With less than three months to go until the election we switched over to likely voters for this week's polls and given the low level of Democratic interest in the election that not surprisingly led to more Republican leaning electorates in the Illinois and Pennsylvania polls we'll be releasing tomorrow than when we polled those states in June.
Our June Pennsylvania poll was based on a D+8 electorate, this week's poll is based on a D+2 electorate. The folks we surveyed in Pennsylvania this week voted for John McCain by 1 point, while Barack Obama actually won the state by 10 points in 2008.
We're seeing an even worse dropoff in Democratic interest in Ilinois. Our June poll there was based on a D+13 electorate, this week's poll is based on a D+9 electorate. The folks we surveyed in Illinois this week supported Barack Obama by 9 points in in 2008 in contrast to his actual 25 point victory there.
That diminishing of Democratic interest in Illinois is probably the worst we've seen anywhere so far in this election cycle but the party's ticket there is pretty clearly the worst they have anywhere in the country too so it's not all that surprising.
The dropoff in Democratic turnout these numbers suggest is similar to what happened to Democrats in Virginia last year, but worse than what they saw in Massachusetts and New Jersey where the problem was not so much Democrats not showing up as it was Democrats showing up and voting for Chris Christie or Scott Brown.
One thing I've noticed throughout this year is that Democratic interest is particularly depressed in states that have unpopular Democratic Governors and Illinois where Pat Quinn has a 23% approval rating and Pennsylvania where Ed Rendell has a 27% approval rating certainly fit the bill. Other states where we saw very low levels of Democratic interest even before switching over to explicitly polling likely voters were Michigan, Iowa, and Wisconsin where Jennifer Granholm, Chet Culver, and Jim Doyle are all among the least popular Governors in the country (if you're seeing a Midwestern theme here you should, the Big Ten states may be the most brutal of anywhere in the country for Democrats this year in terms of lost House seats and Governorships.)
One final note: we do not weight our polls for party or 2008 vote or anything like that- just fixed demographics of gender, race, and age. So the level of Democratic dropoff we show is not determined by our guesses, but by who says they're going to vote this fall and answers our polls.
We'll have the Senate numbers in both Illinois and Pennsylvania out tomorrow and the Governor numbers on Wednesday in each state.
I wonder what the excitement/interest level among Tennessee Democrats is given that Bredesen is probably among the most popular Governors in the country. Then again, we've got an impressively unimpressive Democratic nominee for the state's top job, so that may be driving mood.
ReplyDeleteAll start to poll LVs, and oops, Rasmussen isn't an outlier....
ReplyDeletePlease release the RVs next to the LVs so we can see the shift in the horserace.
ReplyDeleteAmericans are turning against Democrats and labor unions, leaving a golden opportunity for Republicans to fix the Democrats' mistakes. When Democratic policies kill the industrial economies of the Midwest, Democrats lose the ability to buy votes like they always had. And Democrats will then become powerless since they cannot possibly run on their own record or their misguided policies. All substantive issues will speak overwhelmingly in favor of Republican leadership.
ReplyDeleteAnd when likely voters are polled, it should be no surprise to learn that Rasmussen was right all along.
"The dropoff in Democratic turnout these numbers suggest is similar to what happened to Democrats in Virginia last year, but worse than what they saw in Massachusetts and New Jersey"
ReplyDeleteAfter months of PPP saying that the dropoff would not resemble VA, NJ, MA now it seems that only is true because RV were used instead of LV. Now LV polls will show Democrats losing seats from sea to shining sea. This is getting good :)
So, in other words "Kirk and Toomey Have Large Leads"?
ReplyDelete"Please release the RVs next to the LVs so we can see the shift in the horserace."
ReplyDeleteYou can just read the press releases linked at the bottom of today's posts or go to our website and find the previous surveys from June to get a decent comparison.
It's about time you guys started using a likely voter model. It will most certainly get your/PPP poll numbers closer to a real world resemblance to what might be expected this November.
ReplyDeleteI agree with Anonymous (@ 9:06 PM), in providing a numbers comparative between LV and RV results. It could be rather interesting.
relax folks you'll get the RVs from any "mainstream" MSM. of course they wont be described as such. but theres no way they will publish/tout or analyze any LV modelling. it will be impossible to lie about it. so they will continue to discuss RVs. you can always get LVs from rassmussen of course. those right wing nutbags over there, using more accurate modelling and all......
ReplyDelete"I agree with Anonymous (@ 9:06 PM), in providing a numbers comparative between LV and RV results. It could be rather interesting."
ReplyDeleteWe can't do that to begin with because our likely voter screen means we don't poll people who don't intend to vote. As Tom said, we don't fudge the numbers to get what we think the electorate will be; we simply change what we ask to get people to stay on the line who will most likely vote. Anyone (Rasmussen) who weights for party and other demographics to reflect what they THINK the electorate will look like is on dangerous territory.