After two weeks of positive polling for Romney in Colorado, which had him leading Obama in the Real Clear Politics Colorado average by 0.7%, Public Policy Polling (PPP), the polling firm for the Daily Kos, has come out with a new poll with Obama leading by 3. The PPP poll, released on Thursday evening, drove the RCP Colorado average down, giving Romney a even slimmer 0.2% edge.

The partisan breakdown for the poll shows a D/R/I split that doesn’t align with any objective version of reality.

At the end of September the voter registration statistics for Colorado were as follows:

PPP’s September polling of Colorado had Obama with a 6 point lead. The partisan breakdown for that poll actually had a representation closer to what the registration numbers looked like, even if they were still taking a few points from independents and giving them to the Democrats.

The odd thing about this shift in partisan breakdown is that Obama lost support while polling more Democrats than September. When PPP had the Democratic share at 34% in September, Obama was winning by 6 points. The Democrat share is 38% in last night’s poll yet Obama’s lead is half.

Confounding.

Now, of course the active voter registration isn’t representative of a likely voter screen, but I would say it is more accurate than the total of all inactive and active voters. That said, PPP doesn’t even conform to the total registered Democratic voters. PPP does gives the Republicans the same percentage as the actual breakdown of active voters, but PPP arbitrarily takes 5 points from the unaffiliated active voter registration and gifts it plus one to the the Democrats.

Considering the total number of active and inactive registrations show independents with a clear edge, it is again confounding that PPP would choose a model such as they have.

While 2010 was a conservative wave election across much of the country, that wave crashed on the back of a Colorado republican gubernatorial candidate who was not ready for prime time, taking down what until then was a promising senate run from Ken Buck. Despite the disaster occurring in the gubernatorial race, the Colorado General Assembly saw the Republican party take a narrow majority.

Both major parties lost voters since the 2010 election, as some became inactive and many others became independents, that said, Republicans have managed to expand their edge with active voters since 2010. One would think that this would be reflected in some way in the polling models that PPP is using this cycle, but the partisan breakdown doesn’t seem to suggest they are.

Having been on the ground and watching the ground games of both sides, I honestly don’t see the Democrats gaining all that much traction. I went to an event in Denver with Debbie Wasserman Shultz and Lily Ledbetter and the protesters nearly out numbered the attendees. Right leaning independents and Republicans that I have seen have been enthusiastic and eager to vote Romney. Many of the worries that were still there six months ago have now simply melted away, for the most part.

I know that there isn’t a direct correlation between registration numbers and actual turnout, but there are other factors to consider as well. Early voting ballot requests that Media Trackers obtained from the CO SoS show Republicans with a ~20,000 ballot lead. [Disclosure: I am currently the Communication Director for Media Trackers Colorado] Republicans requested ballots for 77% of their entire pool of active voters. Democrats have requested ballots for 81% of their entire pool of active voters.

Taking into consideration all of the factors above as well as multiple reports that Romney won Tuesday’s debate on every issue that voters rate as important this election cycle, I believe PPP is out on a limb with their polling.

DISCLAIMER: This post is my own opinion and does not reflect the opinion of my employer.