of course the final numbers tell the real story.
trump smoked biden in Iowa.
every other pollster missed that by a LOT.
she was the outlier and ultimately proven correct.
Allan
her poll was all over the place. tied to 10 points to 1 pt to 7 points. If you graphed it it would be an N for the Nut the blind squirell found.
While other pollsters were not as close on the spread, they were closer on the final numbers for each candidate.
Emerson and Insider advantage were 1 pt of on Bidenās final number, Selzer was 4 pts off. PPP and insider advantage were the exact same amount off on Trump as she was 5 points.
not close?. they swung and the missed.
selzer hit a home run.
Allan
and if I said Trump would win 8% to 1% I would have been closer on the spread.
Its not the spread that measures accuracy, its how close you get to each candidates actual performance.
i disagree. the spread is the key to me. when i post my polls i never use the percentage number, only the spread. even you use the spread on your posts. how close is the race? percentages are meaningless.
Allan
The spread is great for tracking, but its not a measure of accuracy.
Look
Trump will beat Biden in IA in 2020 8% Trump to 1% Biden
I just outperformed Selzer on the spread. Accuracy is a measure of how close you get to each candidateās actual performance.
trump won in 2020 by 8. unsure where you are getting biden +1
actual results determine accuracy.
Allan
not biden plus 1. Trump will get 8% and Biden will get 1%. The spread is 7% Is it an accurate spread? Yep, sure is.
those are ridiculous numbers.
Allan
its an example of how the spread is useful for tracking but is NOT a measure of accuracy. Accuracy is measured by how close the prediction is to each candidatesā final numbers. MoE guarantees the spread will likely be off. The most accurate pollsters are those who get the closest to each candidateās actual tally within MoE. In 2020 IA, no pollsters measurements were within MoE. The closest 2 were Emerson and Insider Advantage.
Samm
1319
Yes, much depends on who is paying for the poll. My question is, who pays for neutral polls? Does anybody? And to take that one step farther, if not, are there in fact any honest neutral polls?
The ones that colleges and universities do on their own. Some also just conduct them because they know the media will buy the stories. Of course they may be aiming for specific media, and that can skew things
Samm
1321
Colleges and Universities are not unbiased.
not completely, but the most accurate pollsters over the past 10 years are mostly colleges and universities.
Emerson
Suffolk
Marquette
are all among the most accurate. Especially when they do the poll on their own. When they do it for someone, they are not as accurate.
The most accurate polling organization over that time seems to be insider advantage (this surprised me) but each of them is ranked toward the top by both RCP and 538.
1 Like
Electoral Template. The race has tightened post DNC and has mostly frozen where itās at. Basically, a tossup. The debate does not seem to have changed anything. The lead changes from day to day and poll to poll.
Contested states (lead is >0 but <4 and leader (over 3 pts) has broken 50% in fewer than 2 polls)
WI: Harris up 1+ has broken 50% in two polls
MI: Harris up 1 has broken 50% in three polls
PA: Harris up <1 both have broken 50% in two polls
GA: Trump up <1 has not broken 50% in any poll
AZ: Trump up 1+ both have broken 50% in one poll
NC: Trump up <1 both have broken 50% in one poll
NV: Harris up <1 has not broken 50% in any poll
Competitive (lead is >3 but <7 and leader (over 6) has broken 50% in fewer than 2 polls)
MN: Harris up 5+ has broken 50% in one poll
VA: Harris up 5+ has broken 50% in one poll
FL: Trump up 6 has broken 50% in two polls.
OR: Harris up 5 has not broken 50% in any poll
ME2: Harris up 5 has not broken 50% in any poll
Marginal (lead is >6 but <9 and leader (over 8) has broken 50% in fewer than 3 polls)
TX: Trump up 7+ has broken 50% in two polls
NE2: Harris up 8 has broken 50% in one poll
NH: Harris up 7+ has broken 50% in three polls.
Note: TX lead includes older outlier, recent polling lead is 4+ pts.
Note: If Harris wins the states where she has a lead she wins with 277 EVās
1 Like
Nemesis
1324
You have now got me interested in polls and polling so I looked up those universities and interestingly enough they all say Harris is leading Trump.
Now I have not dissected and compared the demographics of the voters they approached or the questions but should Trump be concerned that Harris is obviously as far as the electorate is concerned a much more viable candidate than Biden ever was.
I have to wonder if Trumps campaign now regrets making such an issue of Bidens cognitive decline so early because if he was still the candidate it would have been an easy victory for Trump.
- Emerson: Harris 49% Trump 47%
- Suffolk: Harris 47.6% Trump 43.3% though this was at end of August.
- Marquette: Harris 52% Trump 48%
https://law.marquette.edu/poll/
national polls are meaningless. its the state polls that matter. I donāt even look at national polls
Nemesis
1326
Genuine question - why in your opinion are national polls meaningless?
because thats not how our elections work. just like the popular vote in the election, harris can run up big numbers in CA and NY, but thats not going to win her PA. right now there are some states that āmatterā
PA, MI, WI, GA, NC, NV and AZ
and to a much lesser extent
VA, MN and the 2nd district of NE.
Both candidates have a chance to win any of those states. Though the chance in the final 3 is thin. The others are locked in and they wonāt change.
1 Like
National Polls are definitely not meaningless, but they have a slightly different value.
Since National Polls are released very regularly and State Polls are more released more sparsely, National Polls provide a reasonably accurate measure of where vote is moving.
For example, Harris was +1.5 to +2.5 in National Polls right before the debate. Now she is +3 to +5 in National Polls.
Clearly Harris won the debate and it is showing up in a meaningful move over the course of only a week.
Presumably people in the swing states also saw the debate and you can expect some movement in those polls towards Harris when they start publishing.
