The ABA Section of Dispute Resolution Task Force on Research on Mediator Techniques recently released an excellent report really worth reading. It should be of value to anyone interested in mediation. It also provides useful lessons about what we can learn about ADR from empirical research.
Superstar ADR empirical researcher Roselle Wissler is the principal author of the report, which has all the hallmarks of her meticulous work. Gary Weiner chaired the Task Force, which consisted of a remarkable cast including Alysoun Boyle, Doug Frenkel, Teresa Frisbie, Chris Honeyman, Bobbi McAdoo, Craig McEwen, Jennifer Robbennolt, Jennifer Shack, Tania Sourdin, Donna Stiensta, Beth Trent, James Wall, Howard Herman, Matthew Conger, and former members Kenneth Kressel, Dwight Golann, and Tim Hedeen.
The Task Force identified 47 studies from the past four decades with empirical data analyzing effects of particular mediator actions on certain mediation outcomes. Eight of these studies involved a process instead or in addition to mediation. “The studies covered a range of dispute types, including general civil, domestic relations, labor-management, and community mediation as well as other disputes. A majority of the studies involved court-connected mediation and a single mediator, but there was substantial variation in these and other aspects of the mediation context and mediator characteristics across the studies. … In addition to these differences, the studies also differed in whether they examined specific mediator actions or mediator approaches comprised of multiple actions; how those actions or approaches, as well as outcomes, were defined and measured; and the data sources and research methodology used. This variation contributed to differences in findings across the studies and made apples to apples’ comparisons challenging, making it difficult to draw broad conclusions about the effects of mediator actions.”
Findings About Effects of Mediator Actions on Mediation Outcomes
Here are key passages from the Report’s executive summary:
“The Task Force’s review of the studies found that none of the categories of mediator actions has clear, uniform effects across the studies – that is, none consistently has negative effects, positive effects, or no effects — on any of the three sets of mediation outcomes.” These outcome categories are “(1) settlement and related outcomes, (2) disputants’ perceptions and relationships, and (3) attorneys’ perceptions.”
“Looking at the relative potential for positive versus negative effects, while bearing in mind the substantial likelihood of no effects, the following mediator actions appear to have a greater potential for positive effects than negative effects on both settlement and related outcomes and disputants’ relationships and perceptions of mediation: (1) eliciting disputants’ suggestions or solutions; (2) giving more attention to disputants’ emotions, relationship, and sources of conflict; (3) working to build trust and rapport, expressing empathy or praising the disputants, and structuring the agenda; and (4) using pre-mediation caucuses focused on establishing trust. Some of these actions, however, have been examined in a relatively small number of studies and in only a subset of dispute types, primarily divorce, limited jurisdiction, community, and labor disputes.
“The potential effects of other mediator actions appear more mixed. Recommending a particular settlement, suggesting settlement options, and offering evaluations or opinions have the potential for positive effects on settlement and on attorneys’ perceptions of mediation, but have the potential for negative as well as positive effects on disputants’ relationships and perceptions of mediation. Both caucusing during mediation and pressing or directive actions have the potential to increase settlement and related outcomes, especially in labor-management disputes; but pressing actions also have the potential for negative effects on settlement, and both sets of actions have the potential for negative effects on disputants’ perceptions and relationships” (emphasis in original).
For a concise summary of the report, see the article in the latest issue of Dispute Resolution Magazine.
The report includes useful tables with three columns identifying studies that found negative, positive, or no effects of each of the seven categories of mediator actions on each of the three categories of effects. The report compiles each of the tables in the text into three summary tables which appear starting at page 52. Looking at these tables, you can get quick overall summaries of the findings, though this glosses over nuances in the different studies.
For example, some tables show a considerable number of studies in all three columns, suggesting that the particular category of actions has no consistent effect on the category of effect in those tables. In none of the tables with a substantial number of studies do all of the studies find that the actions have an effect, either positive or negative.
If you are a teacher, trainer, student, or practitioner, what do these results suggest that you should teach or do? Based on these empirical studies, can you be confident that mediators will produce a particular effect if they take any of the actions studied?
In short, the answer is no.
Not surprisingly, the report found that some of the more controversial actions – recommending particular settlements, offering evaluations, and pressing parties to settle – have the potential for both positive and negative effects.
In general, the studies found that some generally uncontroversial actions – such as eliciting suggestions, focusing on emotions and relationships, building trust, expressing empathy, praising disputants, and setting agendas – may or may not produce positive effects. The studies generally did not find negative effects from some of these actions, but some studies did.
So you can’t be confident that any of these actions are going to have particular effects. Rather, the effects of these actions presumably depend on numerous contextual factors such as the parties’ pre-existing relationship, history of the conflict, expectations about the process and outcome, and role of constituents, among many others.
Since the 1990s, when Len Riskin published an article featuring his original mediation grid, many mediation practitioners and academics have had intense feelings about the appropriateness and value of what he called “evaluative mediation.” This single term actually referred to a combination of very different things including assessing strengths and weaknesses of a case, predicting court results, proposing agreements, urging parties to accept a particular agreement, and trying to persuade parties to accept the mediators’ assessments. In practice, various evaluative mediation actions focus primarily on the goal of reaching agreement.
Riskin later identified numerous problems with his original grid and recommended replacing it with a series of new grids, which do a much better job of illustrating dynamics of mediation. Unfortunately, the single concept of evaluative mediation still is commonly used despite Riskin’s critique of the term, and few people refer to the new grids.
In Riskin’s original grid, “facilitative” mediation is the opposite of evaluative mediation. Facilitative mediators help parties to evaluate interests and positions, develop options, and analyze their cases, and anticipate potential court results. Focusing on the goal of promoting party self-determination, mediators identifying with the facilitative philosophy consider it to be real mediation. By contrast, they consider evaluative mediation to be an “oxymoron,” as Lela Love and Kim Kovach famously described it.
I am sympathetic with the concerns of facilitative mediation proponents. When I mediated, I tried to provide the best possible opportunity for parties to communicate and resolve their differences, but I was generally fine if they decided not to reach agreement.
I am also sympathetic with mediators who operate in attorney-dominated markets in which there is a great demand by attorneys (and some parties) for mediators to use “evaluative” techniques. In these contexts, mediators who do not use evaluative techniques and have low settlement rates may not survive in the market. More importantly, some parties greatly value evaluative techniques when done appropriately.
Having served on the ABA’s Task Force on Improving Mediation Quality, which conducted ten focus groups of civil mediators and lawyers, I remember hearing the strong feelings of people who felt that evaluative techniques are risky and problematic as well as the strong feelings of those who felt that mediation isn’t very helpful if mediators don’t use evaluative techniques.
The new Task Force on Mediation Techniques suggests that both sides are right – and wrong. Various evaluative techniques have been found to be helpful and also counterproductive. Although individual studies support each side’s views, the overall body of evidence shows that there is no consistent effect.
The upshot is that no one should claim as an empirical fact that evaluative mediation is inherently good or bad. Much depends on the context and the way that people use particular techniques. For example, the Task Force on Improving Mediation Quality identified various factors suggesting when “analytical assistance” might be more or less appropriate.
What Is Empirical Research on Mediation Good For?
If four decades of empirical research on mediation haven’t proven the general efficacy (or lack of efficacy) of specific mediation techniques to produce the effects analyzed in the Task Force Report, what the heck good is it? This complaint may be especially compelling for people who want empirical research to validate their views about particular techniques.
Indeed, some may hope that empirical research can provide the basis for the field to establish a consensus on “best practices.” That seems unrealistic for several reasons. If there isn’t a clear pattern of empirical results based on the extensive body of research reviewed by the Task Force, it seems unlikely that additional research would do so in the future. Part of the challenge is that there are so many contextual factors affecting mediation practice, that we can’t be confident that using any particular technique is likely to produce a significant benefit across a wide variety of cases. In addition, “best practices” are inherently normative professional judgments that can’t be determined by empirical research. Can you imagine that any amount of future research would resolve philosophical conflicts about the (im)propriety of “evaluative” mediation? I can’t.
Although existing empirical research on mediation hasn’t provided definitive answers, it can be very helpful in developing good theories and questions.
We can’t confidently predict outcomes because there are so many factors that affect mediation across a wide range of situations. The Task Force report includes a very good discussion of difficulties in making general causal inferences.
We can identify some factors that may be significant, particularly in certain contexts. Indeed, research may help us identify especially important contextual factors. For example, the presence and activity of lawyers (or lack thereof) is likely to have major effects on the process and outcomes. So, rather than providing foolproof recipes, empirical research may lead to helpful checklists of factors to consider.
It also can identify some factors to ignore – those that are assumed to be significant but have little or no explanatory value. One of the Task Force’s great contributions is to prompt us to reconsider some plausible but inaccurate assumptions.
Empirical research can help us discover new ideas and develop insights. For example, observing mediations and interviewing participants can lead people to learn things defying their expectations by “seeing” things they hadn’t seen before. Thus it can open people’s minds to better understand how mediations actually work and develop theories about why things work as they do. This is the goal of the Stone Soup Project.
Empirical research also is useful – and, indeed, necessary – to design particular dispute systems well. Rather than seeking to make broad generalizations, such research is tailored to describe specific situations and consider possible effects of different design choices. It is impossible to develop perfect empirical knowledge, but it can help to make plausible inferences that can help to design dispute systems.
Empirical research also can help us improve our conceptual clarity. Just as there is a “Tower of Babel” about negotiation, the Task Force report identifies a similar one for mediation as well. It noted the multiplicity of terms used to describe analogous or related behaviors and it recommended follow-up efforts to increase uniformity.
Improving conceptual clarity would be especially helpful about the thorny set of concepts related to “evaluative” mediation. Developing more uniform concepts could not only produce clearer and more comparable research results, but it also could help clarify communications among ourselves and with the multiple constituents of mediation practice. That, in turn, could contribute to more effective mediation practice.
Roselle, Gary, and all the Task Force members deserve our appreciation for moving our field another step forward.