New Yorker: “Election pollsters sample only a minuscule portion of the electorate, not uncommonly something on the order of a couple of thousand people out of the more than two hundred million Americans who are eligible to vote. The promise of this work is that the sample is exquisitely representative. But the lower the response rate the harder and more expensive it becomes to realize that promise, which requires both calling many more people and trying to correct for “non-response bias” by giving greater weight to the answers of people from demographic groups that are less likely to respond… A typical response rate is now in the single digits.”

“Meanwhile, polls are wielding greater influence over American elections than ever.”

“When voters go to the polls in Kentucky and Mississippi on Nov. 3, The Associated Press will be launching experiments aimed at finding more accurate and less expensive ways to survey them. With funding from the Knight Foundation, the AP has hired GfK Custom Research to identify and survey voters online, instead of conducting a traditional exit poll where precinct-based interviewers ask voters to fill out questionnaires,” the AP reports.

Cliff Zukin: “Election polling is in near crisis, and we pollsters know. Two trends are driving the increasing unreliability of election and other polling in the United States: the growth of cellphones and the decline in people willing to answer surveys. Coupled, they have made high-quality research much more expensive to do, so there is less of it. This has opened the door for less scientifically based, less well-tested techniques. To top it off, a perennial election polling problem, how to identify ‘likely voters,’ has become even thornier.”

“In all my years as journalist and strategist, I’ve never seen as stark a failure of polling as in the UK.”

— David Axelrod, quoted by the Daily Beast.

Wall Street Journal: “Founders of political wagering site Intrade are back with a new forum for electoral prognosticating, minus one big thing: the wagering.”

“Pivit, a web and mobile application, is launching prediction markets this week for the midterm U.S. elections… Pivit’s contracts are based in percentages, not dollars. Participating doesn’t require any money or pay out any rewards. Instead, the new market aims to attract people with rankings of the best predictors.”

Jonathan Bernstein lists five states “most at risk for polling mishaps” in their Senate elections, ranked from most to least likely: Georgia, Alaska, Kansas, Colorado and South Dakota.

“This doesn’t mean any of them will hold surprises once the votes are counted. It just means that right now there are reasons to look at the polls with some skepticism.”

Harry Enten: “Two weeks and two days before Election Day, and we received just two polls over the weekend. The polls, both taken in Colorado by left-leaning groups, had little impact on the FiveThirtyEight Senate forecast — Republicans have a 62 percent chance of taking the Senate. Indeed, we were more interested in the small number of polls than what those two surveys said.”

“The FiveThirtyEight model relies mostly on polls, and without polls, the forecast’s accuracy could suffer. So we looked into it. Fortunately, this weekend was an aberration. We’re getting fewer polls than we did in 2010, but we’re not getting way fewer polls.”

Sam Wang: “Yesterday, Nate Silver and I both examined Senate polling errors. He saw no overall bias; I pointed out that recent bias has been unusually large. Both statements are true. But neither of us pointed out that the biases follow a significant pattern: midterm-year polling is far less accurate than Presidential-year polling.”

“From a practical standpoint, this is good news for those of you who don’t like where things have headed lately: in midterms, Senate polling errors are five times larger than in Presidential years. There is bad news too: the error can go in either direction, and a GOP blowout is also possible.”