Light Seeking Light

Charles Franklin, Professor of Political Science at the University of Wisconsin takes a long, hard look at the wildly diverging polling results in the Pennsylvania Senate Race. [here]

Cutting to the chase: He finds, as illustrated in the accompanying chart, that Casey made major gains and Santorum's support declined greatly in the first half of 2005, and since that time Casey's support has declined a bit while Santorum's has risen, but Casey still has a decided edge.

The two most widely reported polls are the Quinnipiac and Zogby efforts using differing and problematic techniques. Prof. Franklin finds that both track Casey's numbers pretty well, but that Quinnipiac systematically underestimates and Zogby overestimates Santorum's support.

Prof. Franklin prefers to use a trend analysis that currently has the race at 49.7% for Casey and 40.4% for Santorum. That seems quite reasonable to me and suggests that the race is far from over. The postmortems on Santorum are premature, but despite determined efforts he has not been able to close the gap much in recent months. As the campaigns move into their final months Santorum still has a hard row to hoe, but it is not an impossible task as some have argued.

UPDATE:

Alex at the Santorum Blog is also reports this story. It is a rebuke to the "poll jumpers" who are quick to declare the race over on the basis of one poll figure.