Covering only a part of war
▻http://www.washingtonpost.com/blogs/monkey-cage/wp/2014/05/27/covering-only-a-part-of-war
My comparison only focused on events that made it to the international news, but what about those that do not? What determines this selective reporting, and how does it affect the conclusions we draw based on media-based data collections of violence?
About 80 percent of the events are located within 50 kilometers of the where they actually occurred, which roughly corresponds to the diameter of the average district in Afghanistan (area of about 2,000 sq. km). Thus, we note that the spatial error of media-based event data sets is roughly in the same order of magnitude as the size of a district. In other words, we should expect the accuracy of media-based event data to be approximately correct when it comes to identifying the district where an event occurred, but not more precise than this.
Très intéressante approche de comparaison des sources.
La conclusion (très résumée…)
While we have shown that if an incident is reported, the quality of this reporting is, on average, surprisingly good, there may still be many events that are simply left out of the picture. Thus, future research has to determine how events are selected into the news and what effects this selection may have on the results we derive from media-based event dat
a.❞