The random political thoughts thread (Part Deux)

Alex · 10-13-2006, 08:50 AM

If someone issued a study saying that the sun is green, it would be possible to dismiss it without having seen it. I'm sure General Casey has a metric assload of statistics about what is going on in Iraq and if everything he has contradicts this study as ridiculously high I'm sure he is confident in rejecting its results. Doesn't mean he is right.

I find the claim that there are death certificates for almost all of these 600,000 excess violent deaths to be hard to swallow. If they existed then the passive surveillance methods should not be missing approximately 6 out of every 7 of them because they'd be moving through regulatory channels. Maybe not in the outlying provinces (which are also the more settled provinces) but in Bagdhad where a quarter of the popluation is and is the center of violence.

I haven't read the details yet. But I know that there were significant methodological problems with the first report they did (primarily in skewed sampling) and wonder if they were addressed in this second round. Also, I wonder if they address the fact that either the last two years have seen a massive upswing or the first study was way off (implying methodolical problems that, unless corrected for, could still be in the place).

I also tend to be wary of research that has a political goal in mind at the beginning rather than the end. Doesn't mean it is wrong and the technical points need to be addressed rather than the motivations, but it raises concerns.

Alex · 10-13-2006, 10:31 AM

Here is the Fred Kaplan article* from 2004 about the first study that lays out a lot of the problems in the first study and are many of the factors I'm wondering if they were corrected for this time around.

The one that I wonder most about is the argument that the 2004 study used a pre-war mortality rate for Iraq that was 33% too low which would massively inflate the number of "extraneous" deaths.

And the 95% confidence interval for the 2004 study was 92%. That is, they said the number was 98,000 but that they were 95% confident it was between 8,000 and 194,000. A range so as to make the result almost meaningless.

I'm reading the full article now to see if these were addressed (though I early on I see that they are still using what is likely a low pre-war mortality rate).

*And just for anybody not familiar with the source Kaplan is solidly anti-Bush and pretty negative on the war.

10-13-2006, 08:50 AM	#1
Alex . Join Date: Feb 2005 Posts: 13,354	If someone issued a study saying that the sun is green, it would be possible to dismiss it without having seen it. I'm sure General Casey has a metric assload of statistics about what is going on in Iraq and if everything he has contradicts this study as ridiculously high I'm sure he is confident in rejecting its results. Doesn't mean he is right. I find the claim that there are death certificates for almost all of these 600,000 excess violent deaths to be hard to swallow. If they existed then the passive surveillance methods should not be missing approximately 6 out of every 7 of them because they'd be moving through regulatory channels. Maybe not in the outlying provinces (which are also the more settled provinces) but in Bagdhad where a quarter of the popluation is and is the center of violence. I haven't read the details yet. But I know that there were significant methodological problems with the first report they did (primarily in skewed sampling) and wonder if they were addressed in this second round. Also, I wonder if they address the fact that either the last two years have seen a massive upswing or the first study was way off (implying methodolical problems that, unless corrected for, could still be in the place). I also tend to be wary of research that has a political goal in mind at the beginning rather than the end. Doesn't mean it is wrong and the technical points need to be addressed rather than the motivations, but it raises concerns.
	Submit to Quotes

10-13-2006, 10:31 AM	#2
Alex . Join Date: Feb 2005 Posts: 13,354	Here is the Fred Kaplan article* from 2004 about the first study that lays out a lot of the problems in the first study and are many of the factors I'm wondering if they were corrected for this time around. The one that I wonder most about is the argument that the 2004 study used a pre-war mortality rate for Iraq that was 33% too low which would massively inflate the number of "extraneous" deaths. And the 95% confidence interval for the 2004 study was 92%. That is, they said the number was 98,000 but that they were 95% confident it was between 8,000 and 194,000. A range so as to make the result almost meaningless. I'm reading the full article now to see if these were addressed (though I early on I see that they are still using what is likely a low pre-war mortality rate). *And just for anybody not familiar with the source Kaplan is solidly anti-Bush and pretty negative on the war.
	Submit to Quotes