On September 24, Marquette Law School released its presidential poll. See here. Biden 50%, Trump 40%, Jorgensen 3%, Hawkins 2%, someone else 3%, no preference 2%.
Ballot Access News
Edited By Richard Winger
On September 24, Marquette Law School released its presidential poll. See here. Biden 50%, Trump 40%, Jorgensen 3%, Hawkins 2%, someone else 3%, no preference 2%.
Look at the bottom. Democrats were oversampled. Not much of a poll.
They weren’t oversampled. There are more registered or self identified democrats than republicans. Look right above, more people called themselves very conservative than very liberal. Sampling looks pretty representative of voters overall.
Lots of Never Heard of SCOTUS hacks.
I believe many voters won’t tell strangers they’re voting for President Donald J. Trump because they’re afraid of possible violence against them.
Professional polling companies have over 70 years of experience. Their credibility is on the line; their purpose is to be as accurate as possible. Their employees have good communication skills. The polls are anonymous. It is not plausible that any poll respondent, who willingly chooses to participate in a poll, is afraid.
Would you agree that many Trump supporters aren’t sure that the person they’re speaking with works for a polling company?
I don’t know. How could anyone find out? If there were a poll of Trump supporters asking if they feel they are being tricked when someone tries to ask their opinion for a poll, they might think that new poll was also a trick.
Stress level lie detectors for both pollsters and pollstees ???
We all know Winger loves Biden and hates Trump.
We covered this yesterday. Randomized, not self selected internet polls are in line with phone polls. Bertocchi hypothesis is without evidence. Contrary to what people think the polls were pretty accurate in 2016 and 2018 as well. In the Clinton trump race the national popular vote was well within margin of error and indeed about exactly what poll averages predicted. The 80,000 votes in 3 states was too small and fast moving to sample well and swung due to late breaking developments close to election time which moved too fast to keep good track of on a state by state level. There just were not enough state polls being done often enough to keep a good eye on that. Many other races on the ballot that year and since have continued to demonstrate that polls are still pretty good, and sampling is being constantly adjusted to make up for wherever they are shown to be systematically off.
Loving Biden and hating trump has ugatz to do with it. I don’t hate either and greatly dislike both. I suspect Richard is not a big fan of either one either.
Stress level detectors for trumpets desperate for the polls to be wrong?
They’re a pretty minor factor this year and not a good reason to take polls seriously or not. And what does courts doing or not doing anything have to do with how much or little you trust polls?