Mike Spagat sends in an interesting explanation for the noted problems with conflict mortality studies (a topic we’ve discussed on occasion on this blog). Spagat writes:
This analysis is based on the fact that conflict violence does not spread out at all uniformly across a map but, rather, tends to concentrate in a few areas. This means that small, headline-grabbing violence surveys are extremely unreliable.
There is a second point, based on the work of David Hemenway which you’ve also cited on your blog. Even within exceptionally violent environments most households will still not have a violent death. So a very small false positive rate in a household survey will cause substantial upward bias in violence estimates.
I really find it concerning when people used flawed data, because that usually means you’ll get flawed conclusions. Though we live in a world where political science research is rarely used, the idea that a flawed paper may end up being used and leading to horrible results terrifies me. Thanks, Andrew, for publishing this paper…but do you think it is ever possible to have non-flawed data or reasonable reliable data, from which we could do “statistical modeling”?
Er, I made a typo. I meant to type, “Thanks, Andrew, for directing me to this paper.”