It looks like you're using an Ad Blocker.

Please white-list or disable AboveTopSecret.com in your ad-blocking tool.

Thank you.

 

Some features of ATS will be disabled while you continue to use an ad-blocker.

 

Help ATS via PayPal:
learn more

Trump Suppoerters, Can You Just Let Us Know Which Polls Are Valid?

page: 7
90
<< 4  5  6    8 >>

log in

join
share:

posted on Nov, 9 2016 @ 02:07 PM
link   

originally posted by: Myollinir
a reply to: kruphix

the sample sizes are always ~1,000 people and you can tell why the poll is swinging one way or the other by looking at the demographics.

they're all pretty much "duh" indicators. the real poll is the final general election

ALTHOUGH you can see sways - such as if Trump was +7 but they oversampled Republicans above Democrats by 10% or something, then you could see that Trump would actually be at a deficit. This goes vice versa for those polls with Clinton up above Trump and the Democrats are polled at a higher percentage than Republicans. I would pay attention to this data much more than the (CLINTON IS UP +5%)..


Just for the record, election polls do not attempt to get a 50/50 split (or any kind of split) among party affiliation because it's seen as an attitude rather than an immutable characteristic of the voter. What they do try to do is measure LIKELY VOTERS, based upon those immutable characteristics such as gender, race, geographic location, amount of education, etc. etc.)

This is where the polls failed.

There was a huge outpouring of working class white voters (Democratic, Republican, and Independent) who all turned out in droves to vote for Trump. The polls did not predict this.
edit on 9-11-2016 by Greggers because: (no reason given)




posted on Nov, 9 2016 @ 02:08 PM
link   

originally posted by: MotherMayEye
a reply to: Greggers

No reason to trust any of the polls. They're propaganda.


No they're not.



posted on Nov, 9 2016 @ 02:12 PM
link   
a reply to: Greggers

Agree to disagree.



posted on Nov, 9 2016 @ 02:14 PM
link   
I guess we can tell now which polls were valid...the ones that had Trump winning.



posted on Nov, 9 2016 @ 02:15 PM
link   

originally posted by: MotherMayEye
a reply to: Greggers

Agree to disagree.


That's cool. The point I like to stress is that all the reputable polls follow proven statistical techniques, which can be vetted by anyone using various online calculators to determine sample size, margin of error, and confidence level. Then, those pollsters use historical data to match various demographics to predict which demographics will turn out.

As long as voter turnout follows historical trends, the polls are pretty accurate. Looking at these election results, the error can be traced back to a very specific outcome that was not predicted by historical trends.



posted on Nov, 9 2016 @ 02:19 PM
link   

originally posted by: Blueracer
I guess we can tell now which polls were valid...the ones that had Trump winning.


Not necessarily. The nationwide polls that showed Hillary ahead within the margin of error of the actual result will end up being technically correct.

Also, I'm sure we could dig up state polls in various states that nailed it, although I'd have to dig around for that.

I think what it's safe to say, however, is that a massive new voter trend was not identified by historical data, so the voter turnout was estimated erroneously in a huge number of polls, both nationwide and per state.

By the way, just because a poll got the right result doesn't mean it was a valid poll, if they got the right result for the wrong reason.

If you'll excuse my use of a particularly ridiculous example, let's say a poll predicted that only 1% of registered black voters would turn out, and as a result Trump would win big. In that case, the poll would produce the correct result, but it would still be an invalid methodology.
edit on 9-11-2016 by Greggers because: (no reason given)



posted on Nov, 9 2016 @ 02:20 PM
link   
a reply to: Greggers

You are wrong. IDP/TIPP and USC/LATimes both accounted for the correct demographics and both consistently predicted the winner. Rasmussen used a better sampling than ABC/WaPost and also more closely called the election.

It is more than obvious that the media were using polls as a weapon rather than a tool.



posted on Nov, 9 2016 @ 02:20 PM
link   
a reply to: Greggers

I answered the thread title. I stand by my statement.



posted on Nov, 9 2016 @ 02:22 PM
link   

originally posted by: raymundoko
a reply to: Greggers

You are wrong.


About what?



posted on Nov, 9 2016 @ 02:23 PM
link   

originally posted by: Blueracer
a reply to: Greggers

I answered the thread title. I stand by my statement.



Your statement that any poll that predicted Trump as the winner was valid?

That's what you're standing by?



posted on Nov, 9 2016 @ 02:47 PM
link   
a reply to: Greggers

You think the majority of polls weren't propaganda. Yes, they used scientific method to get to where they are, but they used incorrect samples and that was pointed out here on ATS regularly for a year.

We see from the Wikileaks that the Clinton campaign wanted to get their "sample demographics" to "their friends in the media" to "drive perception". The last few WikiLeaks made it all to clear there was collusion between the Clinton Campaign and Media Elites.

Did you forget about the emails?



posted on Nov, 9 2016 @ 03:13 PM
link   

originally posted by: raymundoko
a reply to: Greggers

You think the majority of polls weren't propaganda. Yes, they used scientific method to get to where they are, but they used incorrect samples and that was pointed out here on ATS regularly for a year.

We see from the Wikileaks that the Clinton campaign wanted to get their "sample demographics" to "their friends in the media" to "drive perception". The last few WikiLeaks made it all to clear there was collusion between the Clinton Campaign and Media Elites.

Did you forget about the emails?


If you're talking about the Podesta Atlas report email discussing oversamples, that is not what it referred to at all.

A quick review of the actual Atlas document makes it extremely clear that the report was intended only for internal DNC consumption and that it pertained to the oversamples used by the DNC to better study how to get their message out to key demographics.

Each example of oversampling was clearly illustrated in the Atlas report, along with the reasons why.

It had no relationship whatsoever to public MSM polls. In fact, these internal polls were the DNC's secret sauce and they wouldn't have wanted to share them with anyone.

Again, this was a case of the alt-right media making stuff up. I am happy to walk through that document with you, if you doubt me. Otherwise, I encourage you to locate it and review it for yourself.

As far as people pointing out that the polls had the demographics wrong, mostly this was from people who thought the polls should be split more agreeably between registered republicans and democrats, which, as I've explained, is not something that most polls even TRY to do, for the reasons I gave earlier. Nor would it have been correct to do it in THIS case.

Instead, what would have been correct was to predict record turnout among blue color white voters. Which poll did that? Whichever one did (if any) got it right.

Did the two polls you mentioned earlier predict record turnout amount people with:
1) No college education
2) Blue collar occupation
3) White
4) Male

If so, Kudos to them.

By the way, most of the polls correctly predicted that this group would vote overwhelmingly for Trump. They, however, failed to predict the turnout.
edit on 9-11-2016 by Greggers because: (no reason given)



posted on Nov, 9 2016 @ 03:22 PM
link   
a reply to: Greggers

I did review it, it said our friends in media. Sorry your team got exposed.

I never said polls should be agreeably split. I said several times they were over sampling minorities based on Obama's turnout. Rasmussen said the same thing about other polling firms.
edit on 9-11-2016 by raymundoko because: (no reason given)



posted on Nov, 9 2016 @ 03:23 PM
link   

originally posted by: raymundoko
a reply to: Greggers

I did review it, it said our friends in media. Sorry your team got exposed.


No, it didn't.

I think you're referring to Atlas media, the polling organization they used.

Did you read the Atlas report itself? It clearly has nothing whatsoever to do with any mainstream poll. Seriously go read it.



posted on Nov, 9 2016 @ 03:30 PM
link   
For the record, here are some of the OVERSAMPLES recommended by Atlas Media. Notice how they are all geared toward helping the DNC get the most out of their polling, so they can accurately and effectively message to key demographics. Again, this has nothing to do with MSM polls.

Also, it's called a "media poll" specifically because it provides information on how the DNC can best utilize MEDIA (ie ads and so forth) to reach their target market.

Excerpts:


Focus groups are important in NM, given the importance of authenticity to NM voters. Focus groups should be done early in the campaign and again later to test ads. Given the large percentage of the population living in the Albuquerque metro area, most of the focus groups conducted in New Mexico occur in Albuquerque. The facility there can also attract people from the surrounding areas. Las Cruces and Santa Fe are the other two areas where pollsters often do focus groups. Outside of those three areas, the towns are small and spread out, making it harder to conduct focus groups. Of course, Anglo and Hispanic focus groups should be done separately with appropriate moderators.


Democratic candidates need to persuade conservative Democrats to remain loyal to the party. This might require a persuasion program targeting these Democrats with a specific regional message. Albuquerque Democrats, Santa Fe Democrats, and rural Democrats all respond to unique messages.


Democratic candidates in 2008 should look to increase the Democratic vote from women. Consider doing a research project to determine how best to do that.


Consider doing a Hispanic only baseline survey in early 2008 to accurately capture the mood of the community and draw out the differences between Northern Hispanics in CD3 and Southern Hispanics in CD2. Bush greatly improved in his share of the Hispanic vote, winning 44% as opposed to 32% in 2000. A Democratic priority should again be to win close to 70% of the Hispanic vote in 2008. After the baseline survey, re-assess after the June primary. Poll every two to three weeks. Then, poll weekly for the last five weeks.


Bush scored strong gains in voters under the age of 60, while older voters remained strongly Democratic in 2004. This is an important trend to watch in the future as it runs somewhat counter to national trends. Consider a research project to determine how Democrats can improve their support from young voters in NM, as it may be more difficult to maintain an 11% lead among older voters.


A micro-targeting survey should be considered given the difference between Democratic registration and Democratic voting. The survey should probably take place early in the summer after the primaries, as we would want to generate candidate scores for both presidential and Senate candidates.

21
The survey should take place relatively soon after the primary, as field campaigns will want to use the data in their summer base vote and persuasion programs.

edit on 9-11-2016 by Greggers because: (no reason given)

edit on 9-11-2016 by Greggers because: (no reason given)



posted on Nov, 9 2016 @ 03:36 PM
link   

originally posted by: kruphix
I was about to make a couple threads, one about the current view of the national polls and why they are showing that Clinton is still well ahead in the polls. ... Snip

Or, do you think all polls are invalid and that crowd size/yard signs/Facebook likes are more accurate? If so, please try to give a convincing argument as to why you think this.


TPP Poll is one AZ and TX done. Trump

www.investors.com...

NYT

www.nytimes.com...

RCP

www.realclearpolitics.com...


Enough?



posted on Nov, 9 2016 @ 03:39 PM
link   
a reply to: Greggers

I did read the email. They wanted Atlas to recommend over samples so they can maximize their MEDIA polling, not internal. Sorry your side got exposed.

The polls off by the most were left leaning media organizations. We already know they were colluding with the Clinton campaign.



posted on Nov, 9 2016 @ 03:39 PM
link   

originally posted by: raymundoko
I never said polls should be agreeably split. I said several times they were over sampling minorities based on Obama's turnout. Rasmussen said the same thing about other polling firms.


That's a fair criticism. There were also fair criticisms leveled against both of the polls you mentioned.

And the methodology for that LA Times poll was particularly unusual.

Also, over-predicting black turnout did not, in and of itself, account for what we saw last night.



posted on Nov, 9 2016 @ 03:42 PM
link   

originally posted by: raymundoko
a reply to: Greggers

I did read the email. They wanted Atlas to recommend over samples so they can maximize their MEDIA polling, not internal. Sorry your side got exposed.

.


I asked you to review the Atlas Report. I quoted it above, so you can clearly see this is internal polling.

And as I said, it's called MEDIA polling because it is designed to help the DNC understand how to best utilize various forms of media (ads, campaign letters, etc.) to reach key demographics.



posted on Nov, 9 2016 @ 03:59 PM
link   
a reply to: Greggers

Media polling is polling used/completed by the media. You've never taken marketing classes in college? You are confusing Media Marketing for Media Polling. Often it's just shortened to plain old Media. To throw "polling" after it change the definition.

They are also asking for MORE info from Atlas, NOT what is in the document.

Are you suggesting the email is referring to a marketing campaign based on over sampling?

The document you reference specifically says they need to over sample hispanics and native Americans. It also recommends to make sure they don't get the "old people" in Florida and to make sure they are getting mostly younger voters.

It's a great 37 pages. You should read all of it.

edit on 9-11-2016 by raymundoko because: (no reason given)



new topics

top topics



 
90
<< 4  5  6    8 >>

log in

join