Wednesday, September 20, 2006

New Polls on Senate and Gov Races

Two sets of polls were released on September 20th by SurveyUSA and Rasmussen on the senatorial and gubernatorial races in Maryland.

The results from the two pollsters for the Senate race clash to say the least. The SurveyUSA poll places Michael Steele in the lead with 48% compared to 47% for Ben Cardin. Rasmussen has Cardin at 50%--a full seven points ahead of Michael Steele who is at 43%. The SurveyUSA poll is more at odds with other recent polls while Rasmussen is consistent with the usual finding placing Cardin in the lead.

Rasmussen has Democrat Martin O'Malley with a lead of 49% to 42% over Republican Bob Ehrlich in the gubernatorial contest. These results are very similar to the figures reported by Rasmussen over the past several months. Unlike for the senatorial contest, the results from SurveyUSA confirm Rasmussen's findings. SurveyUSA places O'Malley at 51% and Ehrlich at 44%--another seven-point lead for the Democrat.

The reason SurveyUSA shows O'Malley ahead but Cardin trailing is due almost entirely to much lower support for Cardin among black voters. While O'Malley receives support from 78% of black voters compared to 15% for Ehrlich, Cardin gets the backing of only 63% of black voters compared to 33% for Steele. SurveyUSA furhter has Cardin two points below O'Malley among white voters.

Why do the Senate polls show such different results? Survey USA included 522 likely voters surveyed from September 17-19. Rasmussen polled 500 likely voters on September 13th. Perhaps the Mfume's defeat has generated a backlash and spurred greater support for Steele among black voters. If so, Cardin could be in real trouble as Republicans almost never get much above 15-20% of the black vote, let alone 33%.

The SurveyUSA poll could also be a rogue poll. Actual public opinion will differ from the result reported in the poll by more than the margin of error for one in twenty polls. There are a lot of polls out there--some have to fall into this 5 percent. The pollsters could be weighting parties, races, or regions differently. However, one cannot compare. While SurveyUSA releases at least some crosstabs to the public, Rasmussen only allows premium subscribers a peek at them so I can only speculate about differences in the methodology.

The next few polls to be released should help confirm if the SurveyUSA result in the Senate race is an aberration or a sign of greater support for Steele among black voters, and thus the overall electorate, due to backlash at Mfume's defeat or the success of Steele's clever commercials.