In The Blogs
- ‹ previous
- 4762 of 8830
- next ›
Dep't of Lies, Damn Lies, and Statistics: Iraqi Casualties Edition
Josh Marshall and the crew at Talking Points Memo are looking into this issue of casualties in Iraq — the right and the left are both throwing around statistics on civilian deaths that support their respective positions on the surge, and Marshall wants to figure out what's really going on. But until he does, you should take a look at this very good Washington Post story on why the government's and military's numbers, which they say prove deaths are down by half, are bogus.
So ? Let's not be knee-jerk liberal about this. Let's cut them some slack and say that their numbers are legit.
That means that since 3-19-03 (nearly 4 ½ years), we are down to "overall attacks in Iraq were down to 960 a week in August". That is a good thing? Are they really serious? What is really sad is there are Americans who are going to look at this as if it is progress.
It is sort of like saying that you use to smoke 10 rocks of crack a week and now you are down to 5 and that is OK. Someone's gotta' be smokin' crack to buy this crap.
So ? Let's not be knee-jerk liberal about this. Let's cut them some slack and say that their numbers are legit.
That means that since 3-19-03 (nearly 4 ½ years), we are down to "overall attacks in Iraq were down to 960 a week in August". That is a good thing? Are they really serious? What is really sad is there are Americans who are going to look at this as if it is progress.
It is sort of like saying that you use to smoke 10 rocks of crack a week and now you are down to 5 and that is OK. Someone's gotta' be smokin' crack to buy this crap.
So Let's not be knee-jerk liberal about this. Let's cut them some slack and say that their numbers are legit.
That means that since 3-19-03 (nearly 4 ½ years), we are down to "overall attacks in Iraq were down to 960 a week in August". That is a good thing? Are they really serious? What is really sad is there are Americans who are going to look at this as if it is progress.
It is sort of like saying that you use to smoke 10 rocks of crack a week and now you are down to 5 and that is OK. Someone's gotta' be smokin' crack to buy this crap.
Finally - if you are talking numbers, you got to use rules of scientific method and hypothesis testing to establish a correlation between the "surge" and reduction in violence.
First off, use the same data from the same source. That may be difficult to do since the goalpost keep moving, but that is what science/math deems an acceptable procedure.
Second, you need to look at the size of the numbers and this historical data to determine if any change in numbers is significant. That means that the drop in violence needs to be beyond both historical drops over the same time AND standard deviation of historical numbers.
My stats is a bit rusty, so this will be a good reason for me to brush up. To say "We went from 1700 attacks to 1000" may sound good, but that doesn't mean the reduction is statistically significant.
Then again, people voted for Bush because they hate gays, so I doubt many of them will take any interest in statistics.