I am currently involved in a long-term project to extend and clean the Militarized Interstate Dispute (MID) data, 1816-2010. This page describes the many related projects associated with that research program.
The first part of the project focused on collection of settlement data (agreements, exchanges, etc) for all MIDs that were originally coded as negotiated or imposed. That portion of the project was funded by the National Science Foundation (SES#0923406); the proposal outlining that work can be found here. Data collection for this project is complete, and the data will be released next academic year.
EVALUATION OF ORIGINAL MID DATA
Though not in the proposal for funding, we also used the original grant to begin a systematic evaluation of every militarized dispute in the original data set. This was necessary because we found a relatively high error rate in the outcome and settlement coding of disputes, and we wanted to confirm that our data collection contained all negotiated and imposed settlements. As part of this process, we also checked each dispute for coding accuracy, both for the militarized nature of the incident(s) and also for each individual code in the released dispute-level and participant-level data. The MID replication paper (with Steve Miller and Erin Little) is forthcoming at International Studies Quarterly and also available, with detailed appendices, here.
In a separate paper, I also point out that a large minority of dispute cases involve no actual state-to-state militarized action but instead enter the MID data through state protest. Controlling for these cases is important for proper inference and represents a better way of handling heterogeneity than the popular use of fatal/non-fatal MIDs. This paper (with Erin Little) is forthcoming at Political Science Research and Methods and available here.
The original project created a great foundation from which to collect data on the incidents that comprise each dispute. We used the initial funding to create a bibliography for each dispute and proper coding of each variable. We also compiled individual narratives describing what actually occurred in each conflict. These datasets give us the ability to greatly increase the precision of how we measure interstate conflict, by looking at the within-dispute actions of each dispute participant. My incident collection proposal has been funded by the NSF for the years 2013-2016, and we began data collection in May 2013.
Coding manual for MID incident project