Will Moore posted and interesting post at politicalviolenceataglance.org yesterday, and simultaneously tweeted a question to the UCDP that read: ”Do we try to bring citizen images/video into our data collection?”
The post concerned how certain news companies and human rights groups have made it a skill to analyze and verify video and stills from abuses and fighting in conflicts, and also how social media can be a good source to confirm details on the ground. In the social media age there are certainly lots and lots of video, stills and witness statements flying around that could be used in collecting the type of data that the UCDP is interested in.
But, the answer to Will’s question is “we kind of already do”. The ambiguity in this answer has to do with the fact that the UCDP, in its coding, to a certain extent relies on videos and stills at least, although mainly indirectly. We don’t have the resources to monitor all stills and videos that are out there to confirm actors, deaths or locations for our data. We do, however, rely on such data indirectly and we do at times analyze videos and stills for geographic, actor, contextual or other information.
When I say indirectly I mean that when approaching a source based a lot on videos or stills we would not have our coders delve into that material and attempt to code it. We would also not make use of that source in an indirect manner until we are clear about who the sender is, what its bias is and how reliable the information appears to be. If, after a review, such a source appears to be reliable we would tap it for as much information as possible. So, such data will enter our collection, but we would not interpret it ourselves to any great degree. And in the past couple of years the availability of such sites and such data have increased greatly, as is perhaps most visible on the sites that monitor casualties in the Syrian conflict. It is a never-ending parade of stills and vids of dead people (you literally see dead people).
We do approach vids and stills and use other ‘new’ technology when coding though. Perhaps the most important use is to attempt to verify large-scale or significant massacres, but rarely is their good footage of these events as they happen. More common is to use vids and stills to get a feel for context. We used this method widely when considering if the clashes in Thailand a few years back – between the government and the Reds and Blacks – should be included as an armed conflict. There were many videos and great stills of these clashes, which we used to try to understand the conflict dynamics and how the more violent Blacks would stand in front of Reds and attempt to induce violence from the government (or protect the Reds from government violence from another point of view, see picture below). We’ve done the same in other cases, such as in Sri Lanka and in the Israel-Palestine conflict. Visualization can be important for context. But again, this is not in the SOP’s that we use.
More common use of ‘new’ technology is the extensive use we make of Google Earth in georeferencing. Whilst not the greatest tool on Earth for such activities it has a few important uses for countries that are not well covered in gazetteers or other geodatabases. It has been a handy tool in pinpointing the exact locations of bridges, river crossings, islands, minute and isolated villages and many obscure locations. I remember using it a lot when geocoding Libya, through pinpointing remote airfields and crossroads that would otherwise have been impossible to find.
So, to sum up, we almost solely use vids and stills indirectly, through making use of other’s work (which we feel is ok, since so many make use of our final products for free) in interpreting and evaluating these sources. Should we use it more? Currently we lack those types of resources, as coding fatalities globally every year through our present method swallows all available resources (and more). However, we would gladly welcome global projects such as the ones mentioned in the post; they would make our job easier and more reliable.