Quinnipiac: McCain +2 in CO; Obama +4 in MI, +2 in MN, +11 in WI

[Alcon]

http://www.quinnipiac.edu/x4141.xml?ReleaseID=1195

[Josh/Devilman88]

Obama only ahead by 2% in MN but up by 11 in WI?? Yea, and the toothfairy and bigfoot are having a tea party in Atlantis tonight.

[elcorazon]

interesting.  seems like MN and WI are either backwards or should be averaged out.  MI seems on target and CO really doesn't surprise me.  Still McCain's got a good shot there.  That said, summer polls and how good is quinnipiac?

[Sam Spade]

Wisconsin polls like this, folks.  Even when it's not summer, it's best to just pray and hope that Wisconsin polls are somewhere close to being accurate (and they often aren't).

Some of the Minnesota change looks to be among Indys, but in both CO and MN it looks to be mostly about more Republicans in the sample than the last batch.

[Verily]

interesting.  seems like MN and WI are either backwards or should be averaged out.  MI seems on target and CO really doesn't surprise me.  Still McCain's got a good shot there.  That said, summer polls and how good is quinnipiac?

They're an okay pollster, but this is the first election, I think, where they're doing serious polling outside of NY, NJ, CT and FL, so I'm not sure I trust them outside of that region. Certainly some of their numbers, especially the MN-WI ones, look problematic.

[TheresNoMoney]

I like how they're polling all the swing states.

[Wiz in Wis]

The only real shift here was in MN, even CO seems reasonable given that the last poll was taken during Obama's primary victory glow. MN seems off because everyone else thinks this is a ten point plus race there. Wisconsin and Michigan moved in their margins of error.

Ok, on second thought, I think their CO numbers must have included too many reps, there is no way that Udall and Schaffer are tied. Obama leads dems by the same margin McCain leads reps, and Obama wins independents. Too much R in the sample, same as in MN.

Having said all that.. ON WISCONSIN

[Vsanto5]

I still think that CO is democrat country this year.  Even though this state voted for the Republican in every election dating back to 1968 except of course 1992.  I really feel there is a great awakening of the younger age group, and you could almost say with the convention looming ahead, the large grassroots effort and that favors well with Obama.

[Flying Dog]

Colorado seems plausible, as does Michigan, I suppose. I think they f'ed up Minnesota and Wisconsin, though.

[Ronnie]

Again, I can't see McCain winning CO.  MI, MN, and WI are all obviously lost causes for McCain, so all of these states should be Obama country.

[Wiz in Wis]

Colorado seems plausible, as does Michigan, I suppose. I think they f'ed up Minnesota and Wisconsin, though.

I agree that Michigan is probably not off, Colorado appears plausable, but a 7 point shift away from Obama AND an 8 point shift away from Udall... smells fishy. If they released their party shares this would be academic.

Wisconsin seems right. Every other pollster has it around 10. We're an anti-war state with a candidate from a bordering state. we have leaned Obama pretty much the whole election. +11 seems pretty average to me.

[Tender Branson]

I´ve always said CO will be close, probably within 1%. On the other hand, MN looks way too close, while WI seems to become unwinnable for McCain - but Obama won't win by 11%. MI is about right.

[Dan the Roman]

In all of these samples with the exception of MN Republican outnumber Democrats. This is important because in almost every poll this year up until the last few weeks Democrats have outnumbered Republicans, many times by as much as 15 points. Whats happened is something has changed the likely voter samples to make them vastly more Republican. As a result, the polls that don't weight for party id, ie, Quinnipec, . have swung heavily Republican. Rasmussen, which uses a consistent sample, missed part of the Obama surge last month, and is missing the McCain surge right now.

If you look at the internals of these polls, almost nothing has changed except the sample, and in MI Obama is actually doing better than last month among every subgroup.

[Flying Dog]

In all of these samples with the exception of MN Republican outnumber Democrats. This is important because in almost every poll this year up until the last few weeks Democrats have outnumbered Republicans, many times by as much as 15 points. Whats happened is something has changed the likely voter samples to make them vastly more Republican. As a result, the polls that don't weight for party id, ie, Quinnipec, . have swung heavily Republican. Rasmussen, which uses a consistent sample, missed part of the Obama surge last month, and is missing the McCain surge right now.

If you look at the internals of these polls, almost nothing has changed except the sample, and in MI Obama is actually doing better than last month among every subgroup.

Hmmm....interesting. Thanks for finding that out.

[Sam Spade]

In all of these samples with the exception of MN Republican outnumber Democrats. This is important because in almost every poll this year up until the last few weeks Democrats have outnumbered Republicans, many times by as much as 15 points. Whats happened is something has changed the likely voter samples to make them vastly more Republican. As a result, the polls that don't weight for party id, ie, Quinnipec, . have swung heavily Republican. Rasmussen, which uses a consistent sample, missed part of the Obama surge last month, and is missing the McCain surge right now.

Rasmussen doesn't weigh its state polling.

Also, it is still pretty obvious in these polls that WI and MI have more people "identifying" Democrats than Republicans in the sample, it's just a smaller number than in past polls.

Having more Democrats than Republicans in your Colorado samples means that its going to be wrong long-term, but so...

[Aizen]

Eh, whatever. McCain was bound to lead in a CO poll sooner of later. McCain can still win here although Obama is still favored obviously. Unlike Iowa, Colorado is not yet a lost cause for the GOP. I don't believe the senate race is tied either although it's probably close. Rasmussen is more accurate here. On the other states... MN is completely messed up. MI seems reasonable and WI seems a tad too high.

[Dan the Roman]

I'm not saying they are wrong. I personally think a lot of prior polls were overweighting Democrats earlier this year. What I wanted to do was make the point that a lot of the shift was because the electorate polled changed and not because of the sub group, and the fact remains, earlier polls were estimating Democrats in Colorado as making up 5-7% more of the population than Republicans.

As for CO, I think the GOP advantage is still overated.

Dems  86 - 7 Obama
Indep 47 - 39 Obama
Reps   87 - 7

OK so identical margins among the bases, and Obama leads Independents by 8 but he trails overall by 2? Alright, I got something along the lines of 39% R 33% D 28% I when I tried calculating. That seems a bit lopsided.

[Aizen]

I'm not saying they are wrong. I personally think a lot of prior polls were overweighting Democrats earlier this year. What I wanted to do was make the point that a lot of the shift was because the electorate polled changed and not because of the sub group, and the fact remains, earlier polls were estimating Democrats in Colorado as making up 5-7% more of the population than Republicans.

As for CO, I think the GOP advantage is still overated.

Dems  86 - 7 Obama
Indep 47 - 39 Obama
Reps   87 - 7

OK so identical margins among the bases, and Obama leads Independents by 8 but he trails overall by 2? Alright, I got something along the lines of 39% R 33% D 28% I when I tried calculating. That seems a bit lopsided.

Indies were only 47-39 Obama? Yeah, that's certainly off.

[Dan the Roman]

I'm not saying they are wrong. I personally think a lot of prior polls were overweighting Democrats earlier this year. What I wanted to do was make the point that a lot of the shift was because the electorate polled changed and not because of the sub group, and the fact remains, earlier polls were estimating Democrats in Colorado as making up 5-7% more of the population than Republicans.

As for CO, I think the GOP advantage is still overated.

Dems  86 - 7 Obama
Indep 47 - 39 Obama
Reps   87 - 7

OK so identical margins among the bases, and Obama leads Independents by 8 but he trails overall by 2? Alright, I got something along the lines of 39% R 33% D 28% I when I tried calculating. That seems a bit lopsided.

Indies were only 47-39 Obama? Yeah, that's certainly off.

I don't think Obama's margin among indies is off. Where I differ, is that if I buy that Obama is in reality up 8 points among indies, he is almost certainly not losing the state by 2 points.

[Verily]

I'm not saying they are wrong. I personally think a lot of prior polls were overweighting Democrats earlier this year. What I wanted to do was make the point that a lot of the shift was because the electorate polled changed and not because of the sub group, and the fact remains, earlier polls were estimating Democrats in Colorado as making up 5-7% more of the population than Republicans.

As for CO, I think the GOP advantage is still overated.

Dems  86 - 7 Obama
Indep 47 - 39 Obama
Reps   87 - 7

OK so identical margins among the bases, and Obama leads Independents by 8 but he trails overall by 2? Alright, I got something along the lines of 39% R 33% D 28% I when I tried calculating. That seems a bit lopsided.

Actual registration numbers in CO at last count were 34I-34R-31D, IIRC, with Independents passing Republicans for the first time ever. Colorado is one of the few states that regularly publishes partisan affiliation statistics. (In 2004, it was 37R-32I-30D.) Of course, a fair few registered Independents will probably tell a pollster that they're registered with one party or the other, that being the party they always vote for (but may not have understood partisan registration when they registered to vote, happens all the time in immigrant communities, and not only Hispanics; see registration stats for Palisades Park, NJ for example).

[Person Man]

Well, let's see

Colorado
               O         M         Others
D 31     87         11               2
R 34     11         87               2
I  35    55          43              2

With these numbers, we get-

Obama 49.96
McCain 48.04
Others   2.00