Quadrennial Note on polls and polling

[FillJackson]

Polling averages are more accurate than single polls, because averages use many more data points and because averages minimize the effect of outliers - polls which are far away from what other polls are showing.

Using polling averages keep you intellectually honest, it prevents you from cherry picking favorable polls which is very tempting to do. In the past two weeks, one poll said this was a one point race while a different poll said this was a 13 point race. If you bought into either of those polls you mostly likely fooled yourself

If there is big movement in the race, that will show up in more than one poll and will soon be reflected in the polling average.

For the general election, the Pollster average is here
http://elections.huffingtonpost.com/pollster/2016-general-election-trump-vs-clinton

The RCP average is here
http://www.realclearpolitics.com/epolls/2016/president/us/general_election_trump_vs_clinton-5491.html

I like the way Pollster presents the data. You can see individual polls represented as dots. The average of the polls presented as a line. The margin of error of that line is represented by shading. So you can see if a candidate's lead is within the margin of error or outside it. Sliding your cursor over the chart lets you see it play out in time.

The individual dots are the "noise" and the trend lines are the "signal" in the data.

[fiddy]

intellectually honest

:roll:

[9erempiree]

Surprisingly good graph from the Huffington Post considering how bias they are.

Anyways, if you look at the graph....Clinton down while Trump is trending up.

He's been up since June where Clinton has been down.

[9erempiree]

I expect Bernie supporters to vote for Trump the more this drags out.

As a Trump supporter, I think Bernie is unfairly treated and its a disgrace to his supporters.

Together we can bring down the establishments.

Plus, all Trump has to do is tone it down and show his leadership. He has room to improve. Clinton on the other hand, can't do anything to get the voters' trust.

If I was undecided, there is nothing Clinton can do to win me over where there is still time for Trump to win me over.

[FillJackson]

Surprisingly good graph from the Huffington Post considering how bias they are.

Good pollsters look for accurate date. There is some modelling going on here, but it's mainly an average of other polls. Pollster was its own thing before HuffPost bought them.

I don't think 538 has it's model up, but once of the reasons they have been so accurate is their model tries to account for good vs bad polls. They rank pollsters based on how accurate their track has been and other things like margin or error and number of respondents.

A 1500 person sample from PPP might be weighted more than a 500 person sample from Rasmussen or Gravis.


Anyways, if you look at the graph....Clinton down while Trump is trending up.

He's been up since June where Clinton has been down. Yes, Trump is trending up since started campaigning for President.

[FillJackson]

Looking good. The American voters are taking notice.
https://i.sli.mg/zPdYks.png

I tried.

[Nanners]

Polling averages are more accurate than single polls, because averages use many more data points and because averages minimize the effect of outliers - polls which are far away from what other polls are showing.

Using polling averages keep you intellectually honest, it prevents you from cherry picking favorable polls which is very tempting to do. In the past two weeks, one poll said this was a one point race while a different poll said this was a 13 point race. If you bought into either of those polls you mostly likely fooled yourself

If there is big movement in the race, that will show up in more than one poll and will soon be reflected in the polling average.

For the general election, the Pollster average is here
http://elections.huffingtonpost.com/pollster/2016-general-election-trump-vs-clinton

The RCP average is here
http://www.realclearpolitics.com/epolls/2016/president/us/general_election_trump_vs_clinton-5491.html

I like the way Pollster presents the data. You can see individual polls represented as dots. The average of the polls presented as a line. The margin of error of that line is represented by shading. So you can see if a candidate's lead is within the margin of error or outside it. Sliding your cursor over the chart lets you see it play out in time.

The individual dots are the "noise" and the trend lines are the "signal" in the data.

how do you feel about the fact that trump has officially overtaken clinton in the RCP average and is now only 2 points behind in the huffington post average?

[9erempiree]

how do you feel about the fact that trump has officially overtaken clinton in the RCP average and is now only 2 points behind in the huffington post average?

It's going to get worst for them.

[FillJackson]

how do you feel about the fact that trump has officially overtaken clinton in the RCP average and is now only 2 points behind in the huffington post average?

In terms of the polling, it shows the race is tightening at this moment.

In terms of the whole shebang, I'm not worried. I am more worried that it just set off a round of usual Democratic panic.

Clinton right now is taking hits from her left and her right.

Trump is no longer taking Republican hits, so his support is solidifying. It's not unusual at all for Trump to get a bump because he became the nominee. Republicans are basically coming home to home base.

However, Trump is still below where Romney was at this point. And Republicans coming to home base is still 206 electoral votes. They are going to have to do better than that.

Clinton is two weeks from becoming the nominee on June 7th. I expect to her to get a bump as well after that. There's been polling showing that the Never Obama crowd of Democratic voters in 2008 is bigger than the Never Clinton crowd of 2016.

Sanders has said that he is going to convention with the intention of changing some of the rules of the Democratic primaries

If I were the Clinton campaign manager and I saw these polls, I would basically sit tight and start building the get-out-the-vote structure. And you know what, they started doing that a while ago.

[Nanners]

There's been polling showing that the Never Obama crowd of Democratic voters in 2008 is bigger than the Never Clinton crowd of 2016.


source?

anyway there are some key differences between the never obama crowd and the never clinton crowd.

the never obama crowd was mainly old people and driven by racism/xenophobia, the never clinton crowd is mainly young people and driven by dislike of clintons pro-corporate policies and corporate ties. policy wise there is a MUCH larger gulf between sanders and clinton in 2016 than there was between obama and clinton in 2008. also, in 2008 the DNC did not rig the entire process for obama they way they have for clinton in 2016. obama earned his nomination while clinton is having the nomination gift wrapped to her. if you think the anti clinton crowd is going to forget all of this and turn into clinton supporters after the convention, you are in for a rude awakening.