A Conditional Defense of the Dyadic Approach

We contend that the dyadic approach should be employed as a theoretically informed choice.  The choice of a unit of analysis is a simplifying decision to make the real world understandable and events across time and space comparable. In this sense, the choice of the dyad or network are similar decisions, but the choice of one over the other is a function of what is most useful for answering the research question and the theory being tested. Research should not aim to model all of reality’s complexity for its own sake, but rather model the key processes under consideration. We contend that the dyad is still a useful unit choice for understanding many processes in international research, specifically those processes focused on two-actor interactions.

Despite the appropriate criticism against the assumption of independence across dyads, many conflict processes we observe are primarily between two actors. In our article, we examined how often conflicts were fought between actors.  We examined frequently used conflict datasets: the militarized interstate dispute data and interstate war data from the Correlates of War project and the armed conflict dataset from the Uppsala Conflict Data Program. Based on our count, we found that the vast majority (over 80%) of militarized disputes occur between two states, while over half of interstate wars and intrastate conflicts occur between only two actors. We also examined “linked” rivalries, which are those rivalries that have shared competitors across dyads. We looked at the rate of bilateral conflicts and multilateral conflicts in the rivalries that Israel has or had with Jordan, Egypt, and Syria. Several of their conflicts over time are linked, which is to say that many conflicts between Israel and Egypt were also being fought by Israel and Jordan or Syria. That said, the majority of conflicts between any of these dyads are bilateral. While we cannot say if these numbers are “high” or “low” without a theory of why some conflicts are bilateral or multilateral, we can say that these numbers provide primia facie evidence that thinking about conflict dyadically is a good starting point.

None of the above is to suggest that dyads are the only appropriate unit to use for studying conflict, but that for a non-trivial subset of the conflicts we observe, it may be a useful unit. If the core process a researcher is trying to understand primarily involves two actors (be they states, non-state actors, or groups of states competing against each other), then a dyad is perhaps the most appropriate unit choice. Much of conflict research’s key theoretical developments have come from this dyadic approach. Indeed, much of the game theoretic advances in conflict research have had, at their core, a two-actor game. Thus, at a conceptual/theoretical level, the dyad is a useful simplification of reality. Ultimately, our practical recommendations for research is to focus on which unit best captures the core process the researcher is trying to understand and incorporate other information (e.g. network variables or dyadic variables) when it is theoretically implied.

 

Full Text PDF   Find Replication Data
Discuss this Article
There are currently no comments, be the first to post one.
Start the Discussion...
Only registered users may post comments.
ISQ On Twitter