Over at Michelle Malkin’s site (a really great site). The study purported to show that people in congressional districts that voted more heavily for President Bush also tend to have more bias against black people.
There are many criticisms that can be made of the study, even before it’s published. Hope on over to Michelle Malkin’s post on this. She quotes some criticisms of the study from a Stanford Ph.D. in statistics. They are very salient, and they are also good in a general sense – i.e., that it is very difficult and tedious to do a truly significant statistical study. The methodology has to be flawless and rigorous. It doesn’t appear to be the case with this study, but we’ll find out more when the methodology and results are fully published.
To anyone who wonders why I categorized this as “Math & Comp. Sci.”: it’s because the criticism is a very good outline that should be applied to every statistical study, regardless of content. Most people don’t understand statistics and what makes them useful, and this critique is a good way to get a clue.
Andrew Connell
Contact