Questions for James
By Erick Posted in Miscellanea — Comments (6) / Email this page » / Leave a comment »
James Pethokoukis writes the Capital Commerce blog at U.S. News & World Report, my choice for a weekly news magazine. James is also a senior writer over there covering economics and U.S. economic policy.
So, here's your chance to connect to an MSM journalist who is interested in what you're thinking and wants some interaction. Leave a question for James in the comments. He's going to start answering some of the questions as we head into this final week of the elections. And what does James think about the election? Well, he's looking at a gal named Sallie for some prognosticating on the outcome.
« Today is worthy of celebration — Comments (7) | Redstate in a blue neighborhood — Comments (16) »
Questions for James 6 Comments (0 topical, 6 editorial, 0 hidden) Post a comment »
What is the right number of Democrats to sample? What is the right number of Republicans? The easy answer is to look at the numbers registered in the polling area - state, city, nation, whatever - and take that as your baseline. But not everyone turns out at an election.
What if 40% in a given state are registered Republicans, and half of them usually turn out in mid-term election and 45% are registered Democrats but only 40% of them turn out in mid-terms? Do you sample more Dems, as per the register, or more Rs as per the usual turnout.
But then what happens if the turnout changes? That was the big screw up in the 2004 exit polls. Based on previous elections the sampling was correct. But the turn out was up by a fifth, and more of the extra voters were Republicans.
Good luck in persuading the pollsters they are systematically oversampling Democrats. But it is a complex business, and they may want evidence.
Quentin Langley
Editor of http://www.quentinlangley.net
You mean predicting the wrong outcome over and over isn't evidence?
What I'd say is "Good luck convincing public poll propagandists not to do what their plan is all along."
--
If you're seeing shades of gray, it's because you're not looking close enough to see the black and white dots.
But as I recal the elections of 1998 and 2000 it was the Democrats who outperformed expectations and previous polling. Everyone was raving about how their superior GOTV, based on unions bussing people to the polls, was impossible for the GOP to match. The current consensus that the GOP has superior GOTV is based on two elections.
Quentin Langley
Editor of http://www.quentinlangley.net
Which Old Media journalist will jump off a bridge first when the GOP retains control of both houses?
You know there's a good guess or two here. They're throwing everything they have at us.
"Polls and Thier Errors" -Pejman Yousefzadeh, Red State Blog
I just cannot wrap my brain around why polling companies continue to oversample Democrats and when they are proven incorrect in an election why is it that companies/political parties continue to pay for a inferior product?
Peace through superior fire power:)