What kinds of power-resources do third party interveners employ to influence adversarial parties' behavior in the course of mediation? What factors and circumstances shape the effective use of such resources? This study attempts to answer these and allied questions by examining the dynamics of leverage usage in mediation efforts to end the Northern Ghana (1994-96), Ethiopia-Eritrea (1998-2000), and Peru-Ecuador (1998-2000) conflicts.

Following the lead of Zartman and Touval (2001), the study tracks mediators' use of six sources of leverage (limitation, termination, gratification, deprivation, persuasion and extraction) and draws conclusions regarding their effectiveness within diverse conflict and mediation environments. To gain a better understanding of how leverage works, the study examines how factors such as power symmetry, conflict dynamics, disputants' interests and strategies, mediators' interests, and the interests and actions of other third-parties (independent variables) influence mediators' choice of leverage at various point of the mediation as well as the effectiveness of the selected leverage (dependent variables).

The study identifies conflict dynamics and mediators' interests as key reasons for the predominant use of persuasion and extraction and for the rarity of termination as leverage in the cases. It also finds that limitation is a catalyst for sources such as gratification and extraction, which also appear to be effective only when mediators are willing to attach appropriate conditions. In the case of gratification, this often requires mediators to be willing and able to withhold promised goods, an act that imposes huge political costs on interveners with contractual obligation to supply such goods, or those mediating in conflicts accompanied by severe humanitarian crises. With one or both of these conditions present in all three cases, it is perhaps not unusual that none of the mediators were able to employ gratification effectively as leverage.

These and other findings provide valuable insights into choice and outcomes of leverage usage in mediation. The lessons they highlight serve as key inputs for the dissertation's conclusions regarding the suitability of various types of leverage to mediators operating within specific conflict environments.