For our final 2015 Conflict Consortium Virtual Workshop (CCVW), we focused on Prof. Yousef Munayer‘s database on Settler Violence in Israel-Palestine from 2000-2014. The data serve as the primary source for the 2012 report When Settler’s Attack (PDF), authored by Munayer. Joining us for the conversation was Will Moore, Christian Davenport, Devorah Manekin and Javier Osorio. You can watch the video here.
Munayer and a team of coders have produced an event-like data base by human content analysis of the reports produced by the Palestine Monitoring Group (PMG), a consortium of human rights NGOs distributed throughout the relevant territories, supplemented by news accounts. The resulting data are rather disaggregated at the longitude-latitude day for individual incidents with information about perpetrators, victims/targets, actions and outcomes. In some ways, it fit the standard who did what to whom, where and when format commonly known as events data, but in others it did not. Some of the discussion centered on what could be done with the existing data to convert it to the more standard events data format, thereby enhancing its potential use by academic researchers, including increasing the potential for making it readily mergeable with other sub-national datasets that covered the West Bank and Gaza Strip.
Guided by the Conflict Consortium’s Standards and Best Practices document on data creation, the session began with a discussion about what was being coded. Munayer maintains that settler violence constituted a distinct form of political violence undertaken by Israeli settlers against Palestinians in Israel proper, in places under Palestinian authority and in zones that fall under neither’s jurisdiction.
While focused on settler activities, Munayer’s data also includes activities undertaken by the Israeli and Palestinian authorities against diverse targets. This broadens the focus of the database to be more concerned with conflict and violence in the relevant territory. Rather than adopt the language or label that was more broadly targeted, however, Munayer highlighted one of the forms of violence collected. This differentiates it from databases on the region such as the Levant created by Deborah “Misty” Gerner and Phil Schrodt. It also differentiates the database from distinct research projects focused on peaceful/cooperative activities.
Discussion quickly moved to the network of PMG affiliates: geographically where were they, how does one qualify to be part of the network, are some offices/staffs larger than others, how does PMG information compare against that provided by newspapers, government reports, satellite or crowd sourcing? It was clear in this case that the comparison across sources beyond the PMG is crucial and needs to be done. This is not only to check for biases but to assess “perspectives” as it is clear that different sources would likely focus on only specific events, for specific audiences and for specific reasons (see Davenport 2007). The latter point is especially important for in this case information should not be compiled together but it should be viewed as it emerges from individual sources. At present the database did not offer this type of comparison but it could and should be done.
Another focal point of discussion concerned some interesting maps produced in the report that outlined the complexity in jurisdiction throughout the major. In different locales, Israel, Palestine and open/contested zones can be found of a dizzying variety. Highlighted in the report are arrows indicating where attacks generally came from as well as against whom they were directed. This revealed the importance of geography within the conflict and one is immediately led to wonder (as we did): why specific locales were likely to see highly aggressive settler violence whereas others were much less likely to see such activity? This is one of the questions that the data were created to prompt and facilitate, but greater attention needs to be paid to assessing the extent of the undercount bias across space, time, event type, actor and target. Academic researchers could then try to either redesign data collection to address that variation, or to build statistical models of the heterogeneity (see Conrad, Hill & Moore 2014).
We also discussed the possibility of leveraging some of the automated coding expertise of consortium members to determine whether the data might be collected that way moving forward. Munayer is very interested in collecting the data in near-real time.
Finally, we established that Munayer would like to make the data publicly available, but first needs to properly document the data collection, produce a User’s Guide, and “clean up” the somewhat messy format that he, as the producer, can work with, but would not be appropriate for other users. He shared that it was not a top priority, but something he hopes to do down the line. We expressed interest in assisting in whatever ways we could. We will provide updates as this progresses.
@engagedscholar and @WilHMoo