Just remember, Zogby had Obama ahead by 13 in California, and in a dead heat in Ohio, on primary eve.
Take the Unholy Trinity of pollsters - ARG, PPP, and Zogby - with a large shaker of salt.
I’ve been wondering what effect the surge in young voter registration might be having on the polls. I’d hazard to guess that a large percentage of new registrations falls in that group, and a lot of them don’t have land lines. It could be a surprisingly large number that don’t show up in the polls at all.
RealClearPolitics has the pre-primary polls for all the primaries, and you can go directly to SurveyUSA’s website and get their poll breakdowns by age, sex, etc. And CNN has the exit polls on its site.
So if you’ve got a theory, this is one year for which there’s a ton of data to test it with.
Good idea, but what I was thinking about wouldn’t show up in that. There’s an awful lot of young people who have cell phones only — no land line — who might have (1) registered to vote, and (2) not been polled. They can’t use auto-dialers for cell phone polls, and so they almost always poll only land lines.
See this article from Pew:
She later explained to Bill that, when she responded to his question about lower-class white folks by saying “Screw 'em,” she was not literally suggesting or authorizing a course of action.
Pardon the hijack, and this might be better suited for ForkIII, but in regards to California, does anyone know of a place that shows the final splits for early voting? It undoubtedly benefited Clinton, but to what extent I’m curious but do not know. I have heard nothing of it since the primary nor has my Google-fu netted me any answers. I think this is one of the biggest ‘non-stories’ of the primaries, and perhaps, had California held their full primaries on that Tuesday, or even just closer to it, we might already be in the general election stages by now.
I hate to be a wet blanket, but tonight the entire Hardball panel predicted a 14 point + win for Hillary in PA. You never know until they vote, but I worry about those that will tell a pollster they’re for Obama just because they don’t want to appear racist and then marking Hillary’s box once they’re in the booth.
Are you kidding Bob? It wouldn’t be a run-up to a primary without you worrying!
Taking Hardball predictions seriously? Nah. My only reason to worry is that SUSA has usually been good and they are the ones giving her the biggest margin.
Excellent! Let a 14+ point margin become the conventional wisdom, and let Obama keep it under 10 points – then it will be perceived as a win for him.
One of these days, that sky really IS gonna fall!
Why wouldn’t it show up? Young people with cell phones wouldn’t show up in pre-primary polls, but they’d show up in exit polls. If there’s a consistent discrepancy between one and the other, you ought to be able to find it.
I don’t know anything about the official tallies, but SurveyUSA asked its CA respondents whether they’d already voted, and if so, for whom. Since they called California on the freakin’ nose (52-42 for Clinton), chances are their breakdowns are in the ballpark.
The short version is: early voters were more for Clinton than primary-day voters, but not by much. 34% of their sample on Feb. 3-4 had already voted; 66% had yet to vote. The already-voted group favored Hillary by 54-37, and the primary-day voters favored her by only 51-45. One big difference was that 7% of the early voters reported voting for someone else besides Clinton or Obama (mostly Edwards in all likelihood, though they don’t say), and that was down to 2% in the primary-day voters.
If CA had gone for Hillary by 51-45 rather than 52-42, it’s hard for me to see how that would have caused Hillary to drop out instead of stay in.
I’m talking about Pennsylvania, as in the thread title. Has Pennsylvania had an exit poll already?
So you’re asking whether cell-phone-using young people in Pennsylvania are exceptionally unlikely to show up in polls, compared to other states?
You’re right: that can’t be answered.
I’m just trying to understand why it makes sense as a question. Is there any reason under the sun to expect that that might be the case?
I’ve been wondering what effect the surge in young voter registration might be having on the polls. I’d hazard to guess that a large percentage of new registrations falls in that group, and a lot of them don’t have land lines. It could be a surprisingly large number that don’t show up in the polls at all.
They won’t show up in the land line polling, but they will show up in the stalls. As evidenced by the last several primary and caucuses where Obama did so well - against popular polling.
Stalls? Like where Republican senators hang out?
Rasmussen only has Clinton up by 3 points following the debates (poll taken 4/17). A bit of a shift from their previous 9pt difference on the 14th.
I’m trying to make sense out of the divergent polling data Most notably SUSA has been fairly consistent in their numbers - Tuesday’s numbers giving Hillary a 14 pt lead. Meanwhile the poll of polls place it less than 6. SUSA has a post which they purports addresses the polling inconsistencies but it really fails to do so. They merely point out that in three recent PA elections the pollsters erred wrong to the Republican side once, wrong to the Democratic side once, and once nailed it. One respondant noted that each error was underestimating the losing sides share whatever the ideology. But each of those times the pollsters were at least fairly clustered together.
Why is SUSA, which as has been noted here before, generally has a very good track record, a relative outlier in their call? It has been noted that they push the undecideds to decide more, but is that really it? Is their something else different about their techniques? Do the different pollsters define “likely Democratic primary voter” differently than each other? How do even define that in this primary which is closed but which had huge pushes to get independents and Republicans to register to vote Democratic early on, by Teams Obama, Clinton, and Project Chaos all?
I’m not sure why SUSA is so far out there - it’s got to be something about their sample and where they are getting their data. There are so many factors that could effect the polling, I think their polling skews one way for some reason, but that would not make too much sense because they are running blind surveys right?