via Business Insider:
The last few days have brought a wave of good news for President Obama in national and swing-state polling. But according to one pollster — and only one pollster — Mitt Romney is still the one with the advantage.
That pollster is Dean Chambers, the founder of UnSkewed Polls. His site, which has earned endorsements from Texas Gov. Rick Perry and The Drudge Report, has picked up steam this week, becoming a rallying cry for conservatives who cite it as proof that polls are slanted in a effort to paint a much rosier picture for Obama than reality.
“Imagine how different the narrative would be if they adhered to proportionate sampling,” Chambers said in a phone interview, pointing to the CBS/New York Times and CNN/ORC polls as the main culprits of “skewing.”
“They’re just creating the illusion that Obama is doing much better than he is. They hope it will drive voters.”
Others polling agencies, he says, are simply doing it wrong without any bad intentions. But Chambers’ main problem with the recent polls that, on average, give Obama a 4-point lead over Romney, is that they oversample people that identify as Democrats. That reflects a popular conservative consensus as of late.
What Chambers, a 45-year-old from Virginia, does on his site is “unskew” the results. He uses Rasmussen’s party registration numbers — 37.6 percent Republican, 33.3 percent Democrat and 29.2 percent Independent — and re-weighs the polling data according to those numbers. If, for example, 90 percent of Republicans vote for Romney in a poll and 5 percent go for Obama, he’ll reweigh that in accordance with the revised percentages.
There’s no happy medium: The results give Romney disproportionately gigantic leads. So a recent Reason/Rupe poll that gives Obama a 7-point lead all of a sudden becomes … a 7-point Romney lead. A Reuters/Ipsos tracking poll giving Obama a 5-point lead turns into … a 10-point Romney lead.
Chambers can see where this might be a little implausible.
“It’s only an estimate,” he said in a phone interview Wednesday. “My unskewing is an attempt to repair their badly done polling data. Romney not might actually lead by 8 percent. But he’s definitely not behind by the 7 percent they report.”
But there’s a reason why almost every polling firm does not reweigh its polling data according to party ID.
Jim Williams, a pollster from the Democratic-leaning Public Policy Polling, outlined a few:
- Party ID on polling is an attitude, not a demographic. It’s easy to change your party identification — even from one election cycle to the next. What doesn’t change is a person’s age, race and gender, demographics for which pollsters do adjust.
- Attitude can even change within the race. For example, a registered Republican disenchanted with Romney as the party’s nominee might identify as an independent.
- Williams gave this example: A voter might register as a Republican or Democrat at a younger age and then vote more for the other side as time goes on. But the voter might not change registration because it’s not a necessity and it takes time.
Are the polls being unreasonable to Mitt Romney?
Williams explained that overall, pollster
s don’t reweigh for party ID because it forces the pollster to project what he or she thinks the turnout will be, which is simply unreliable.
Chambers’ answer to that, though, is that he can reasonably predict it. He was spurred into action, he said, by a July ABC/Washington Post poll that had a plus-9-point Democratic split.
“It just didn’t look right compared to a few other polls that I’d seen,” Chambers said. “I don’t think anyone can really honestly argue that 2012 electorate is going to look like 2008.”
But the question, of course, is whether Rasmussen reflects the electorate’s party makeup as a whole in 2012.
Gallup, for example, found this year that 4 percent more people identify as Democrats than Republicans. Talking Points Memo’s database found an average of 32.2 percent Democrats and 24.4 percent Republicans over the past two years.
Chambers concedes that Rasmussen’s numbers might not be absolute. But he said it’s “common sense” to think that Democratic enthusiasm can’t match what it was in 2008. And therein lies the theory that the pollsters and media are playing the results to discourage Republicans.
“There’s no harm in questioning them,” Chambers said of polls. “No one should ever look at any polling by anyone and assume it’s completely accurate. After all, they’re just polls — they’re not perfect. The only poll that really, truly counts, of course, is the actual votes.
“Voters should realize the effort on the part of the media and some of these pollsters. The mainstream media is playing them. They should realize they’re being played.”