2016 Postmortem
Related: About this forumStan Greenberg: Are Polls Missing Cell Phone Users?
Are Polls Missing Cell Phone Users?
Stan Greenberg: " I've seen tracking polls saying that Mitt Romney is either tied or leading in the presidential race, but we think that they are simply wrong. It's not a conspiracy theory; those other polls are just simply missing a critical segment of President Obama's coalition: cell phone users. Failing to survey those who don't have land-lines -- who tend to support the president by a significant margin -- those polls are blind to the fact that Obama is on track to win re-election on November 6."
http://politicalwire.com/archives/2012/10/29/are_polls_missing_cell_phone_users.html
ffr
(22,670 posts)That's okay, I cast my ballot for Obama-Biden, the only poll number that will matter.
Marsala
(2,090 posts)The odd thing was that it was a robo-poll that reached my cell phone. Must have been a mistake.
vi5
(13,305 posts)And everything came out pretty close to what was being polled, or at least the polling averages. I remember we hung our hopes big time on that being how Kerry was going to pull it out (and yes, I know stolen election Ohio, etc......I mean pull it out with a big enough margin).
LisaL
(44,973 posts)Which were people already voted. Then all of these states went to Bush.
Apparently Bush voters were too ashamed to admit they voted for him.
Which still makes no sense to me.
if we had the raw numbers.
But even they wouldn't be very useful. All the exit poll numbers, however preliminary, are fitted to a model.
As they honed their models they reduced the number of people polled, it made it less and less likely that their voter sample was anything near random. In other words, a small error in the model could easily turn into a large error in the results.
The models were wrong in how they altered the raw data to produce the preliminary exit polls.
CTyankee
(63,912 posts)and Chris Hayes mentioned this on his show a few weeks ago. I remember being absolutely CERTAIN that the polls were bogus because of the polling not calling cell phones and those were all Dem voters!
Please, I don't want to go down that path again, thank you...
Doctor Jack
(3,072 posts)democraticinsurgent
(1,157 posts)I remember in 2008 there not being that much difference in their behavior, but I have recently seen a poll that broke out cell-only users. Much more D friendly.
I assume that the pollsters who are not polling cell phones are making adjustments on that basis, but there's no substitute for actually polling a true random sample.
tabbycat31
(6,336 posts)JiminyJominy
(340 posts)In the past .."Oh but their poll doesn't do cell phones so it should be discounted"
It never seems to come to fruition though.