Thanks to Geoff Sharp at mediator blah blah . . . for turning us on to this Ohio State Journal of Dispute Resolution article entitled Will this Case Settle?  An Exploration of Mediators' Predictions.

Written by Harvard Law School Assistant Professor, Michael L. Moffitt, this is the best exploration of the perils of evaluative or "predictive" mediation practice I have ever seen. 

The most provocative issue addressed by Moffitt is the mediator's choice to send false signals. 

Moffitt gives four possible reasons a mediator would misrepresent his own evaluation of the likelihood of settlement:  (1)  a simple aversion to delivering unwelcome news; (2) desire for personal gain, i.e., by predicting settlement too early, the mediator could seek to extend the mediation when billing by the hour rather than the day; (3) desire to "do one's job" of being optimstic, i.e., the mediator who believes that his role is to be a "cheerleader" for resolution would predict settlement even if he thought the case were unlikely to settle; and, (4) belief that predicting success will influence the parties.

I have a lot to say about this but no time to say it today.  I invite comments from my readers on both "sides" of the medition table -- mediation advocates (litigators) and mediators.  Better yet, how about hearing from that too often silent or muffled party, the client.