New polling Data. What does it mean?

Since my last post, the two polls that have been following the Connecticut Senate race (Quinnipac University and Rasmussen) have released new polling data pertaining to the current state of the race. In this post, I will present this new data as well as try to explain the significance.

The Rasmussen Poll released its new data on September 9. They conducted a poll of 500 Likely Voters in the state of Connecticut. The results showed Mr. Blumenthal leading the race with 53% of the vote opposed to his opponent Ms. McMahon who garnered 44% in the new poll. This is a 2 point shift in favor of the Democrat who only led by 7% (47-40) in the previous poll conducted by Rasmussen. The shift isn't extremely large, but it does put Blumenthal over the 50% mark for the first time in a couple months which definitely is significant because it is an upward shift, which as I stated in my last post, the shifts are more important than the raw numbers.

The poll that has received the most attention in the media is the Quinnipac University poll that was conducted between September 8 and September 12. The Quinnipac Poll shows Blumenthal leading McMahon by a margin of 51%-45%. This is a shift that is in favor of Ms. McMahon being that she was down 10% in the last poll conducted by the University. However, the most recent poll was conducted among 750 Likely Voters in Connecticut as opposed to their last poll that was conducted amongst Registered Voters.

The change in the sample from Registered voters to Likely voters is one that many would argue favors the Republicans in a year that they are poised to recieve a net gain in both houses of Congress. Therefore, it could be said that the most recent Quinnipac poll may be more accurate come November than Rasmussen's poll, but juxteposed to their previous poll (Blumenthal up 10%) it is irrelevant because of the difference in sample. On the other hand, Rasmussen's poll was conducted among Likely Voters in both cases, therefore it is more reliable to note the shift. The only way we will know if any sort of shift has occurred like Quinnipac suggests is to wait for their next poll conducted among Likely Voters.