The fine folks at Rasmussen Reports have dipped their toes into the political waters of Vermont. Rasmussen, as politics watchers already know, is a polling firm with a longstanding reputation for favoring Republican candidates.
And surprise: Rasmussen says the race for Governor is closer than you thought. It gives Governor Shumlin 48% and Republican Scott MIlne 36%. The survey, which combined robo-calls and an Internet component designed to capture voters who don’t have landline telephones. And it processes the results through a “weighting program” designed to, says Rasmussen, “insure that the sample reflects the overall population.”
Or, given Rasmussen’s track record, perhaps it’s really designed to “insure that things look good for Republicans.”
Anyway, the Milne campaign pounced on this bit of good news like a starving hyena on some rancid roadkill. The Milne news release compares the Rasmussen survey to a July poll from CBS News/New York Times that gave Shumlin a 25-percent lead, and concludes that the race is getting closer. Of course, comparing a real live news organization to Rasmussen is like comparing apples to wax fruit. To Milne, though, Rasmussen is a sign that “Vermonters are ready for fresh ideas” etc., etc.
But according to Nate Silver’s notoriously accurate Five Thirty Eight, Rasmussen is reliably unreliable. In its review of 2014 Presidential polls, it said this about Rasmussen:
For the second consecutive election — the same was true in 2010 — Rasmussen Reports polls had a statistical bias toward Republicans, overestimating Mr. Romney’s performance by about four percentage points, on average.
Okay, so here’s what we’ve got: a single poll with a self-described +/-4% margin of error, from a polling firm known to favor Republicans by four percentage points, that shows Milne trailing Shumlin by “only” 12 points.
I can see why Milne is excited, but this is hardly evidence of a tightening race. Let’s wait until another non-Rasmussen poll comes out.