>Luke Weiger wrote:
> So, people, how convincing is this research?
>
> -- Luke
...........
>>The researchers said they used a widely accepted method of study
>>known as Multiple-Regression Analysis.
.>Ravi:
"in answer to your question (how convincing), a few more details are necessary. it says above that they considered the effect of many variables, but those are mostly quantitative measurable (some by
grouping) ones (median income, previous support, etc). what about the effect of pro-bush sentiment given 9/11 and the war(s)? it has been said that the e-voting areas coincide with rural areas (or areas where support has been swinging away from democrats and towards republicans). how do the results (in florida) correlate against similar results from other states? how about with other races in the same counties? "
Is not one of the problems with Multiple Regression that the factors are independent of each other? Most factors loaded into this analysis are unlikley to be so. H