NOTES

1 Because of space limitations, this chapter does not review the valuable contributions made by constructivist researchers to methods for studying conflict. Studies on discourse analysis and grounded theory are particularly relevant.

2 Perhaps the earliest attempt to perform a field experiment on conflict resolution is the popular rob­bers cave study performed by Sherifand his colleagues (1961). Their demonstration of the positive impact of a superordinate goal has captured the popular imagination. The problem with the study is that it was a non-replicated demonstration experiment. The value of the study is that it stimulated a debate during the cold war about alternative strategies for reducing tensions as well as considerable research on the topic. (See Johnson and Lewicki [1969] for an example of earlier research; see Brown and Wade [2006] for a recent study on the concept. The research shows both positive and negative impacts of superordinate goals on resolving conflicts.)

3 Despite the differences in methodological approach, the various studies in this volume converge on the general conclusion that there are no assurances that international mediation will lead to better outcomes or a successful conclusion of negotiations. The effects of mediation are contingent on a variety of factors in the situation as well as the broader context of the conflict (see, in particular, the chapter by Bercovitch and Regan, 2004).

4 The idea of methodological pluralism differs from cumulation in research. A key difference is that pluralism is based on methods variance, while cumulation depends to a large extent on methods invariance. The former emphasizes robustness, while the latter “adds up” results obtained from replicated studies. These different aims are also reflected in technical approaches to doing research: comple­mentary methods that address the same research question versus effect sizes calculated with meta- analytic techniques. Both approaches are concerned with the issue of generality but approach that issue from different angles. Just as replicated findings provide evidence for general effects, similar results from different methods also provide evidence for generality. However, the multi-method approach allows for the possibility of contingent findings. A number of multi-method studies illustrate both types of findings (e.g. Wilkenfeld et al., 2003). The meaning of cumulation in an age of methodological pluralism is an issue that merits more attention. It is relevant to the debate over internal versus external validity.