Do conflict parties’ characteristics influence mediation onset? Using the veto player theory, this study addresses the question why and under what circumstances mediation occurs. The predictions made by veto player theory correspond with why conflicts are, or become more intense and longer (Cunningham 2006). I argue that the number of veto players influences the cost-benefit calculations of conflict actors, as low numbers of veto players limit concession making, while high numbers of veto players exacerbate information asymmetries. Therefore, I test the hypothesis that the relationship between the number of veto players and mediation onset probability is curvilinear, with medium numbers of veto players increasing mediation incidence likelihood. Using data on mediation onsets in civil conflicts for the period 1946-2003, I find across different statistical model specifications that low and high numbers of veto players impede mediation onset. The models explain and predict mediation occurrence well, but are sensitive to model specifications, i.e. the exclusion of observations does not allow the model to reproduce the same results. The findings confirm the explanatory value of conflict costs and the benefits of a dyadic conflict analysis approach, yet suggest that more research on conflict actors’ characteristics is necessary to understand mediation.

After the Cold War, mediation as an armed conflict resolution process was thought to become an increasingly common tool to address even the most complex armed conflicts. Recurring conflicts especially should be the focus of mediation. However, with regards to armed conflicts that reoccur, there is a gap in the research of mediation in relation to the rounds of violence. This thesis argues that the more rounds of violence a recurring armed conflict experiences the less likely it will experience mediation. To examine the relationship between the rounds of violence of a recurring armed conflict and mediation onset, this thesis employed Structured Focused Comparison (SFC) to compare two recurring armed conflicts. One with a low mediation occurrence (West Papua) and one high mediation occurrence (Aceh). While the case comparison provides limited support for the hypothesized relationship between rounds of violence and mediation, the results instead may provide support that recurring armed conflicts are not increasingly resistant to mediation the more rounds of violence they experience.