The Denver Post SurveyUSA poll has some interesting demographics that appeared in their marijuana story. Looking closely at them shows why Romney is highly likely to win Colorado.

Compare the Post demographics to other sources about Colorado’s demographic makeup – focusing on Anglos, Hispanics and all other racial/ethnic groups. Pundits and pollsters use previous elections’ exit polls and US Census data. Census itself has two analyses: a “Voting and Registration” report for 2008 and the American Community Survey 2011 report.

Post Poll Exit Poll 08 Census 08 ACS Citizens ACS Adults
Anglo 75.0% 81.0% 84.7% 78.1% 73.5%
Hispanic 17.0% 13.0% 8.5% 13.7% 17.7%
All Others 8.0% 6.0% 6.8% 8.2% 8.8%

Unrealistic Share of Hispanics

The Post poll substantially undervalued the Anglo share of its sample.

SurveyUSA, the Post’s pollster, tried to match (through weighting techniques) the population shares of racial and ethnic groups and, it appears, did not exclude the 196,000 Colorado Hispanic non-citizens ineligible to vote. Given 284,000 non-citizens residing in Colorado (it’s not just Hispanics), concern about aliens voting in our election is well founded.

Add in this. Hispanic citizens are less likely to vote than Anglo citizens. For 2008 Census reported 71.5% of all adult Anglo citizens voted, while just 51.3% of adult Hispanic citizens voted. That means that, instead of 13.7% of 2012’s Colorado voters being Hispanic, it’s much more likely to be about 10% (a share Census reported in the past). NO WAY will it be 17%, the Post’s number.

For whatever reason, the Post did not report the Romney-Obama breakout by race or ethnic groups. American Research Group’s (ARG) Colorado poll (completed the day before SurveyUSA started the Post poll) has the following near mirror image breakout:

Obama Romney
White 42% 53%
Hispanic 54% 40%

IF the Post over-estimated the number of Hispanic voters (as seems highly probable), that would reduce Obama’s final results. And we’ll see something, I’d wager, like the ARG result: Romney up 4% over Obama.

Can Obama’s Ground Game Shift the Results?

DU’s poli sci prof Seth Masket, an activist Democrat, studied the question of how useful Obama’s ground game was in 2008. He says that an “Obama field office was associated with a 0.8 percentage point increase in the Democratic vote share in the county.” Giving Obama that boost would not close a 4% gap. He further found “no net boost to the [Democratic] vote” if a field office was put in a county where there had already been a field office in the previous presidential election. Given the very robust field office operation in 2008 in Colorado, it seems – per Masket’s findings  unlikely the Obama 2012 ground game could alter who wins Colorado.

If other swing states have similar polling issues, Romney – assuming he holds the gains he’s made since October 1 –  could win several other states that are now too close to call.