The Briefing Room
Briefing Room Polls (Guests Welcome!) => The Briefingroom Polls => Topic started by: Dexter on October 31, 2018, 10:17:16 pm
-
:pondering:
-
Should we take them seriously? Sure, they help to shape public opinion. Are they usually accurate? No, and less so recently.
-
When actual early voting returns are completely different than their sampling models, I'd say no.
-
Considering today is the anniversary of Dewey beating Truman, I honestlydont know.
-
Considering the number of times in the last month that I have made things up or simply pressed random buttons...
-
Considering the number of times in the last month that I have made things up or simply pressed random buttons...
And I know the numbers of the polling firms and don't answer at all, and by this time in the election alot of people don't.
-
Nobody ever asks me. But where I live is almost entirely Democrat, so I guess my opinion wouldn't count anyway.
-
I'd say that people tend to take the polls seriously when their preferred candidate is ahead, and dismiss them when their preferred candidate is lagging behind.
-
When actual early voting returns are completely different than their sampling models, I'd say no.
What early voting returns? Who gives results of early voting before election day?
@Free Vulcan
-
What early voting returns? Who gives results of early voting before election day?
@Free Vulcan
FL doesn't publish, AFAIK, how they voted, but they do publish who voted:
https://countyballotfiles.elections.myflorida.com/FVRSCountyBallotReports/AbsenteeEarlyVotingReports/PublicStats
The breakdown this year was around 42/39/19, so one might argue that any poll that oversamples one party by much is questionable (if we assume early voting is an indication of overall voting).
-
I guess we'll find out tomorrow won't we?
-
What early voting returns? Who gives results of early voting before election day?
@Free Vulcan
A number of states show early voting ballots requested and returned by party. The GOP is leading or neck and neck in most.
I don't know why people can't understand that this is not good for Dems. At all.
-
Political poll are rubbish! Let me explain. In order for these polls to have any meaning requires a probability sample. Without a probability sample no "margin of error" or any other statistical information can be done. These days the average political poll has a response rate of 0.9%. That's right....0.9%. Pew claims it's higher. It is not. These polls are self selecting samples due to the incredibly low response rates. In Wisconsin the Marquette Law School poll that is so widely acclaimed had to make over 90,000 phone call to attain 1400 responses on it's last poll before the midterm elections. This poll had a 1.4% response rate. Better than most but still rubbish. I analyzed the data.....rubbish. Nearly 50% of the responses came from just two areas while these same two areas comprise only 16% of registered voters. They claim they adjusted the data....they did it wrong. Political polls are in crisis mode.
-
Yeah, sorta. There is always a margin of error.
-
Polls don't matter at all. They have no impact whatsoever in my voting habits, and I tend to disregard them altogether. Popularity is a poor (beggardly) standard.