« 10 Florida Republicans face blame for recent voting issues | Main | Voting rights groups to Scott: We need to fix that -- or else »

Sizing up the pollsters: Who got it right and who didn't in prez race

Nate Silver, the wiz-kid behind the New York Times "Five Thirty Eight" blog, not only precisely called the outcome of the presidential race, he statistically nailed the performance of every state.

He has now analyzed the pollsters with a no-holds-barred accounting of their biases and average errors. For the record, the Herald/Times' pollster, Mason Dixon, didn't fare too well. According to Silver, it had a 5.4 percentage point average error rate in the eight statewide polls it conducted and it favored Republicans by a 2.2 percentage point margin. (We took a look at that here.) 

"Polls by American Research Group and Mason-Dixon also largely missed the mark. Mason-Dixon might be given a pass since it has a decent track record over the longer term, while American Research Group has long been unreliable,'' Silver wrote.

The winner, with the most accuracy measure, was TIPP, a national tracking pollster for Investors’ Business Daily, followed by -- no surprise -- Google Consumer Surveys. Because most polling firms underestimated Mr. Obama’s performance, Silver wrote, polls like TIPP's that had what had seemed to be Democratic-leaning results had the best final outcome.


Feed You can follow this conversation by subscribing to the comment feed for this post.


Gallup & Rassmussen had some of the most inaccurate polls during the last 8 wks of the Presidential election campaign. Their polls showed a Romney-Republican winning margin over Obama/Democrats of b betw 3-6 percent. According to their polls, a growing Republican wave was gaining, giving credence to the claims that the Romney campaign -- and the media, incl the Miami Herald -- heavily promoted and publicised during the past 4 wks or so. The reason? Those pollster really believed their "enthusiasm" and "likely voters" surveys that supposed showed Repubs & indeps chomping at the bit to run to the election booths and kick the Dems out of office -- while Dem voters supposedly sulking at home, dispirited at how things were going in the US. ALL -- I mean, ALL of that was serious fantasy. In fact, Dems and Dem-leaning indep voters were furious at attempts by the right wing and Repubs party to make it harder for them to vote -- and came out to vote in ENORMOUS numbers in early voting. Some web-sites -- George Mason Univ's site in particular -- documented the huge numbers -- in Ohio, almost 40% of all regis voters in Cayuhoga County (Cleveland) and Franklin County (Columbus) had voted by the day before Election Day. In FL, almost 55% of all those who had voted in 2008 had already voted by Election Day, almost all of that in Dem heavy SE FL & the I-4 corridor. In Iowa, approx 38% of ALL regis voters had already voted by the Mon before the election, and an amazing 43% of those early voters were Dems, while only 32% of those early voters were Repubs. The facts were all there for the pollsters, but most of them chose the ignore it. (Only Survey USA & PPP asked the voters that they surveyed IF they have already voted, and if they had, WHO they voted for -- and, as a result, were much closer to the actual vote totals at the end of Election Day.

MORAL for pollsters = Stop living in the right-wing, superconservative Repub alternate reality fantasy world of FOX news & AM talk radio -- and deal with the REAL world that most of us Americans live in.


it never ends !

The comments to this entry are closed.