[classicman]

If one wants a polling study conducted on a much larger scale than the Lancet one, the Iraq Family Health Survey data published in the New England Journal of Medicine last year would be a place to look. It estimated a violent death toll of anywhere from 104,000 to 223,000 between March 2003 and June 2006. Even that range of figures, however, includes a significant amount of guesswork: the authors of the study increased the number of reported deaths in the survey by 36% to compensate for estimated under-reporting by poll respondents. But unlike the Lancet’s shotgun statistics, the data upon which these estimates were made are at least available to someone other than the study’s own authors.

From UT's link - very interesting read.