How to Judge a Poll
Over the next couple of weeks, we will be deluged by polls. Most of them will be honestly conducted. Others will constitute an attempt to discourage the voters on one side or another. How can one separate the sheep from the goats?
You should start with one key question. What are the pollsters’ turn-out expectations, and do they make any sense? Jay Cost went to the heart of the matter in his post this morning. As he put it, “the partisan composition of the electorate remains the critical unresolved issue of this cycle. Every pollster is making a guess as to what the electorate will look like, and these guesses are at least as important as their final numbers.” He should have added that these guesses determine their final numbers.
Here is an example. There is a Democratic outfit called Public Policy Polling. They recently took a look at the Senatorial race in Pennsylvania and announced that Joe Sestak was now in the lead over Pat Toomey. In his column in today’s Washington Examiner, David Freddoso isolates what the poll tells us about the expected electorate in Pennsylvania – to wit, that it will be even more Democratic (48%) than it was when Barack Obama took Pennsylvania in 2008 (44%).
Freddoso does not call out PPP for this, but he did ask Dean Debham at PPP how this could be, and this is what he reports:
“It’s a shift in those who are choosing to participate in the poll,” [Debham] said, and a sign that Democrats are now more likely to vote. “Democrats are starting to pay attention.”
Has Republican enthusiasm suddenly fallen off? “They got way out in front — and I’m seeing this around the country, where Republicans got way out in front…so they locked in their core fairly early…Democrats start pulling up as more people make up their minds.”
Well, okay, but does PPP really expect Democrats to make up a larger share of the electorate than they did in 2008? “This far out, all you’re doing is taking a snapshot,” Debnam said, adding that PPP will be polling that race again the weekend before the election.
Freddoso does not cry foul. Instead, he observes that PPP has “a good track record.” Then, he adds an oblique warning: “If Democrats turn out in record numbers in two weeks, surpassing the Obama surge, they still will.”
My bet is that, with this poll, PPP is doing what it can to encourage Democratic turn-out but that, if there is no Democratic surge (and it is hard to believe that there will be), they will soon back off. If such a surge does not show up on Rasmussen’s polling, you can be confident that my suspicions are correct.
- Comment (1)
- · Quote
- · UnfollowFollow (1)
Comments :
Jun '10
Re: How to Judge a Poll
There are liars, damn liars, and statisticians.