Quote:
Originally posted by KSig RC
I didn't see the piece, obviously, but the 'weighting' referenced here (as well as contextualization) is a heated research topic in every field. For instance, weighting in my field has been disallowed because Bayesian workup of numbers adds yet another bias to the figures.
These are tenuous bases for attack on studies; the empiricism behind the numbers is always up for discussion, no matter how they're done. This is the real world of statistics.
|
Well to be honest a lot of the stuff went over my head in the small segement... mostly dealing with the mathematical model used. Basically the host/commentator argued that without weight or context, using left/right sources becomes problematic to say the least - sorta a looks good on paper, but not in practice. When they refered to weight they ment how left/right of centre was the source they subject was citing... which then gets into the prickly arguement of where centre is... something that I didn't (and the Fifth Estate) want to touch.
What I found interesting is that say if I was a reporter, and I was doing a piece say on the Iraq War - if I used sources such as Amnesty International and the Department of Defense once each my piece would balance according to the formula - even if the piece was an article attacking one side or the other in the arguement over body counts or prisoners of war... that is where I found the application of the Groseclose-Miylo study problematic when analyizing the media (or politics for that matter).
Threaded Mode