Today's Polls

I'm betting 2.9% of within the margin for error. Cripes I can't believe he may actually be catching her. Electing him just shows how frustrated people are with the gov't.

See my previous post about how relying upon polling averages can be statistically misleading. It appears that some posters (I'm not referring to you) think there is some value in repeating the same debunked nonsense, over and over. That's why they are best ignored.
 
See my previous post about how relying upon polling averages can be statistically misleading. It appears that some posters (I'm not referring to you) think there is some value in repeating the same debunked nonsense, over and over. That's why they are best ignored.

They are relying on it because they were pretty good predicters previously. I think 538 had a 90% accuracy rating one cycle, and 85 on another for all elections.

They actually take into acct (depending on the site) other factors as well, and have been pretty good.

What I'm wondering though is whether changes in technology have changed the accuracy of previously good methodology. I'm not a numbers geek this election is so far out of left field.

And btw the owner and chief geek at 538 admitted as much when he mea culpa'd how wrong they were during the primaries.

So I don't blame them for going with tried and true information. Whether it remains as such by reputation after this year's results are in is another matter.

I expect good results down ticket. And have absolutely no clue now as to the presidency.
 
They are relying on it because they were pretty good predicters previously. I think 538 had a 90% accuracy rating one cycle, and 85 on another for all elections.

They actually take into acct (depending on the site) other factors as well, and have been pretty good.

What I'm wondering though is whether changes in technology have changed the accuracy of previously good methodology. I'm not a numbers geek this election is so far out of left field.

And btw the owner and chief geek at 538 admitted as much when he mea culpa'd how wrong they were during the primaries.

So I don't blame them for going with tried and true information. Whether it remains as such by reputation after this year's results are in is another matter.

I expect good results down ticket. And have absolutely no clue now as to the presidency.

To be clear, in the quote you referenced, it was not 538 but the RCP average you were referred to. The RCP average immediately preceding prior elections does not have a remarkably accurate track record. How could it? It captures polls released on different dates and dealing with polling data sampled over different periods. Over the last 3 cycles, the most accurate poll was IBD. It wasn't "perfectly" accurate, and it might not be as accurate this time out, but those are the historical facts. Let's see which one is most accurate this time out.

There is some admiration for approach taken by 538, but as you point out, they were significantly wrong as recently as the GOP primaries.

Everyone is entitled to think certain polls are more reliable than others. That's why I post them ALL, as of the time of my posts.

But only the disingenuous will think that an average is necessarily more accurate than any specific poll. That's not even logical.
 
Everyone is entitled to think certain polls are more reliable than others. That's why I post them ALL, as of the time of my posts.

You post only the early morning results, never the full day's reports.
And that is usually just IBD, LA times and maybe one other, hardly a legit sampling.
 
Today's polls, thus far:

General Election: Trump vs. Clinton vs. Johnson vs. Stein ABC/Wash Post Tracking Clinton 46, Trump 45, Johnson 4, Stein 2 Clinton +1
General Election: Trump vs. Clinton ABC/Wash Post Tracking Clinton 49, Trump 47 Clinton +2
General Election: Trump vs. Clinton vs. Johnson vs. Stein IBD/TIPP Tracking Clinton 45, Trump 44, Johnson 4, Stein 2 Clinton +1
General Election: Trump vs. Clinton IBD/TIPP Tracking Clinton 45, Trump 43 Clinton +2

http://www.realclearpolitics.com/epolls/latest_polls/
How much did they oversample these polls??
 
How much did they oversample these polls??

Not sure, but this would be a good thread to post that sort of information.

For most voters (i.e. those who lack academic training as statisticians) what likely matters most are the track records of the various polls for accuracy in comparison to election results, as well as consideration of what factors may be historically unique in this election that might cause polls that were conducted with integrity to be, nevertheless, inaccurate.

Science becomes practical when it produces tools that can be used by ordinary people to predict results. Ordinary people tend to use those tools, without understanding how they were built or how they work, until such time as they stop correctly predicting results.
 
Updated list of polls released today, thus far:

General Election: Trump vs. Clinton LA Times/USC Tracking Clinton 43, Trump 47 Trump +4
General Election: Trump vs. Clinton vs. Johnson vs. Stein ABC/Wash Post Tracking Clinton 46, Trump 45, Johnson 4, Stein 2 Clinton +1
General Election: Trump vs. Clinton ABC/Wash Post Tracking Clinton 49, Trump 47 Clinton +2
General Election: Trump vs. Clinton vs. Johnson vs. Stein IBD/TIPP Tracking Clinton 45, Trump 44, Johnson 4, Stein 2 Clinton +1
General Election: Trump vs. Clinton IBD/TIPP Tracking Clinton 45, Trump 43 Clinton +2
General Election: Trump vs. Clinton vs. Johnson vs. Stein Rasmussen Reports Clinton 45, Trump 42, Johnson 5, Stein 2 Clinton +3

All are welcome to update as further polls are released.

http://www.realclearpolitics.com/epolls/latest_polls/
 
It takes the polling organizations a couple of days to collect data followed by a day to analyze it. It's going to be at least until Wednesday before the polls being published surveyed people after Comey's announcement, and say least Thursday before they reflect people having heard Clinton's response as well.

Polls lag significant events by 3 to 5 days.

So for the next few days I'd discount the polls.

Trump is up to 33% in prediction markets, which reflect news faster than polls. He was at 18% prior to Comey's interference in the election.
 
Today's polls, thus far:

General Election: Trump vs. Clinton vs. Johnson vs. Stein ABC/Wash Post Tracking Clinton 45, Trump 46, Johnson 3, Stein 2 Trump +1
General Election: Trump vs. Clinton LA Times/USC Tracking Clinton 43, Trump 47 Trump +4
General Election: Trump vs. Clinton ABC/Wash Post Tracking Clinton 48, Trump 47 Clinton +1
General Election: Trump vs. Clinton vs. Johnson vs. Stein IBD/TIPP Tracking Clinton 45, Trump 44, Johnson 4, Stein 2 Clinton +1
General Election: Trump vs. Clinton IBD/TIPP Tracking Clinton 45, Trump 44 Clinton +1

http://www.realclearpolitics.com/epolls/latest_polls/
 
Today's polls, thus far:

General Election: Trump vs. Clinton vs. Johnson vs. Stein IBD/TIPP Tracking Clinton 44, Trump 44, Johnson 4, Stein 2 Tie
General Election: Trump vs. Clinton IBD/TIPP Tracking Clinton 44, Trump 44 Tie
General Election: Trump vs. Clinton LA Times/USC Tracking Clinton 42, Trump 48 Trump +6
General Election: Trump vs. Clinton vs. Johnson vs. Stein ABC/Wash Post Tracking Clinton 46, Trump 46, Johnson 3, Stein 2 Tie
General Election: Trump vs. Clinton ABC/Wash Post Tracking Clinton 48, Trump 47 Clinton +1

http://www.realclearpolitics.com/epolls/latest_polls/
 
All tied up down the home stretch.
 
Close, but not tied.
Clinton is still up according to the recent polls from bud's link as well as their averages.

It's tied. Those are margin of error differences and residual from over sampling.
 
It's tied. Those are margin of error differences and residual from over sampling.

There are some other posters who appear to have a disdain for uncontroversial mathematical principles. I posted 5 polls. The net difference between the candidates over those 5 polls is +5 for Trump (3x ties, 1x +6 for Trump and 1x +1 for Clinton). That averages to a +1 Trump lead over 5 polls. However, that number is well within the margin of error of any of these polls. Therefore, statistically, it is a tie.

To come up with an average of polls that gives Clinton a lead, you would have to include other polls (either older or newer, which would necessarily be predicated on a different sampling period and a different set of information available to those being polled). However, if that average produced a number lower than the margin of error (which it is likely to), you'd still have a statistical tie.

That part is just math. The rest is statistical manipulation.

Of course, this just looks at the national polls, and it is the state votes that will determine the election.
 
That's closer than I thought it would be frankly. It will depend on turnout, are Trump supporters fired up and Clinton's dejected by the trifecta (Wikileaks, Obamacare rates, FBI Wiener-mail) or not
 
That's closer than I thought it would be frankly. It will depend on turnout, are Trump supporters fired up and Clinton's dejected by the trifecta (Wikileaks, Obamacare rates, FBI Wiener-mail) or not

Cnn has been covering the advanced vote and it seems Clinton is not doing as well with the Obama coalition as expected. And the supposed states in play like Texas just aren't.

Trump's path is still a bit narrow. But so now are the polls.
 
Cnn has been covering the advanced vote and it seems Clinton is not doing as well with the Obama coalition as expected. And the supposed states in play like Texas just aren't.

Trump's path is still a bit narrow. But so now are the polls.

That's what I've been reading as well.
Chances are Clinton still wins at this point.
 
Today's polls, thus far:

General Election: Trump vs. Clinton vs. Johnson vs. Stein CBS News/NY Times Clinton 45, Trump 42, Johnson 5, Stein 4 Clinton +3
General Election: Trump vs. Clinton CBS News/NY Times Clinton 47, Trump 44 Clinton +3
General Election: Trump vs. Clinton LA Times/USC Tracking Clinton 43, Trump 48 Trump +5
General Election: Trump vs. Clinton vs. Johnson vs. Stein ABC/Wash Post Tracking Clinton 47, Trump 45, Johnson 3, Stein 2 Clinton +2
General Election: Trump vs. Clinton ABC/Wash Post Tracking Clinton 49, Trump 47 Clinton +2
General Election: Trump vs. Clinton vs. Johnson vs. Stein Rasmussen Reports Clinton 42, Trump 45, Johnson 4, Stein 1 Trump +3
General Election: Trump vs. Clinton vs. Johnson vs. Stein IBD/TIPP Tracking Clinton 44, Trump 44, Johnson 4, Stein 2 Tie
General Election: Trump vs. Clinton IBD/TIPP Tracking Clinton 44, Trump 44 Tie

http://www.realclearpolitics.com/epolls/latest_polls/