Public Policy Polling is out with its latest Colorado poll on the presidential race. Properly analyzed – that is, adjusting for the pitiably bad sample – this poll shows a statistical tie between Romney and Obama. Remember, a bruising Republican primary has just ended, and Republicans haven't had time, yet, to kiss and make up.
The PPP poll was promoted by Curt Hubbard, the Denver Post Editorial Page Editor (how fitting!) under the title “Support for Obama Building…” Hubbard even claimed he dug “deeper into the poll results.” Not deep enough, oh fellow nerd! (That's his word for himself, and it fits me, too.)
So, what's wrong with PPP?
First, PPP customarily overstates Democratic voting strength – as others have reported in the past. This poll splits partisans 36% D, 35% R and 30% Independent. Colorado's Republicans are currently up 5.5% among active voters over the Democrats. In the Democratic landslide year of 2008, Colorado's exit poll had Republicans up by 1%. Ask yourself if Democrats are really as enthusiastic in 2012 as they were in 2008? (National data say “no.” PPP doesn't claim, or even mention, voter enthusiasm in this poll or its commentary, despite Hubbard's statement about it in his write-up.)
Second, this poll substantially overstates the Hispanic share of Colorado's voters. In 2008, US Census (Table 4b) reported that Hispanics were 8.4% of Colorado's voters; PPP has them at 14%. They show Hispanics giving Obama 53%, Romney 38% with 10% undecided. If accurate, that means Romney already has exactly the vote share that McCain got in 2008 while Obama is down 8%. The Velasquez Institute suggests, based on Census data, that Hispanic turnout will be down one million nationally in 2012. We, and PPP, should be cautious about accepting a radically larger turnout.
But the real sampling boo-boo shows up in the “Ideology” category. PPP reports 35% of Coloradans are liberals. Maybe on Happy Thursday on the Boulder Mall … when college kids ride their bikes topless and mostly bottomless. But not in real Colorado.
Gallup reports the following ideological split for Colorado: 22.4% Liberal, 34.5% Moderate and 40.2% Conservative (some “no response” here). Using the Gallup ideological split and accepting the “very” liberal and conservative outcomes reported by PPP for all liberals and conservatives, here's what the adjusted PPP poll shows:
- Obama 44.3%
- Romney 45.3%
- Undecided 7.6%
'Nuf said.
Thank you for your analysis of this poll. My 12 year old has been looking at the Electoral College map on Real Clear Politics, and has been quite dismayed to see the most recent poll numbers for Colorado. She hasn’t had a whole lot of statistics yet in school, so polling samples are still a new concept to her. I have pointed out to her that PPP has a D behind its name for good reason, but actually being able to show her numbers as to why those PPP numbers are not right will help explain it to her much better.
It would also help if there was more than one polling group in this state. I would like to see how Rasmussen poll numbers would look. Until then, thanks.
but the sample is fair.
Colorado is a centrist state, and Rethug radicalism is either scaring the hell out of the middle, or pissing them off for focusing on reproductive rights instead of the economy. Say what you will about the Dems, but at least they’re focused on the things Americans care about the most.