I notice that this (warning: PDF file) analysis by the Brookings Institute is getting a lot of press in the pro-war community.
What is interesting is that while the numbers on coalition hostile deaths are close to those of Icasualties, and the numbers on Iraqi police/military fatalities agree exactly, the civilian figures (on same page as Iraqi police/military fatalities) are totally different. As they are greater for most months, this could be attributed to better knowledge on the part of the Brookings people in many cases (Icasualties relies entirely on news reports, which likely miss some deadly incidents). However, the numbers for March from the Brookins Institute are half that of Icasualties (446 vs. 901) unless you seriously want to suggest that half of the news reports are fake or that they greatly exaggerate the death toll, the Brookings Institute's numbers for March must be bogus (perhaps they only counted the first half of the month or something).
So don't trust reports on how calm Iraq is that rely on the Brookings Index.
That is all.