Data were collected on public sector negotiators who bargained under the Public Employment Relations Act, a statute which does not permit strikes. Questionnaires sent to mediation users within 6 months of the mediation contained two measures of mediator effect as independent variables: mediator strategy and mediator quality; dependent variables were settlement and nonsettlement. All responses were marked on a five-point Likert-type scale, separated into union and management data groups, and then subjected to discriminant analysis. The strongest determinants of settlement for union negotiators were devising a framework for negotiations, changing expectations, and maintaining the neutrality and confidentiality of the mediator. Management negotiators responded positively to face-saving activities by the mediator, discussions of the cost of disagreement, and the mediator's expertise, trust, and impartiality. No evidence showed that testing proposals or emphasizing other settlement patterns was related to settlement in public sector mediation. Tabular data, footnotes, and 37 references.