Monday, March 10, 2008

More on the Liberal-Conservative Rankings

Yesterday, fellow blogger El Presedente presented some good analysis of the Denver Post article on the comparative rankings of the Bob Schaffer and Mark Udall.

Any set of statistics can be made to say anything that the statistician or author wants them to say. While we thought the Denver Post article balanced and informative, it did not point out some relevant facts about the data set.

There were some very liberal congresses and thus some very liberal congressmen in the 1930's during the New Deal, in the 1960's after the Goldwater election, and in the 1970's after watergate. Ranking Mark Udall as the 509th most liberal congressman from 1937 onward is likely accurate, but it is also deceptive given the competition.

Conversely, there were very few conservative congresses in that time, so Bob Schaffer's ranking is also deceptive, though it may be less deceptive than the Mark Udall rankings.

While the Denver Post did not fully identify the data source, it did name the author. We found Dr. Kieth T. Poole's email and asked him for a link to the source data table. He was kind enough to provide it. We haven't had much time to look at the data but we expect it to be a gold mine for both campaigns.

More later.

No comments: