The Iraqi Death Study
Echidne conducts a guided tour of the methodology they used for the study, and it is very standard stuff without any trickery or cherry-picking. The confidence level of 95% reflects that it isn’t all that the researchers might have wished for, but it is the best that can be expected under the existing conditions in Iraq.
Dave Johnson covers another major misunderstanding: the report doesn’t say that the US killed all of these people directly. It states the death rate before the invasion and the death rate since the invasion. Anyone who thinks you can have an invasion, a complete break down in law, an insurgency, and, now, a civil war without the death rate showing a significant increase needs their meds adjusted.
When you destroy the electrical grid, the water and sewage systems, the medical care system, law enforcement, and there’s a guerrilla war going on, more people are going to die than in a stable society, even if that society is repressive.
Thinks about it: if the medical system worked, fewer of those sick or injured would die; if the water and sewage systems worked, there would be less disease; if the criminal justice system worked, there would be fewer murders.
8 comments
A 95% CI is excellent. Nothing to sneeze at at that level, even if 98% would be (slightly) better.
Reasonable people who deal with statistics understand that, Michael, but if you are ready to complain you pick on every point. You can’t get a 98% CI without people dying under current conditions, so you go with the best model you can.
It’s very frustrating to read what some people are writing, because they are missing the point of the study. Some of those who think they are supporting the study seem to be claiming that the numbers represent those directly killed. Having never done any real research they don’t understand that the norm is for a larger death toll from indirect causes, like famine and disease, than direct causes.
Reasonable people who deal with statistics understand that, Michael…
I keep forgetting just how small a subset of the population that class is.
When the political class makes a point of using statistics almost exclusively to lie, it’s very difficult to find useful,valid research.
Mostly I just filter that shit out. I expect politicians to lie, or at least to obfuscate by omission, and I’m rarely disappointed on that score–though I would very much like to be.
On the other hand, I spend most of my working days, and quite a bit of my time in class, surrounded by people who do an awful lot of research involving statistics, and for whom at least a basic understanding of how they work is as elementary as needing to water a plant if you want it to keep growing. It would appear that this supersaturated environment has temporarily blinded me to the overall dearth of statistical understanding in the United States citizenry at large.
Chalk up another success for No Change Left Behind, I guess.
One of the reasons this administration refused to take numbers of the Iraqies killed is so they could refute the numbers given… and so the people back home wouldn’t be upset to learn how much “collateral damage” has been done in the quest for oil.
I had to take two different statistics courses, a general course for social science and a separate business course for my IT degree. You can’t work in business, technology, science, or much of anything else without a clear understanding of statistics.
When these people get hauled on to talking head segments and they obviously don’t understand the difference between the median and the mean, you really have to wonder why they are in anyone’s rolodex.
The people in charge have a pretty good feel for the numbers of the direct deaths, even if they bury them, but they haven’t been gathering data on the secondary deaths at all.
That, OWL, is what makes this study important, because it is looking at all of the deaths, not just those directly caused by the violence.
Millions died in the Ukraine from famine under Stalin. They were shot, but they died as a result of Stalin’s policy of exporting food for hard currency at a time when their were crop failures.