The polling guy seems to be inferring that he is getting his turnout model from the D vs. R responses from this specific poll. I really don’t think this is the the standard polling methology.
The normal.scheme is to poll the R’s and D’s and I’s and determine the best estimate of how each of these 3 groups will split their vote between Obama and Romney. Then they prorate this result based on their independently developed turnout model.
Am I right about this?
“The polling guy seems to be inferring that he is getting his turnout model from the D vs. R responses from this specific poll. I really don’t think this is the the standard polling methology.
The normal.scheme is to poll the R’s and D’s and I’s and determine the best estimate of how each of these 3 groups will split their vote between Obama and Romney. Then they prorate this result based on their independently developed turnout model.
Am I right about this?”
Sort of right.
If they’re really doing it right, it all starts with the actual voter turnout numbers for previous elections, Then they’d look at voter registrations since those elections and look for trends that would change the balance. Then they’d do some polling to try to determine voter “enthusiasm” and compare it to the “enthusiasm” polls that they took before other elections and determine if there’s any trend in those polls that change the balance. Then they stir that all together with some secret sauce and they come up with the the numbers they think will match the voter turnout for this election.
BTW, they don’t just go by party affiliation. There’s a bunch of other demographics in the mix too. Age, sex, race, income, blah, blah, blah. So, for example, with Nobama on the ticket, they’ll bump their black turnout number up a bunch of points to account for him drawing more blacks to the poll. If Hillary had won the primary instead of Nobama, they would have left the black number as is and bumped the female turnout number a bit.
I think all those pollsters are hoping that Mitt is such a wishy-washy, milquetoast, barely just a step or two to the right of a flaming-liberal candidate, that he won’t draw any voters to the polls and their D+9 number will hold up.
Yes, you are correct. The standard procedure is to weight your raw results based on the predicted turnout. This interview makes it sound as if Quinnipac skips that step completely.
You set your populations however you want. You hopefully true your internals first (as Hewitt hammered home), then sample until the picture emerges (or the money runs out).
Morris did a monster DvR national LV sample in May that I have a lot of respect for:
"From May 5-11, 2012, I conducted a survey of 6,000 likely voters. On such a mammoth sample, the margin of error is less than 1 percent. I found that Romney has amassed a sizable lead over Obama of 51-42, far in excess of what published polling and surveys of registered — as opposed to likely — voters are indicating."