We don’t often agree with Aurora Sentinel‘s editorial page editor Dave Perry. We’re pretty transparent about that. But the rhetorical gymnastics that Perry has to do to support a completely debunked poll from the University of Colorado are ridiculous, even for him. To revisit the poll, remember, PeakNation™, this is the poll that suggested Republican U.S. Senator Cory Gardner had just a 25% approval rating. If it were true, sure, it would be cause for alarm.
We identified last Friday our primary concerns with the poll and, subsequently, received an email from a loyal reader who identified several more issues:
- Between the first and the second poll, they changed their methodology and sample sources. As one polling expert noted, that’s “nuts.”
- 2016 exits polls were +8 Democratic, not the +14 in this poll. Actual party registration from the Secretary of State is even, Dem 31%, GOP 31%. How this poll got to Dem +14 is a mystery.
- Here’s the most bizarre: The demographics of the survey are way off — on page 29 of the report they have a breakdown of a question by race which reveals that the pollsters conducted 213 interviews with white Coloradans and 586 with minority Coloradans. That would make this survey 27% white, 73% minority. That is completely off and doesn’t match Colorado at all, much less actual voters. Latest census estimates peg CO at 69% white.
- There were two other polls in the field at the same time as this one in mid-November. One, Democratic Keating Research, found 44% with a favorable impression of Senator Gardner while 40% had an unfavorable impression. Another, from Morning Consult, conducted from October through December, found Senator Gardner with a 41% approval rating and a 36% disapproval rating.
Here’s Perry’s half-assed defense of the poll:
“To his credit, he point outs that the CU study seemed to include responses from more minorities than you might expect from a random sampling in Colorado, but it is what it is.”
Um, actually, that’s the point – it’s not “it is what it is” rather, it’s not what it’s not. This is not an accurate poll because the sample does not at all reflect the voters in the state of Colorado. Why not just poll 100% Democrats? Why even bother with Republicans or Unaffiliateds?