Previous research found that violence against civilians contributes to mediation onset. With mediation taking place, the conflict parties and their behavior receive generally more international attention. Sticks and carrots used by the mediator and the international community do not only keep the parties at the table, but also move the disputants towards an agreement. We assume that continued acts of violence against civilians are not approved by the mediator. Conflict parties that continue to abuse civilians might push the mediator towards favoring their opponent. Consequently, we expect the level of one-sided violence to be smaller when mediation takes place. If the mediation process does not produce an agreement, or if the parties are dissatisfied with the talks, though, the level of violence against civilians might increase again. Whether civilian victimization remains more confined during and after mediation or whether parties discharge their discontent and try to put even more pressure on their opponent and the mediator through increased atrocities needs to be investigated systematically. We will test these expectations with time-series analyses of data on events of one-sided violence and mediation processes for selected conflicts in Africa.
