The latest polls show Hillary Clinton has garnered a pretty good lead in Iowa, while Bernie Sanders is again edging ahead in New Hampshire. It possible that the two candidates come out of those first two states roughly equal in delegates. But if that is true, then the third state to vote (South Carolina) seems poised to slam the door in Sanders face.
South Carolina Democrats prefer Clinton over Sanders by a 44 point margin (65% to 21%). This could easily be the state where the dream of "Feel the Bern" begins to die.
The chart shows the results of the Fox News Poll -- done between December 5th and 8th of a random sample of 364 likely South Carolina Democratic voters, with a margin of error of 5 points.
Ted: For a while I've been wanting to make this comment, not about any specific poll, but about how the whole idea of polling seems to be becoming less and less meaningful. )It's enjoyable, and I love seeing your graphically brilliant way of showing them, but political polling may be falling into some of the same problems that have long mage tv ratings enjoyable but unrelated to actual tv viewing.
ReplyDelete'
I'll gladly, arguendo, accept that polls do give accurate predictions about the 'universe' they are investigating. The trouble is that the universe is not 'the universe of all voters' or a subset like 'all Republican voters in Idaho.'
'
The 'universe' that is actually being surveyed is the universe of 'those people who have and are willing to spend the time answering a political poll.' We treat these two universes as congruent but are they? Various factors seem to be separating them more and more. (Though this year the amount of attention that, yes, Trump, has brought to the race makes the "I don't even bother thinking about politics until Labor Day" crowd much smaller than usual.)
'
But, with the increase of Caller ID there are more and more people who just won't pick up the phone from a caller they don't recognize. I don't know if people are busier these days, and less likely to spend the fifteen minutes a poll takes, but I wouldn't be surprised if they were. There are more and more Hard Right paranoids who won't answer a survey because it might really be the gummint collecting names for the FEMA camps. And elated to those are the 'home-schooled' generation, which probably actually has fewer phones than average as well as being unwilling to answer these polls. The complications of mobile phones -- where does that owner live? -- and of internet polls add to the blurriness of the congruence. And you have to have both an interest in and a knowledge about politics to even have an interest in answering the polls. (I couldn't answer a poll on movies because I love tv but rarely watch a movie made after 1950, so I couldn't pick some of the main superstars -- not to mention the 'celebrities because they are celebrities' types (I couldn't even tell you what a Kardashian IS) -- out of a line up. Glenn Close, maybe, but if you showed me Sharon Stone and Meryl Streep, I'd have to flip a coin before I guessed which was which.)
'
A person who feels like that about politics may actually, eventually, come around to vote -- in some unpredictable way -- but they aren't going to waste their own time answering a poll.
'
On the other hand, political junkies like us -- and even those on the other side -- who know the issues, know the people, maybe even remember what Chaffee and Patacki were like as Governors (bragging, but even I have to admit Jim Gilmore stumps me. Governor of Virginia, but nothing else) will be eager to answer -- if they happen to have the time at that moment, and happen to be the household member who fits the poll's requirements. But they are hardly representative of the electorate as a whole.
'
There are other factors that are even less predictable. Poll answerers are likely to have less active in-person social lives, because the calls usually come in early evening, and the active ones are getting dessed, in the shower, or already out the door -- and that might even give a clue to a socio-economic facto, since you need more money for in-person socializing.
'
All these factors skew the results in different ways, and we take it as probable that the skewing cancels out, but I have no idea if this is true. So, while I will continue to enjoy your polls, and the others I see listed, I think we have to have the salt cellar handy every time we find ourselves too tempted to take them too seriously.
You make some excellent points, Jim. We political junkies understand that a scientific poll is an educated guess, not an absolute truth -- and it is good to remind others of that.
Delete