CO-Project New America/Keating Research (D): Obama up 4

[Tender Branson]

48-44 Obama

Project New America, formerly Project New West, is a private company that provides cutting edge tools to understand and communicate with a rapidly changing America. Through our unique model, PNA develops, conducts aggregates, and disseminates research, messaging and on-going strategic guidance with the nation’s leading progressive stakeholders. Since 2007, Project New America has conducted over 20 statewide surveys in Colorado.

Keating Research surveyed 601 Likely 2012 Colorado voters from May 21-24 on behalf of Project New America. The margin of error for the survey is +/-4.0%. Live interviews were conducted. Keating Research is a leading CO polling firm whose previous clients include Hickenlooper for Colorado and Michael Hancock for Denver Mayor.

http://www.projectnewamerica.com/PNACOPresidentialMemo5-29-12.pdf

[Tender Branson]

Keating was Hickenloopers internal pollster in 2010 and got his double-digit victory right, while many other pollsters incl. Rasmussen, PPP, Marist and Magellan showed a close race:

http://www.chieftain.com/news/local/article_8e7685e8-ea2a-11df-b62d-001cc4c03286.html

https://uselectionatlas.org/POLLS/GOVERNOR/2010/polls.php?fips=8

[Tender Branson]

I'm pretty sure that Obama has around 85% to 90% of Democrats in Colorado, because it's not a crap state like Oklahoma, Arkansas or West Virginia.

And Romney also has around 85-90% I guess.

Which means, if Obama leads by 20 among Independents, then this poll has to be made up of quite a lot more Republicans than Democrats to get an overall 4% Obama advantage.

Probably 5-10% more Republicans in that sample than Democrats.

[Tender Branson]

I'm pretty sure that Obama has around 85% to 90% of Democrats in Colorado, because it's not a crap state like Oklahoma, Arkansas or West Virginia.

And Romney also has around 85-90% I guess.

Which means, if Obama leads by 20 among Independents, then this poll has to be made up of quite a lot more Republicans than Democrats to get an overall 4% Obama advantage.

Probably 5-10% more Republicans in that sample than Democrats.

Could someone call Bawlexus?

We don't need Bawlexus for this, it's pretty obvious:

If the sample would look like the PPP sample a month ago, Obama would be up by double digits in this poll too.

No Democrat in Colorado has gotten less than 90% among Democrats since at least 2004.

Even John Kerry got more than 90%. Obama got 92%. Bennet got 94% in 2010, Hickenlooper got even 96%. PPP had Obama at 90% too a month ago.

With these numbers it's hard to imagine how Obama would lose the state right now.

Unless 10% more Republicans turn out in November ...

[Tender Branson]

I emailed Ethan Axelrod (Ethan@ProjectAmericanCentury.com) http://www.projectnewamerica.com/PNACOPresidentialMemo5-29-12.pdf to ask about the partisan sample of the poll. Here is the email I received back:

"Hi Brandon,
Can't provide crosstabs, but 37% of respondents were registered Republicans. 33% registered Democrats., and 30% unaffiliated. That's registered, not self-ID."

If anyone wants to verify the email is from who I say it is, email me at ballenwhite@gmail.com and I'll be glad to forward it to you.

So, the poll sample Project American Century is using is R+4 (37R, 33% D, 30% I). It's not hard to envision Colorado turnout looking like this in November, considering that it was R+1 in huge Dem turnout year of 2008, and was R+9 in 2004. So as you can see, the sample of this poll isnt necessarily overly GOP friendly. Seems almost spot on.

If the sample was R+4 and we assume that both Obama and Romney got 90% among their own parties, then Obama lost about 5% of Democrats to Romney and Romney lost about 3% of Republicans to Obama, while Independents went 57-30 for Obama (just like the release said).