2016 Postmortem
Related: About this forum538 on These New Fangled Online Panel Polls
This split, however, between live-interview polls and everything else, is something we keep our eye on. When we launched our general election forecasts in late June, there wasnt a big difference in the results we were getting from polls using traditional methodologies and polls using newer techniques. Now, its pretty clear that Hillary Clintons lead over Donald Trump is wider in live-telephone surveys than it is in nonlive surveys.
We dont know exactly why live-interview polls are getting different results than other types of surveys; there are a lot of potential causes and its something well be digging into. But its harder to measure this gap in the first place than you might think; pollsters make a lot of choices whether to use a registered or likely voter sample, for example, or whether to poll Gary Johnson and Jill Stein that could account for the differences between live and nonlive surveys. But one method to get at the difference is to use our forecast models, which account for these things.
As of Tuesday morning, Clinton led Trump by 6 percentage points and had a 79 percent chance of winning, according to our polls-only forecast. But running our polls-only model using only live-interview surveys, Clinton leads Trump by 7 points and has an 86 percent chance of winning. Running it with only nonlive-interview polls, Clinton leads Trump by 5 points and has a 71 percent chance of winning.
It would appear that one of the major reasons why Clinton has been "slipping" in the polls the last week or two is because of these largely untested online panel polls and real polls with tried and true methodology are showing a much different (i.e. better) race. Sounds like the folks at 538 are having some concerns about what these online panels are doing to their numbers and are planning on looking into this issue.
http://fivethirtyeight.com/features/live-polls-and-online-polls-tell-different-stories-about-the-election/
molova
(543 posts)The Hill loves these linkbait polls. They get tons of retweets and click from their shocking headlines about wild swings.
Doctor Jack
(3,072 posts)Polls for the past 50 years have been pretty good including the last election just 4 years ago. There is no good reason why we should be throwing out the methodology and starting from scratch with these highly experimental techniques. My only guess is that these online polls are cheap to conduct, you can turn them out rapidly, and they have discovered they create that horse race feeling when one probably doesnt actually exist.
Im weary about trying to unskew polls but if 538 is raising the alarm about these and they have no established track record then maybe they should be ignored.
Cicada
(4,533 posts)Pollsters who use live human calls have to make a huge number of calls to get a thousand willing to be polled. Pollsters have to find other methods. That polls still work at all is very surprising given that nobody will answer their calls anymore.
Doctor Jack
(3,072 posts)I would like to read about it
Stellar
(5,644 posts)apnu
(8,756 posts)Its the opposite of science. And since the tried and true method of data collection isn't lining up with the pre-conceived story, they have to invent new methods of data collection in an attempt to make the data fit the story.
Online polling has never seriously tried to improve the old phone and interview methods.
Demsrule86
(68,556 posts)Doctor Jack
(3,072 posts)Maybe that wasnt the original intention but it seems that once they realized how unstable they are, they jumped all over them