Why Polling Can Be So Hard

You want me to pick up the fone when you call? Let me know who the heck you are. We get more scam, "survey", and political robo calls out here than would have been believable a decade ago that we don't even considering picking it up unless caller ID says it's someone we actually want to talk to. And 95% of the time, they won't even leave a message on our (archaic, I know) answering machine. However, I did get perhaps my favorite robo-spam message of all time on our answering machine last week: "HELLO! I'm (so-an'-so), President Donald Trump's daughter in law, calling to PERSONALLY invite you to an EXCITING event coming to your area!" What kind of people actually attend these Trump rallies, anyway?

Like tom midwest states below-- and other commentators- "Unknown Caller" and indeed, any out of area "Area Code"- gets ignored in my household. And the Times is actively calling in my Congressional District (TX 32). Same goes, whether it's my cell or landline.

In stock markets there is a mechanism called momentum. Investors are driving the price of a stock simply by following the herd. What you pollsters seem not to understand is that you are creating momentum. As readers study the polls they change opinions. Polls do not exist in a vacuum. Pollsters are shaping the elections.

Polling does a disservice to voters. They provide politicians with information as to how to pander to constituents. In some cases polls affect election results instead of predicting them. We cannot prevent polling but we can certainly make them less dependable. My personal policy is "Always lie to pollsters". Make politicians say what they believe in instead of what polls tell them you want to hear. My 2 cents

Polling is an art. Not all organizations have great polling artists, unfortunately. But there is also how some media organizations and journalists have chosen to report on the news. Nate Cohn, in December 2016, in two separate articles, wrote that the signs were there all throughout the general election. He never explained why he and others didn't report it. https://www.rimaregas.com/2016/12/26/progressives-liberals-arent-into-yo... The day after the election, Gallup interviewed older white voters in the middle of the country. Had more organizations interviewed and reported on like-minded people throughout the election cycle, maybe the Clinton campaign would have changed its strategy? Maybe not. Politico, in a very comprehensive post-mortem, published an article on the media bubble. They're right. Polling focuses too much on the coastal centers and not enough on the rest of the nation. It is no more difficult to get accurate polling in rural Wisconsin than it is in Brooklyn. But this, to an extent, flies against the mea culpas of right after the election. I guess, two years later, some want to modify the narrative. Oh, well...

NYT takes commendable care in polling. But there could be selective response rates by political preference within demographic categories and also selective turnout. Are trumpets likelier to refuse calls from untrusted media outlets and universities than dems? Privacy concerns may vary by age and rurality as the commenters here reveal. Do minority voters favoring dems succumb to voter suppression tactics, or turn out above your expectations to defy them, more so than trumpers in their groups? And,the sampling error and cumulative weights within the tiny subgroups are rather large, aren't they. What are the ranges?

Considering the onslaught of pre-recorded spam, neighbor spoofs, live money scams, and bait-and-switch caller ID calls on both landline and mobile phones, I've begun to wonder just who even answers a pollster's call. No matter how the voter registration pools are assembled beforehand, that someone answers the call at all casts serious doubt on the breadth and accuracy of the poll. It's no surprise that the recent live polling reports on this site show hundreds of thousands of calls yielding less than a thousand answers — does not seem like a representative cut of anything. And what Caller IDs do pollsters use these days, if any?

Agree that polling (and response rate) is a difficult and onerous task. Part of the problem is the readily available call screening, call blocking and caller id systems. Ours is so tight that barely 10 calls from surveys, sales or donations get through in a month. Having used telephone surveys, we, at least, when one gets through, do answer the questions. The worst the public can do is to refuse to answer. If you don't answer, your opinion is lost and ignored.

Polling is like marriage and democracy, terribly flawed but we have not yet come up with a better way to get the benefits out of all three institutions.

You pollsters need to stop. Polling generates so much feedback (in electrical engineering terms) that it has become active interference with the electoral process. The good news is that the public sees it this way. I am glad that the response rate to these live polls is 3% or less. I look forward to the day when polls are viewed with same respect as stories about four-armed babies in the National Enquirer.

Reading this discussion, especially in regard to how misleading polling in Wisconsin may have played a role in the 2016 outcome, makes me feel only more strongly that the electoral process for president needs change. I am (somewhat cautiously) hoping that the National Popular Vote Interstate Compact becomes successful, nullifying the traditional electoral college, so candidates have to campaign more broadly for a popular vote victory, and not target selected "key" states based on electoral college vote counting (info on the compact: https://en.wikipedia.org/wiki/National_Popular_Vote_Interstate_Compact). The primary process also badly needs revision, with primaries spaced more closely in time (and more simultaneously), reflecting voter distributions, not traditions of influence. And (with my statistician's hat on) I wish pollsters would not just ignore or complain about polling challenges, but more honestly report *uncertainty* in their results—or not post *any* results if they can't get a handle on the uncertainties. No candidate should be tempted to repeat Clinton's mistakes because of the hubris or ignorance of pollsters.

Nate, you're watching two congressional districts in Kansas (the 2nd and 3rd). However Kansas has an even more interesting three-way governor's race where millennials could make a big difference, locally and nationally, if only they'd vote. It would be great if you or Monmouth would take an interest in the Kansas governor's race because so far the polling has been lackluster, in the number of people called and apparently in methodology.

Maybe a better idea is to just stop polling. Then people running for office cannot ignore entire states based on an assumption of guaranteed win or loss. Similarly entire constituencies will not be ignored for a similar reason. While it was fantastic to watch Nate and 538 correctly call an election, I really do not need to worry so much in advance about what the outcome will be.

Polls are no longer reliable. The only poll worth anything is the Poll Booth when you vote on election day. I would refuse to answer a poll if contacted. Polls artificially sways voters either to not vote because they think their candidate doesn't have a chance or because they are so far ahead that one vote doesn't matter. Every vote matters. Ignore the polls. Vote.

What I believe... is that when we look back at the 2016 election (and 2018), we will see that the ubiquity and exponential nature of social media accounts and platforms, which can be echo chambers full of lies, we will see that traditional polling techniques have been rendered obsolete. (The echo chambers of lies are further amplified by the 'hell-if-I care' fact challenged propaganda spewed out on TV by Fox.) What I fear is ....that the discussions we may be having will be in Russian. (Why is there ZERO leadership from the top on informing and educating citizens to identify and call out mis-information and lies?? (rhetorical question.))

I wonder if the age of accurate polling, the ability to predict election outcomes, is coming to an end. Not because of the factors mentioned, but because when the balance of media coverage swings in one direction, people are just much less likely to contradict the conventional wisdom. We saw it with Trump and with Brexit. In both cases, the media weighed heavily in one direction and that direction was validated by the polling. The outcome however was much different.

My problem is that I get anywhere from five to fifteen phone calls daily - more on weekends - from "unknown" numbers, listings in DC, Maryland, and other cities nationwide, and most recently "local" numbers with no names. I don't answer any of them anymore, since almost all of them are or claim to be various fund-raising organizations for my registered political party that are in dire need of more, more, more money for local election campaigns, elections nationwide, and various real or fake bills that may or may be not going to local or national legislative vote. Let's not even mention the blind or scam calls for roof or home repair and non-political donations. If there are legitimate polls hidden in this endless barrage of calls, too bad - only callers that leave valid messages get a return call from me, and that happens only a few times a year at best.

Are they assuming that people who registered as Democrats and Republicans have stayed Democrats or Republicans? I registered 45 years ago. I don't remember if I specified a party or whether the question was even asked. I certainly wouldn't feel bound by the answer now.