openness and willingness to compromise. Good mediation skills are important to facilitate conflict resolution. They include good listening habits, reflexive and active listening, reframing, reality checking by representatives with their clients, the ability to articulate issues, and respect for others. Preparation is another important variable that mediators report makes a difference.

Mediators engage in both facilitative and evaluative techniques to bring about resolution. Listening, reality checks, and caucuses are three key tactics used by mediators. Mediators seldom mention the use of transformative mediation tactics. One potential reason for this frequency might be due to the nature of the issues being discussed, the need for a solution, and the relatively low probability that these parties will be working with one another in the future.

D. Turning Points of the Mediation

In our analysis of mediations that result in a resolution, in addition to the mediator's perception as to the conduct of the parties that facilitate resolution, we sought another mediator perspective. If the dispute is resolved, what do mediators attribute as the turning point or points in the mediation? Mediators provide a unique perspective on the incidents and behavioral patterns that are critical for the resolution of the dispute. As process oriented participants or process experts, the mediators are not only conscious of process dynamics, they often deliberately manage the dynamics to ensure that every opportunity is made to keep the process moving. Mediators are afforded the opportunity to identify more than one turning point.

Mediators' responses to the open-ended question on turning points can be classified into seven groups: communication, attitudes, the offer, the mediator, the representative, feelings, andother variables. While coding the data we observed that although we had expected the identification of a certain discrete act or acts as turning points, the mediator responses included a wide range of behavior. Most of the responses to this question are not a single discrete act as one might expect from use of the phrase "turning point."

1. Communication

Communication and discovery-related (CDR) variables are turning points in 46% of the cases. Among the different CDR variables, changes in charging party and respondent position, demands, or behavior due to information obtained at mediation is the single largest turning point category (44%). This information is influential in cases involving 27% of the charging parties and 17% of the respondents. Thus, charging parties appear to benefit more from the information obtained at the mediation. This makes sense in that employers often have information that has not been provided to an employee that influences whether or not the employee believes that he or she has been the victim of illegal discrimination.

When parties come to mediation and gain new knowledge, facts, and information the process acts to empower the individual party. The literature on empowerment demonstrates that information exchange leads to new perspectives and therefore new information can positively impact how the parties solve the problem.⁹

How information is communicated and received plays a major role in moving the process toward resolution. A well thought out opening statement (1%), a charging party's offer of evidence to support a claim (1%), and pre-mediation dialogue between the parties (1%) are also cited by mediators as turning points. These codes, while reported separately, join with the information providing conduct data discussed above to complete the category of communication and discovery-related variables.¹⁰

2. Attitudes

Based on the words used by the mediators we classified a group of turning points in the category of parties' attitudes (39%). While the mediators report attitudes as turning points, it is clear from the coding that these attitudes usually did not develop at the mediation, but are more likely aimed at one's mindset, worldview or perspective. In some instances, attitudinal attributes are connected to the resultant behavior and so we code one part of it attitudinal and the other under the behavior category. Among the various attitudinal variables, both parties' openness to compromise and resolution (26%) is cited as the number one turning point. Other attitudinal variables include the charging parties' willingness to compromise (5%), respondents' willingness to compromise (5%) the parties' mutual trust (2%) and their attitudinal change as seen in a shift from positional to interest-based discussion (1%). All of these attitudinal measures have one major theme in common: the desire to focus more on common ground than on those things that divide the parties. Seeking compromise requires an attitude conducive to the search for common ground.

3. The Offer

In 11% of the cases, presentation or discussion of "the offer" is cited as a turning point. While this may be the result of information obtained at mediation or due to party attitudes, the presentation or discussion of "the offer" represents a distinct chronological point in the mediation. The offer includes conduct such as the respondent offering more money and/or other benefits, or an offer to hire/or rehire (5%), and the charging party or the respondentmeeting all demands (3%). In other instances the parties' exchange of a reasonable initial exchange or offers (1%) or the parties' descriptions of the resolution offer (1%) and the communication of the final offer (1%) are turning points.

4. The Mediator

Mediators sometimes cite their own mediator behavior (12%) as a turning point. Strategies and tactics that they identify as a turning point include reality checking (5%), the use of a caucus with the parties (5%), encouragement by the mediator to settle (2%), and keeping the parties focused (0.3%). When the mediator acts as a process resource, such as a reality checker, or a reflecting board, or otherwise manages the parties' dialogue effectively, this conduct can aid in reaching a solution. By keeping control of the process dynamics through the use of caucusing, the mediator can keep the process on track. Another technique mediators use is to keep the parties focused. Given all the other possible reasons why disputing parties often lose focus or worse get so far down into the specifics of their case that they are mired in detail and can not see the general reasons for the dispute, it is imperative that the mediator focus and refocus parties from time to time.

5. The Representative

In 12% of the cases the behavior of legal and non-legal representatives is identified as a turning point. Charging parties' lawyers (6%) and respondents' lawyers (4%) play an important role that can lead to the resolution of the dispute. Lawyers who are familiar with or experienced in mediation understand the dynamics and intention of the process. As such these lawyers often act as stabilizing forces by keeping their clients on tract, checking expectations with the potential outcome that the process can deliver, and supporting clients through a host of other behaviors. The influence of external parties, such as the co-owner and spouse (1%) and the direct participation of key players from the respondents' side in mediation (0.5%) are also turning points.

6. Feelings

In 9% of the cases, acknowledgment of feelings and culpability serves as a turning point.¹¹ More specifically, respondents' recognition of charging party's hurt (3%) parties' recognition of each other's position (2%), respondents' admission of some liability/apology (2%), charging parties' admission of some liability (0.4%), charging parties' venting/catharsis (2%), and the charging parties' specific expression of need (0.3%) are cited as turning points in mediation. Attending to the others' feelings, empathizing and demonstrating concern in a sincere manner can be a turning point. Admission of some liability may also act as a turning point because it may reduce the tendency of one party to act as a complete victim and therefore require the other to assume the role of the villain. Also, by apologizing one not only takes responsibility but also sidetracks the potential for a long-winded scenario where one person attempts to corner the other into being the source of his/her problem.¹²

7. Other Variables

Other turning points include the parties' need to preserve future relationship (1%), which can be a good place to begin the search for common ground. In other instances it is the exact opposite whereby the charging parties "moving on" with their life (1%) is the primary motivating factor in obtaining a resolution. One unusual code is the strong desire to leave the mediation for reasons, such as time pressures, personal schedules, bad weather, need to travel, getting to another meeting, and other non process oriented reasons (1%). As the literature demonstrates repeatedly, if there is a time line and cutoff where the mediation must end then this can begin to act as turning point.

Included in the miscellaneous category (2%) are turning points such as the use of a co-mediator, current relationship with lawyers, moment of silence, and removal of a person from the process. The data in the "miscellaneous category" needs further examination to tease out patterns and/or to place some of these codes in other coding categories. This information is summarized in Table X.

Table X
Turning Points of the Mediation

^{Note: Total number of resolved cases is 1223. Since some mediators have given multiple responses, the total number of responses differs from actual number of resolved cases. The percentage given is a function of the resolved cases and not of the responses or participants.}

E. Non-Resolution

1. Reasons/Causes and Perceived Responsibility

Mediators are asked to list explain in detail the three main reasons for the non-resolution of the dispute. This results in a narrower coding range. In 53% of the cases, mediators attribute the non-resolution to a range of behavior codes in the category of the charging party or parties' actions. This does not necessarily mean that the charging party was acting unreasonable, although many of the responses indicate this typically happens. In 46% of the cases, the respondent's actions are held responsible. In 31% of the cases, the non-resolution was attributed to the actions of both parties. In coding this data where up to three reasons are listed, it was possible to have a first response that identified both parties, and a second response that identified one party or the other. Thus, these particular codes are not mutually exclusive and our reported results are cumulative.

Other reasons for non-resolution include the actions of the charging party's lawyer (13%) and the actions of respondent's lawyer (7%). The parties sometimes used nonlegal representatives who the mediators indicate are a barrier to resolution - 1.8% for charging parties and 1.4% for respondents.

Outside factors (8%) that are not under the control of the mediator also blocked the resolution of the dispute. Here both charging parties (3%) and respondents (5%) are influenced by outside events or actors, such as unions and outstanding grievances, insurance companies, jurisdictional questions, and other legal causes of action outside of the scope of the EEOC dispute. The lack of legal representation is another reason cited (1%). Other (1%) reasons are mediator self-blame and the lack of urgency to settle on the day of the survey. Table XI presents this information.

Table XI
Specific Reasons Cited for the Non-Resolution of the Dispute

^{Note: Total number of unresolved cases is 811. Since some mediators have given multiple reasons, the total number of responses differs from actual number of unresolved cases. The percentage given is a function of the unresolved cases and not of the responses or participants.}

The next step in the coding protocol for this survey question is to identify the primary obstacle to the resolution of the dispute. This is accomplished by our analysis of the specific reasons cited by the mediator for the non-resolution of the dispute. The coders then attribute responsibility to the most culpable party or other factor.¹³ This coding requires that the coder select one choice from the list of possible codes. While in many cases the mediators readily point to the party or reason, this was not always the case. Here, the coder made the final decision based on the mediators' comments and, in harder to discern cases, often in conjunction with input from the other coder.

Mediators perceive that both parties are equally responsible for non-resolution in 33% of the cases. The perceived responsibility is placed on the charging party 28% of the time, charging parties' legal representatives 8% of the time, and their non-legal representatives less than one percent of the time (1%). Mediators report that the respondent is responsible for non-resolution 22% of the time, their legal representatives 3% of the time, and their non-legal representatives less than one percent of the time (1%).

Sometimes the mediator needs more time. This was attributed as the reason for non-resolution in 3% of the mediations.

Table XII
Perceived Responsibility for the Non-Resolution of Disputes, Imputed From Mediators' Comments

Note: Total number of unresolved cases is 811. Since some mediators have given multiple reasons, the total number of responses differs from actual number of unresolved cases. The percentage given is a function of the unresolved cases and not of the responses or respondents.

2. Conduct that Interferes With Resolution

After identifying the three main reasons why the dispute was not resolved, we asked the mediators to identify any conduct by the parties that interfered with the resolution of the dispute.

a. Charging Party Behavior that Interferes With Resolution

Mediators identify participant behavior that they believe interferes with the resolution of the dispute. As shown in Table XIII, in 73% of the cases, the mediators indicate that the charging parties engage in conduct that interferes with the resolution of the dispute. In particular mediators state that the charging party engages in some type of interfering behavior, positional conduct or both in 55% of the cases. This category includes behavior such as unrealistic, rigid, not prepared, inflexible, passive, language problems, need for an attorney, overconfidence, inability to make a decision, not knowing what one wanted, failure to address the real issue, failure to distinguish unfair from illegal, and other similar behaviors.

We separately code conduct that we deem to be driven by the charging parties' emotion (15%). These behaviors include blaming the other party, anger, intent to "make the company pay," and being too upset to effectively participate.

In addition, the charging parties can be influenced by external events, which can either block resolution or otherwise interfere with the mediation. These external issues include other legal causes of action (e.g. defamation, wrongful discharge, breach of contract, worker's compensation – 3%), the charging party's financial situation, and other events or factors outside of the mediation.

b. Charging Party Representative Behavior that Interferes With Resolution

Mediators also cite specific behavioral attributes of the charging parties representatives' (legal and non-legal) that interfere with resolution. This includes the unrealistic/unreasonable evaluation of the claim (18%), adversarial/arrogant/

argumentative conduct/hijacking of the mediation process (11%), and demonstrating in some way a lack of commitment (5%) to the mediation process or to the resolution of the dispute. Mediators also observe that the charging parties' lack of preparation (4%), and use of mediation as a discovery tool (2%) are other factors that cause the dispute to unresolved.

c. Respondent Behavior that Interferes With Resolution</5>

Mediators observe that the single largest grouping of respondent behavior that interferes with the resolution of the dispute is an unrealistic and/or rigid position or related behavior (32%). This behavior includes descriptions of behavior such as unrealistic, entrenched, wanted to much, failure to address the real issue, refusal to make an offer, and similar conduct. Respondents' can also be too adversarial, emotional, or adopt a "holier than thou" attitude all of which we classified in the attitudinal cluster (10%). Respondent lack of decision-making authority interferes with the resolution in many mediation cases (6%). Respondents also can be influenced by external events (1%) that interfered with the resolution of the dispute.

d. Respondent Representative Behavior that Interferes With Resolution

Mediators' perceptions of the conduct of the respondents' representative (legal and non-legal) that contribute to non-resolution include their unrealistic and/or unreasonable evaluation of the claim (15%), adversarial, arrogant or argumentative attitude including hijacking of the mediation process (10%), limited decision-making authority (3%), lack of preparation or knowledge (2%) and overt hostility to the charging party (1%). In a few cases (1%) the fact that the EEOC dispute was part of a larger litigation action also interferes with the resolution of the dispute. Lack of adequate client preparation (1%) and other factors, such as the use of scare tactics and tactics that limited the mediator are also cited as behavior that interfered with the resolution of the dispute.

Table XIII
Mediator Perceptions of Participant Behavior that Interfered with the Resolution of the Dispute

Note: Total number of unresolved cases is 811. Since some mediators have given multiple reasons, the total number of responses differs from actual number of unresolved cases. The percentage given is a function of the unresolved cases and not of the responses or participants.

3. Barriers to the Resolution of the Dispute

We ask the mediators to list the five most important barriers to the resolution of the dispute. The question asks that they list these barriers in order of importance. This open-ended question allows the mediators to identify anything that they perceive to be a barrier to resolution. While the survey requests the mediator to list up to five barriers, in many cases they identify less than five.

The barriers to resolution question requires the coders to once again add an additional summary code that requires a subjective determination of the response as to which party or representative is seen by the mediator as the major barrier to the resolution of the dispute. We attempt to ascertain whom the mediator identifies as the party or parties that caused the mediation to stall. The same coding protocol we use for the previous question (on why the dispute was not resolved) is also used here. The one difference between these questions is that this one requires the mediator to identify the barriers in hierarchical order. Thus we are able to discuss how often a barrier is ranked within the five-scale range.

Mediator feedback regarding the barriers can be broken into the following major categories: 1) positional conduct of the parties; 2) the "table" conduct of the parties; 3)3) the conduct of the parties' attorneys or other representatives; 4) emotion/attitude; 5) lack of respondent authority at the table; 6) external factors; and 7) miscellaneous factors.

The positional conduct of the parties is the most significant barrier to the resolution of the dispute. It is listed as the number one barrier in 65% of the cases. A further breakdown shows that this is reported as the charging parties' positional conduct in 56% of the observations. Moreover, this is reported as the number one barrier in 37% of the cases. This category includes behavior such as an unrealistic or unreasonable position, refusing to settle for reinstatement only, seeing the other party as wrong at the expense of the process, refusing to go along with the process, insisting on reinstatement as a condition for resolution, the inability to make a decision, not knowing what one wanted and similar conduct.¹⁴ The respondents' positional conduct constituted 46% of the barrier observations. However, it was only identified as the number one barrier in 20% of the cases. This appears to signal that the mediators saw the charging parties' conduct as the number one barrier much more often.

Within the category of respondents' positional conduct, in addition to the broad category of taking a position that was not conducive to a mediated resolution, we identified particular conduct and/or positions that acted as barriers. In 5% of the cases the respondent took a position of innocence/did nothing wrong; in 4% of the cases the refusal to pay any money was the barrier. In 3% of the cases the respondent would not offer anything! In other words, the respondent attended the mediation and refused to make any concession -- monetary or nonmonetary.

Both parties' positional conduct is identified in tandem as a barrier. This barrier is reported by mediators in 16% of the observations and is ranked as the number one barrier in 9% of the cases. In addition to the conduct identified above for each party, this category also had a reciprocal element including the inability of the parties to work together to resolve the problem or to value the case, and a zero sum factual face-off that handcuffed the parties.

The second major category of behaviors that act as a barrier is the "table " conduct of the parties (41%). Here we identify actual negative bargaining/mediation conduct that we see as more than merely positional conduct. While it is sometimes difficult to distinguish between a position and conduct, here we identify acts that serve as barriers at the mediation table. The charging party table conduct was identified in 24% of the observations. Such conduct includes the refusal to explore options, listen, negotiate, discuss all issues, detail one's damages demand, or to show any interest in settling the dispute. This charging party table conduct was the number one barrier in 6% of the cases.

The respondent table conduct was identified as a barrier in 17% of the observations and as the number one barrier in 6% of the unresolved cases.

The third most common barrier is the conductof the parties' attorneys or other representatives; thisconstitutes 27% of the barrier observations. Note that attorneys could be a barrier due to their aggressiveness, unreasonableness, lack of preparation, positions, or even be a barrier just by doing their job and advising their clients against settling. Thus, this measure includes both destructive conduct and professional conduct. The same could be said for the parties' other representatives. The conduct of the charging party's attorney is identified as a barrier 13% of the time and is the number one barrier in 4% of the cases. The conduct of the respondent's attorney is observed as a barrier 9% of the time and as the number barrier to resolution 3% of the time.

The conduct of the charging parties' other representative is observed as a barrier to resolution 2% of the time and in 1% of the cases is the number one barrier. For the respondents' other representative this constitutes 1% of the observations and the number one barrier in a half percent of the cases. Here is a good example of a non-legal representative having difficulty. "Never represent your mother. Charging Party attorney couldn't be objective and couldn't provide important advice. I didn't feel as free to ask probing questions due to their relationship. Charging Party attorney was not a labor lawyer."

The fourth category of barriers is the emotion/attitude of the parties. The charging party's emotion/attitude is a barrier in 14% of the observations and was the number one barrier in 4% of the cases¹⁵. This conduct includes being distraught over the discrimination at issues, anger and other emotions. The respondent's emotion/attitude is identified as a barrier in 7% of the cases and the number one barrier in 1% of the cases. This respondent barrier was often described as the lack of respect for the charging party or other such language that strikes at the dignity of the charging party. In addition, a closely related to emotion/attitude is the category of personality clash. Some type of personality clash, which includes ethnic tension, dislike, and related behaviors are identified as a barrier in 2% of the observations. It is the number one barrier 1% of the time. While we report this as a separate category, it could be included in the emotion/attitude barrier category.

These findings indicate that all it takes is one person to derail the process through a counter-productive attitude or emotional state (angry, sad, fearful, or even violent). Here is a good example of how one party can stop the process. "Mediator could not keep the Respondent from attacking Charging Party verbally which continued to upset Charging Party. It was determined to end the mediation because it appeared that there could have been physical violence." Here is another instance. "Charging Parties combative attitude. She did not work to listen to the person which had direct knowledge about the dispute. This process is voluntary. Parties must be willing to participate and listen to each other and be willing to compromise. Charging Party was unwilling to compromise even if Respondent was accommodating."

The lack of respondent authority constitutes 9% of the barrier observations and is the number one barrier 4% of the time. This indicates that we see this barrier almost 10% of the time but that when we do it is the number one barrier in half the cases where it appears. This indicates that this is a particularly pernicious barrier. We find that it is most prevalent, but not exclusive to, public entities. There seems to be nothing more irritating and yet more preventable than to have a respondent at the table who cannot make the final decision. This, according to some mediators, comes close to acting in bad faith. By sitting through the entire mediation to only announce at the end that the proposal will be taken back for consideration is seen by mediators as frustrating, improper, underhanded, and sometimes unethical.

External factors are a barrier in about 6% of the cases. These factors include other lawsuits, other EEOC charges, probate proceedings, relationships and other factors external to the EEOC charge/dispute at mediation, but which exercises influence over the dispute and its resolution. External factor(s) is the number barrier in almost 3% of the cases. Similar to lack of authority, when it appears it is often the primary barrier. Other lawsuits, relationships between management and unions, developing case law, and other external forces or actors play a decisive role in whether or not the case will be resolved at the table or perhaps in the courts. The external context is an important issue for future intake procedures. Identification of such external issues may assist the mediator prepare for these tougher cases.

Miscellaneous factors also provide insight into mediation barriers. These include: the interference of legal issues/use of mediation for discovery (4%), lack of witnesses and legal representatives (5%), the role of time (3%), the respondent as a state entity¹⁶ (2%), lack of trust between the parties (2%) mediator self blame (1%), party walkouts (1%), and language/communication difficulties (1%).

Example of lack of trust includes "Lack of trust on the part of CP." Another more detailed response is "Respondent not interested in offering Charging Party anything. Charging Party had entered into a prior agreement and then filed an EEOC charge after Respondent and him signed off on another agreement."

Regarding the use of mediation for discovery, it is often not possible for the mediator to discern this intent. Acting in bad faith by using the mediation process as a form of discovery is a serious concern and charge so it is somewhat a relief to know that mediators indicate this happens rarely. Then again, this may be the instance where the parties are overt as to their motives while in other cases they are more covert. We note that the overall climate of EEOC mediation, from the point of first contact to the disposition of the case, is heavily influenced by a legalistic culture. This context may therefore help to produce more instances of bad faith discovery than in other arenas that incorporate mediation.

The role of time has appeared in the various parts of our study. We note that in many instances time is short because one of the parties, including attorneys and other representatives, only schedule two hours for the mediation or fly in for half the day and need to get back to the airport. The implication here is to advise intake personnel to have the parties' plan for at least four hours (just above the mean for the mediations). This may remedy the situation.

We find it interesting that of the 811 unresolved cases and the over 2000 responses by mediators about barriers, only 7 report their conduct was a barrier to resolution.

Table XIV
The Five Most Important Barriers to the Resolution of the Charge

Note: Total number of unresolved cases is 811. Since some mediators have given multiple reasons, the total number of responses differs from actual number of unresolved cases. The percentage given is a function of the unresolved cases and not of the responses or participants.

As discussed above, based on the analysis of the mediator comments regarding the five most important barriers to the resolu