snafu Posted September 1, 2020 Posted September 1, 2020 4 hours ago, Tiberius said: It's a good poll. This isn't 2016. Sure, favorability measurement is a great poll. I view Biden favorably -- like my old grampa. Doesn't mean I want to vote for old gramps because I don't think old gramps is really cut out for the job.
Tiberius Posted September 1, 2020 Author Posted September 1, 2020 Just now, snafu said: Sure, favorability measurement is a great poll. I view Biden favorably -- like my old grampa. Doesn't mean I want to vote for old gramps because I don't think old gramps is really cut out for the job. Voters think he is. And no Jill Stein this time, that helps even more
The Frankish Reich Posted September 1, 2020 Posted September 1, 2020 8 hours ago, SectionC3 said: This is a really good point. “The polls” were late to the Trump undercurrent in 2016. I think they might have the opposite problem this time through. I suspect that the universe of likely voters is too small in a lot of these polls (because we’re going to have heretofore unseen turnout), and the “unexpected voters” will break heavily for Biden. There really wasn't an "undercurrent." For all the grief they've taken, Nate Silver's 538 blend of polls came awfully close to spot-on for the national popular vote: Clinton ahead by 3.6 percentage points. https://projects.fivethirtyeight.com/2016-election-forecast/ The actual result: Clinton won the popular vote by 2.1 percentage points. That's damn good forecasting. Obviously the state-by-state results were different, but you're dealing with smaller sample sizes and more variability there, so it's always a wild card. So where we stand today is almost precisely where we were on election eve in 2016: Biden (as with Clinton) has about a 70% chance of winning, Trump 30%. That's not overwhelming odds in Biden's favor; 538 says he's "slightly favored." Could Trump pull off that inside straight again? Absolutely. Is it likely? No. Too many ways for it to go wrong. 3 1
SectionC3 Posted September 1, 2020 Posted September 1, 2020 (edited) 2 hours ago, The Frankish Reich said: There really wasn't an "undercurrent." For all the grief they've taken, Nate Silver's 538 blend of polls came awfully close to spot-on for the national popular vote: Clinton ahead by 3.6 percentage points. https://projects.fivethirtyeight.com/2016-election-forecast/ The actual result: Clinton won the popular vote by 2.1 percentage points. That's damn good forecasting. Obviously the state-by-state results were different, but you're dealing with smaller sample sizes and more variability there, so it's always a wild card. So where we stand today is almost precisely where we were on election eve in 2016: Biden (as with Clinton) has about a 70% chance of winning, Trump 30%. That's not overwhelming odds in Biden's favor; 538 says he's "slightly favored." Could Trump pull off that inside straight again? Absolutely. Is it likely? No. Too many ways for it to go wrong. I agree except for the undercurrent. I worked the streets in 2016. I had a bad vibe in September. I sensed the persuadables were going to break for Trump. Close the book on the Clintons, I thought, was a great line at the end. but otherwise, completely agree. Edited September 2, 2020 by SectionC3 1
Unforgiven Posted September 1, 2020 Posted September 1, 2020 POLITICS Polling Firm Rasmussen Warns About “National Poll Suppression” From August 1-15 there were 16 polls that came out. From August 16-31 there have been only 8 and none covering a later date than the 25th. If you fully believe in polls like the dimwitted op. 1
Tiberius Posted September 2, 2020 Author Posted September 2, 2020 27 minutes ago, Unforgiven said: POLITICS Polling Firm Rasmussen Warns About “National Poll Suppression” From August 1-15 there were 16 polls that came out. From August 16-31 there have been only 8 and none covering a later date than the 25th. If you fully believe in polls like the dimwitted op. Oh, you think it’s a conspiracy to hurt Trump with ? fake polls?
B-Man Posted September 2, 2020 Posted September 2, 2020 34 minutes ago, Unforgiven said: From August 1-15 there were 16 polls that came out. From August 16-31 there have been only 8 and none covering a later date than the 25th. If you fully believe in polls like the dimwitted op. NO polls in the last SEVEN days ? Yeah, thats not a giveaway. 3
Taro T Posted September 2, 2020 Posted September 2, 2020 1 hour ago, Unforgiven said: POLITICS Polling Firm Rasmussen Warns About “National Poll Suppression” From August 1-15 there were 16 polls that came out. From August 16-31 there have been only 8 and none covering a later date than the 25th. If you fully believe in polls like the dimwitted op. Ummm, RCP is showing 4 GE polls on August 26 including Rasmussen w/ Biden +1. It also shows a GE poll yesterday w/ the typical Biden lead & an Emerson GE poll w/ Biden +2. Not sure why your info is showing nothing since 8/25.
Brueggs Posted September 2, 2020 Posted September 2, 2020 1 hour ago, Tiberius said: Oh, you think it’s a conspiracy to hurt Trump with ? fake polls? Do you know of a better way to make people think a certain candidate is going to lose, so that people decide not to bother voting because it won't make a difference? 1
Tiberius Posted September 2, 2020 Author Posted September 2, 2020 General Election: Trump vs. Biden USA Today/Suffolk Biden 50, Trump 43 Biden +7 General Election: Trump vs. Biden Grinnell/Selzer Biden 49, Trump 41 Biden +8 President Trump Job Approval Grinnell/Selzer Approve 43, Disapprove 51 Disapprove +8 President Trump Job Approval Politico/Morning Consult Approve 42, Disapprove 57 Disapprove +15 2020 Generic Congressional Vote Grinnell/Selzer Democrats 48, Republicans 41 Democrats +7 Direction of Country Politico/Morning Consult Right Direction 28, Wrong Track 72 Wrong Track +44 Tuesday, September 1 Race/Topic (Click to Sort) Poll Results Spread Georgia: Trump vs. Biden WSB-TV/Landmark* Trump 48, Biden 41 Trump +7 North Carolina: Trump vs. Biden East Carolina U. Trump 49, Biden 47 Trump +2 North Carolina Senate - Tillis vs. Cunningham East Carolina U. Cunningham 44, Tillis 44 Tie 9 hours ago, Brueggs said: Do you know of a better way to make people think a certain candidate is going to lose, so that people decide not to bother voting because it won't make a difference? Do you know a way to get all the pollsters to cheat like that?
Brueggs Posted September 2, 2020 Posted September 2, 2020 9 minutes ago, Tiberius said: General Election: Trump vs. Biden USA Today/Suffolk Biden 50, Trump 43 Biden +7 General Election: Trump vs. Biden Grinnell/Selzer Biden 49, Trump 41 Biden +8 President Trump Job Approval Grinnell/Selzer Approve 43, Disapprove 51 Disapprove +8 President Trump Job Approval Politico/Morning Consult Approve 42, Disapprove 57 Disapprove +15 2020 Generic Congressional Vote Grinnell/Selzer Democrats 48, Republicans 41 Democrats +7 Direction of Country Politico/Morning Consult Right Direction 28, Wrong Track 72 Wrong Track +44 Tuesday, September 1 Race/Topic (Click to Sort) Poll Results Spread Georgia: Trump vs. Biden WSB-TV/Landmark* Trump 48, Biden 41 Trump +7 North Carolina: Trump vs. Biden East Carolina U. Trump 49, Biden 47 Trump +2 North Carolina Senate - Tillis vs. Cunningham East Carolina U. Cunningham 44, Tillis 44 Tie Do you know a way to get all the pollsters to cheat like that? Answering questions with questions. Typical. Its not necessarily the pollsters cheating. I wonder what would happen if those conducting the polls chose a certain demographic to administer their questions to? I wonder if they could target specific groups that would be more likely to have similar opinions? Wow, I can't believe no one ever thought of that before.
Tiberius Posted September 2, 2020 Author Posted September 2, 2020 Political conventions and violent protests have not done much to change the race between President Donald Trump and Democratic nominee Joe Biden in 11 crucial battleground states, according to Morning Consult polls released Tuesday. The polls found the race little changed in Colorado, Minnesota, North Carolina, Ohio and Texas, while Biden gained ground in Arizona, Georgia, Michigan and Wisconsin. Trump gained 2 percentage points on Biden in Pennsylvania and 3 points in Florida, though the president still trailed in both of those states. Nationally,Biden was found to be leading Trump by eight points, 51%-43%, exactly what Morning Consult found earlier this month. Biden's 8-point put him in a better than Hillary Clinton, who had a 3-point advantage in the poll after the 2016 conventions. https://www.usatoday.com/story/news/politics/elections/2020/09/01/biden-gains-trump-wisconsin-leads-other-key-states-poll/3455027001/ 10 minutes ago, Brueggs said: Answering questions with questions. Typical. Its not necessarily the pollsters cheating. I wonder what would happen if those conducting the polls chose a certain demographic to administer their questions to? I wonder if they could target specific groups that would be more likely to have similar opinions? Wow, I can't believe no one ever thought of that before. Maybe they are master bating while conducting the poll. I can’t believe people are not thinking that.
Koko78 Posted September 2, 2020 Posted September 2, 2020 2 hours ago, Tiberius said: Do you know a way to get all the pollsters to cheat like that? Yes, hire Democrats.
The Frankish Reich Posted September 2, 2020 Posted September 2, 2020 (edited) 3 hours ago, Brueggs said: Answering questions with questions. Typical. Its not necessarily the pollsters cheating. I wonder what would happen if those conducting the polls chose a certain demographic to administer their questions to? I wonder if they could target specific groups that would be more likely to have similar opinions? Wow, I can't believe no one ever thought of that before. You need to read up on this. fivethirtyeight does a nice analysis of what they call the "house effect" of various pollsters. Some consistently trend in favor of Dems; some Repubs. Why don't they fix this since it is obviously a sampling issue? I don't know. I suppose there's an audience for their polls - call it confirmation bias. At any rate, these house effects can be controlled for, and that's what the good aggregators (like fivethirtyeight) do. General comment: I'm actually amazed at how good national polls (not state wide -- those have a lot more variability, and of course that matters given the electoral college!) are when you control for these effects. I steadfastly refuse to answer calls from unknown people, and most people I know do the same. So you've got to be getting a biased sample from any poll, right? You'd think so, but then there's no way the RCP and 538 projections could be this good. I have to just say I don't understand it, but the proof is in the results. Every 4 years the side that's losing in the polls claims there's something wrong or even a conspiracy to push out faked numbers. McCain's supporters (I believe it was then) actually started a site they called "unskewed polls." They added a few points to McCain based on perceived oversampling of Dems vs. Reps. It was trash, and turned out to be wrong, whereas the mainstream polls were right. EDIT: it was Romney 2012, not McCain. And the unskewed guy fessed up that he got it all wrong: https://www.businessinsider.com/unskewed-pollster-dean-chambers-nate-silver-election-dick-morris-michael-barone-2012-11 Edited September 2, 2020 by The Frankish Reich
The Frankish Reich Posted September 2, 2020 Posted September 2, 2020 Oops. Another conspiracy theory shot to ... smithereens. Added today DATES POLLSTER SAMPLE RESULT NET RESULT President: general election AUG 31-SEP 1, 2020 B- Ipsos 1,089 RV Biden 47% 40% Trump Biden +7 President: general election AUG 30-SEP 1, 2020 B YouGov 1,207 RV Biden 51% 40% Trump Biden +11 President: general election AUG 29-SEP 1, 2020 A/B IBD/TIPP 1,033 RV Biden 49% 41% Trump Biden +8 President: general election AUG 26-SEP 1, 2020 B/C USC Dornsife 2,650 LV Biden 51% 42% Trump Biden +9 President: general election AUG 26-SEP 1, 2020 B/C USC Dornsife 2,650 LV Biden 52% 42% Trump Biden +9 President: general election AUG 26-SEP 1, 2020 C+ Rasmussen Reports/Pulse Opinion Research 2,500 LV Biden 49% 45% Trump Biden +4 President: general election AUG 28-31, 2020 A Suffolk University 1,000 RV Biden 47% More Biden +5 President: general election AUG 28-31, 2020 A Suffolk University 1,000 RV Biden 50% 43% Trump Biden +7 President: general election AUG 26-30, 2020 A+ Selzer & Co. 827 LV Biden 49% 41% Trump Biden +8 President: general election AUG 21-25, 2020 Opinium Biden 56% 41% Trump Biden +15 KEY A = ADULTS RV = REGISTERED VOTERS V = VOTERS LV = LIKELY VOTERS Yesterday President: general election AUG 31-SEP 1, 2020 Redfield & Wilton Strategies 1,835 LV Biden 49% More Biden +8 President: general election AUG 25-31, 2020 B/C USC Dornsife 2,636 LV Biden 52% 42% Trump Biden +10 President: general election AUG 25-31, 2020 B/C USC Dornsife 2,636 LV Biden 51% 42% Trump Biden +10 President: general election AUG 30, 2020 B/C Morning Consult 4,000 LV Biden 52% 43% Trump Biden +9 President: general election AUG 28-30, 2020 B/C Morning Consult 12,966 LV Biden 51% 43% Trump Biden +8 President: general election AUG 28-30, 2020 Léger 861 LV Biden 49% More Biden +7 President: general election AUG 24-30, 2020 B/C AtlasIntel 4,210 LV Biden 49% 46% Trump Biden +3 President: general election AUG 29, 2020 B/C Morning Consult 4,000 LV Biden 51% 43% Trump Biden +8 Aug. 31, 2020 President: general election AUG 30-31, 2020 A- Emerson College 1,567 LV Biden 51% 49% Trump Biden +3 President: general election AUG 24-30, 2020 B/C USC Dornsife 2,574 LV Biden 54% 40% Trump Biden +13 President: general election AUG 24-30, 2020 B/C USC Dornsife 2,575 LV Biden 53% 41% Trump Biden +12 President: general election AUG 27-29, 2020 B/C RMG Research 915 LV Biden 48% More Biden +4 President: general election AUG 25-28, 2020 C HarrisX 2,862 RV Biden 47% 38% Trump Biden +9 Aug. 30, 2020 President: general election AUG 29, 2020 John Zogby Strategies/EMI Research Solutions 1,007 LV Biden 45% More Biden +3 President: general election AUG 29, 2020 John Zogby Strategies/EMI Research Solutions 1,007 LV Biden 48% 42% Trump Biden +6 President: general election AUG 23-29, 2020 B/C USC Dornsife 2,555 LV Biden 54% 40% Trump Biden +15 President: general election AUG 23-29, 2020 B/C USC Dornsife 2,556 LV Biden 54% 40% Trump Biden +14 Aug. 29, 2020 1
ziltoid Posted September 2, 2020 Posted September 2, 2020 Remember the Tulsa rally TicTok experiment? Remember 2016?
Tiberius Posted September 2, 2020 Author Posted September 2, 2020 Slew of new polls! Wednesday, September 2 Race/Topic (Click to Sort) Poll Results Spread General Election: Trump vs. Biden USA Today/Suffolk Biden 50, Trump 43 Biden +7 General Election: Trump vs. Biden Grinnell/Selzer Biden 49, Trump 41 Biden +8 General Election: Trump vs. Biden Rasmussen Reports Biden 49, Trump 45 Biden +4 General Election: Trump vs. Biden Economist/YouGov Biden 51, Trump 40 Biden +11 General Election: Trump vs. Biden Quinnipiac Biden 52, Trump 42 Biden +10 General Election: Trump vs. Biden IBD/TIPP Biden 49, Trump 41 Biden +8 Pennsylvania: Trump vs. Biden Monmouth* Biden 49, Trump 46 Biden +3 President Trump Job Approval Quinnipiac Approve 43, Disapprove 54 Disapprove +11 President Trump Job Approval IBD/TIPP Approve 39, Disapprove 55 Disapprove +16 President Trump Job Approval Grinnell/Selzer Approve 43, Disapprove 51 Disapprove +8 President Trump Job Approval USA Today/Suffolk Approve 45, Disapprove 52 Disapprove +7 President Trump Job Approval Politico/Morning Consult Approve 42, Disapprove 57 Disapprove +15 President Trump Job Approval Rasmussen Reports Approve 49, Disapprove 50 Disapprove +1 President Trump Job Approval Economist/YouGov Approve 43, Disapprove 55 Disapprove +12 2020 Generic Congressional Vote Economist/YouGov Democrats 51, Republicans 38 Democrats +13 2020 Generic Congressional Vote Grinnell/Selzer Democrats 48, Republicans 41 Democrats +7 2020 Generic Congressional Vote USA Today/Suffolk Democrats 48, Republicans 42 Democrats +6 Direction of Country USA Today/Suffolk Right Direction 30, Wrong Track 62 Wrong Track +32 Direction of Country Politico/Morning Consult Right Direction 28, Wrong Track 72 Wrong Track +44 Direction of Country Economist/YouGov Right Direction 25, Wrong Track 67 Wrong Track +42 Congressional Job Approval Economist/YouGov Approve 15, Disapprove 59 Disapprove +44
The Frankish Reich Posted September 2, 2020 Posted September 2, 2020 16 minutes ago, Deranged Rhino said: In what way is it harder to push out fake numbers than real numbers? I feel like I could manufacture a fake poll of, say, 1,000 individuals in literally one day. I'd just take the data from an old poll and tweak it a little bit to fit my "narrative" if I wanted to show, say, Biden leading by 11 points. Or 15. It's nothing. Doing real polling takes time. In other words, this is an utterly asinine comment by someone calling him/herself "The Last Refuge." 1
Recommended Posts