The Experiments
NMB programs provide an integrated dispute resolution process including face-to-face mediation processes typical of many in person mediation processes. We are using a rigorous process definition language to development and exhibit the first rigorous definition of the process used by NMB. We have been using the Little-JIL process formalism as the basis for this definition. Little-JIL is one of a family of process definition languages that are defined through formal semantics that create the possibility of process definitions that demonstrate precision and rigor. Little-JIL employs graphics and abstraction thereby also improving clarity and comprehensibility. A Little-JIL process is represented as a hierarchical decomposition of process steps with a graphical syntax. The interface includes declarations of the “agent” (here a mediator, the disputants, or a computer) that is to carry out the step, resource requirements of the step, exceptions that may take place during step execution, and messages that may be sent from this step to another step.
Using the process model as a specification, we have first concentrated on a prototype tool for supporting a brainstorming process that is at the core of the NMB’s grievance mediation method. With the formal NMB process model serving as the requirements, the ODR system we will demonstrate will serve the specific governmental and public clients of NMB.
The ODR system has been demonstrated on numerous occasions and subjected to several rounds of experimental evaluation. Three University classes on the University of Massachusetts Amherst campus have simulated work as parties in disputes using an early prototype of the ODR system. NMB professional mediators have used this ODR system in simulations based upon cases used to train the NMB’s parties in traditional face-to-face mediation, and an online University class with members across the U.S. has used the ODR system to run an extended asynchronous mediation case study. From these we have begun to see guidance for the future development of ODR systems and the methods of developing them.
At a high level, we have found that participants have little trouble adopting the online tool, effectively generating many ideas and moving fairly directly towards solutions. At the same time, the volume of text generated has presented a challenge to effectiveness. Participants and instructors noted that the anonymity of posts lowered inhibitions for questionable contributions. In addition, inhibitions for creative engagement in brainstorming were also lowered – and this enhanced the quality and content of the posts. We have received suggestions on enhancements to the process that are either required by the move to online support or now made possible because of the online support. Suggestions of the first sort range from screen layout to text edit functions. The new opportunities include the possibility of concurrent discussions and support for reorganizing lists of ideas.
We are currently analyzing the initial responses to the brainstorming tool. We are evaluating the trade offs of expanding the functionality of the system. At the same time, we are exploring ways to realize the promise of the computer as a Fourth Party to fully “assist in identifying and evaluating interests, options and solutions”.
We also have seen the need to quickly connect the online systems operation so that it is driven directly by the underlying grievance model. We are concerned that rapid changes made to the online system will violate in either spirit or letter the process NMB has exposed. For example, the disputants make commitments to the privacy constraints they will maintain in any mediation. While our process models might demonstrably enforce these privacy constraints, if our prototype tool is not directly driven by that model, the tool might allow these commitments to be violated. This can be destructive of the trust that the NMB is building up. By constructing the online system so that it is driven directly by the process model, we should be able to identify promptly the presence of such potentially destructive characteristics and to focus our efforts to deal with them by modifying the process model, and hence the operational tool