Amazing how this doesn't go with the press release.
I read through the first 4 pages. Biggest hole in the entire report? If they visited a home in their sample where no one was home, they never went back or attempted to contact them again.
So let's say in Falluja they had 30 homes. During one of the sample months 20 of those homes evacuated the city. They were never contacted again. Does this not seem to be a problem?
They also estimated 7 people per residence. Based on what? Do you know what such a high household sample estimation can do to a study if you apply that to an entire country? Lets say those 30 homes in the Falluja cluster are the sample, of those more than likely a few were unable to be contacted at some point when the city was evacuated last summer. Is it possible? Yes.
Let's further reason that the homes in Falluja reported that 10 people died in the conflict in the last year. Thats out of a sample size of 210 people (30 homes, 7 people per home.) If you have a population of 500,000 in the region divided by 7 people per or 71, 400 households. Does that mean that there were 23,800 dead in Falluja?
Without disintegrating this into a thesis about research methodology and how this study doesn't hold up.... let's ask one basic question. Is it possible for any study to be done with a reasonable accuracy without a fluctuating sample size through invasion, war, insurgency, rebuilding, evacuations, returning from evacuations? That's a pretty basic no brainer.
BOMBS GOING OFF, GUNFIRE! HELLO! ACHMED! WE'RE HERE FOR THE STUDY! BOOM! BANG!
WHAT IS YOU CRAZY! HELIDOPTERS FLY OVERHEAD, RPGH FIRE, AUTOMATIC GUNFIRE IN THE BACKGROUND.
NO! WE NEED TO KNOW IF SOMEONE DIED!
YOU'LL DIE ALLAH WILLING UNLESS YOU HIDE!
BOOM BOOM BOOM!