Tuesday, November 23, 2004

More on Statistics

I didn't expand or explain my previous rebuttal of evildrclams attack on the figures I was using because my following argument I feel to be the height of obvious common sense (at least for such a well read intellectual as he is). There are a number of ways in which I filter statistics that I read or hear about from time to time. The number one factor in determining its relevance to me is impartiality of the original source. This I often determine by proxy with the impartiality of the source I obtain it from. I always check the given assumptions, relevant researched details as a second check, but usually I don't bother to write them down, bookmark them to pass on. Like with cash or cheques, I don't often check to make sure it's not counterfeit. I take it on trust and I might have a quick look if I don't trust the source implicitly. The secondary source I obtained this particular data was "The Economist" (what! again). Now from years of reading it, I have learnt that it is one of the most impartial magazines I have seen. There is almost no political bias, and very little country bias as well. The reason it is more able to say fairly controversial things is that every journalist signs off on every article. Thus individuals can't be targeted because they've offended one or another political entity in their articles. In this particular case, if the Economist was partisan, it would have certainly rejected this statistic. It's support for the war has been unwavering. It explained in detail how the researchers got their information. The researchers used various methods to determine exagerrations of the kind evildrclam talks about. It even rejected certain areas as being not representative because of the huge attacks in some of the randomly chosen suburbs. The most reliable statistic uncovered is the difference in death rates before and after hostilities started regardless of cause of death. The calculated violent deaths form the data is actually less reliable than the general death rate calculations. The figure of 100,000 is the extra deaths comparing death rates before and after the date the attacks started, regardless of cause. I would almost certainly have taken this data with a pinch of salt from any source which at some point indicated even a minor inclination of being against the war or the US.

1 comment:

Dr. Clam said...

'Non-partisan' is not a good description of 'The Economist'- it is, f'rinstance, one of the most virulently anti-Israel mainstream periodicals in English.