The Denver Post's John Ingold reports that Romney is gaining on Obama in Colorado in the latest poll (PDF) by Democrat polling firm PPP, yet he leaves out one glaring issue — the sample. We honestly don't understand how reporters can cover a poll without disclosing who was actually polled.
PPP's poll has Obama over Romney 49-46, up from Obama's 49-43 lead last time PPP polled Colorado. But if you were to weight the poll properly, according to a polling source of ours, Romney would actually be beating Obama by a point.
Why is that? PPP's sample of 37% Democrat, 37% Republican and 27% Unaffiliated/Other is not in line with what every other pollster who has looked at Colorado assumes will happen. That sample assumes a higher percentage of Democrats will turn out than Republicans. If you have been sticking your head in the sand since 2008 you might believe that, but everyone else knows better.
Ingold also reports that Obama is "cleaning up with women" based on a 52-44 lead. Here he is forgetting to add some context — namely that this represents the smallest lead Obama has had among women all cycle in Colorado.
For some context about the actual Colorado electorate, the active registration totals in Colorado are, as of August 1, 31.5% Democrat / 36.5% Republican / 32% Unaffiliated/Other. Secondly, 2008 was the only time in Colorado history that we can remember where more Democrats voted than Republicans, and in that wave year it was by barely less than a point more, according to Secretary of State records. In 2010, the electorate was 6 points more Republican than Democratic.
Most pollsters assume turnout will be somewhere between 2008 and 2010 — but we haven't heard from any legitimate source that Democrats will repeat anything close to 2008 turnout. PPP's sample suggests they believe something no respected observer in Colorado believes.
PPP's struggles with sampling an accurate representation of the Colorado electorate have been long documented by us at Colorado Peak Politics. A number of reporters, like Eli Stokols, usually include the requisite disclaimers about PPP's issues when covering their polls, but other reporters have not been quite so good.
Even The New York Times polling expert, Nate Silver, has found that PPP has shown a three-point oversampling of Democrats on average this year.
Next time you read a poll on Colorado, the first question you should ask is: who was polled? It matters.