How did the pollsters do?
On Sat Dec 29, 2007, realjohnboy wrote:Given the calendar, Nimh, it is too late too get any meaningful polls. [..]
I would repeat, the time for polls is over.
So were those last polls out indeed meaningless? Or did they after all capture some important last-minute trends that the pre-Christmas polls were not showing yet?
Now that the caucuses have come and passed, we can compare how the last polls out did in comparison with the actual results.
It's a bit unfair really; after all, pollsters never claim to
predict the outcome, just to represent the mood at the time they poll. But in reality, coming close to the actual results yields some real brownie points, and can establish your reputation - just like being far removed from them can damage it.
So let's look at the numbers, focusing first on the Democratic race. But, as always, it's not as straightforward as you might think.
First things first: Iowa caucuses, End results vs. Entrance survey
First off, we cant go on the actual caucus results, strange though that might seem, at least not in the Democratic race - because those are the end product of a process where caucusers of candidates who didnt make 15% in their precinct had to change their vote to another candidate. Whereas all the pollsters measured, the occasional additional question on second choices excepted, was first preferences.
So instead, to make the comparison, you'd need to look at whom the caucusers
came in to vote for. What their initial preference was. Thats more tricky, since the Iowa Democratic Party does not release vote counts, let alone numbers for what initial preferences were. The official result it publishes is only that of the eventual number of delegates for each candidate, and percentages derived from
that.
So here, we're basically back to a survey rather than hard outcomes. There was a so called "entrance survey", in which caucusers were asked whom they were going to vote for as they entered the precincts. This way, it also compiled a big overview of how supporters of the different candidates broke down by age, gender, location, etc. You'll already have seen numbers from this entrance survey posted by different people on different threads here -
here's the MSNBC page with the entrance poll data for the Democratic race. This entrance poll - a variation of the exit polls done in the general elections - was done by the National Election Poll (NEP) and used by all the network broadcasters. This would thus be the fairer thing to compare the poll results with, though again, it being a survey does add an element of doubt and error:
"even exit (or entrance) polls are still only polls".
The difference between the initial preferences people came in with and the end result can be quite substantive. As pollster.com recounted: "In 2004, both Kerry and Edwards did better in the reported results than the entrance poll. Most observers attribute much of the six-point gain for Edwards to a deal struck on caucus morning between the Kucinich and Edwards campaigns that sent most Kucinich supporters into the Edwards camp on the second round."
So how did it work this time? Let's put the results from the entrance poll and the official end results next to each other
Code:
MARGIN MARGIN
OBAMA HILLARY EDWARDS OTHERS OBAMA- OBAMA-
HILLARY EDWARDS
END RESULTS 38 29 30 3 9 8
(%age of delegates)
ENTRANCE SURVEY 35 27 23 15 12 8
This is an interesting table in two respects concerning things we talked about before:
- We talked a lot about where the voters would go whose initial preference was Richardson, Biden, Dodd or Kucinich, when their candidate was not viable in their precinct. These numbers seem pretty unambiguous: while the percentage for those candidates dropped from 15% of initial preferences to 3% in the end result, Edwards' picked up 7 points, Obama 3, and Hillary 2.
- I have to make a retraction of sorts. After the caucuses, I repeated several times, on different threads, that Edwards did much better than expected "or polled for". Looking again at these numbers, this is at best only partly true. Yes, despite all the volatility in the race, his second place result was a surprise. He had been stuck at third place in the polls for months, if at an increasingly small distance. But his second place finish was wholly thanks to the number of supporters of non-viable candidates walking over to his camp in the second round. In terms of initial preferences, Edwards trailed in third place with 23% of the vote - which was in fact slightly less than the polling trends had him at on the eve of the caucuses.
More like getting pestered by mosquitoes. This jerk's been following me around, making absurdly idiotic claims, and attributing them to me. And I'm getting pretty sick of it.
Then this should be a relief; I was wrong. If you really regret that decision you can get almost 50 to 1, right now, at Intrade.com :wink:
