Free Republic
Browse · Search
News/Activism
Topics · Post Article

Skip to comments.

Pollster Oversample Chart
BigEdLB ^ | 8/13/16 | BigEdLB

Posted on 08/14/2016 2:07:27 PM PDT by BigEdLB

Here is my analysis of oversampling during polling vs Gallups analysis of that actual party ID in a state is.



TOPICS: Politics/Elections
KEYWORDS: 2016polls; deceit; fraud; oversample; polling; polls
Navigation: use the links below to view more comments.
first previous 1-2021-4041-6061-76 next last
To: BigEdLB

Thank you very much


21 posted on 08/14/2016 2:26:20 PM PDT by Steve Van Doorn (*in my best Eric Cartman voice* 'I love you, guys')
[ Post Reply | Private Reply | To 11 | View Replies]

To: Verbosus; Steve Van Doorn

Close. What it shows is the percent of oversample of Dems and Republicans VS what Gallup says are actual persons registered as D or R in each state......


22 posted on 08/14/2016 2:26:20 PM PDT by Red Badger (Make America AMERICA again!.........................)
[ Post Reply | Private Reply | To 10 | View Replies]

To: BigEdLB

Thanks.


23 posted on 08/14/2016 2:27:38 PM PDT by mrsmith (Dumb sluts: Lifeblood of the Media, Backbone of the Democrat/RINO Party!)
[ Post Reply | Private Reply | To 17 | View Replies]

To: BigEdLB

Here is a link to the most recent Gallup Poll data on party affiliation:
http://www.gallup.com/poll/15370/party-affiliation.aspx


24 posted on 08/14/2016 2:27:46 PM PDT by Nero Germanicus
[ Post Reply | Private Reply | To 17 | View Replies]

To: Nero Germanicus

He has one that is state by state. that is the national one over time


25 posted on 08/14/2016 2:30:34 PM PDT by BigEdLB (Take it Easy, Chuck. I'm Not Taking it Back -- Donald Trump)
[ Post Reply | Private Reply | To 24 | View Replies]

To: BigEdLB

bookmark


26 posted on 08/14/2016 2:30:50 PM PDT by dadfly
[ Post Reply | Private Reply | To 1 | View Replies]

To: BigEdLB

Thanks. Based on your calculations where you show 79 data points, 15 points with a Republican oversample, the likelihood of that kind of skewed result (or more skewed) based on approximation via normal distribution (Vassarstats.net) is:

< 0.000001 ie 1 in a million.

So either the pollsters are crooked or GOPers are more shy than Dems.


27 posted on 08/14/2016 2:37:09 PM PDT by ScaniaBoy (Part of the Right Wing Research & Attack Machine)
[ Post Reply | Private Reply | To 18 | View Replies]

To: ScaniaBoy
So either the pollsters are crooked or GOPers are more shy than Dems
28 posted on 08/14/2016 2:42:18 PM PDT by BigEdLB (Take it Easy, Chuck. I'm Not Taking it Back -- Donald Trump)
[ Post Reply | Private Reply | To 27 | View Replies]

To: Nero Germanicus; jamese777; BigEdLB

LoL People’s Press?

They should call themselves the Communist’s Press. That’s Pew Research once run by Madeleine Halfwit. You’re still plying your trade Zot Jamese777/Nero for Obama and Hillary.


29 posted on 08/14/2016 2:43:04 PM PDT by Red Steel
[ Post Reply | Private Reply | To 19 | View Replies]

To: BigEdLB

Why, in 29 states where oversampling is recorded, 26 are oversampled to favor the ‘rats in the polls - only 3 are oversampled in favor of the ‘pubs - assuming a chance oversampling based on the random selection of participants, we would expect by chance that oversampling would favor the ‘pubs 14 times out of the 29 - instead we get only 3/29 in their favor - it’s been too long since I had statistics to remember how to figure the probability of this outcome, but it is quite small, probably well below the 5% level - we may thus reject as being too improbable the nulhypothesis that this sampling occurred randomly, and conclude that the pollsters are deliberately oversampling their polls to favor ‘rats - who said conservatives don’t like science......


30 posted on 08/14/2016 2:43:39 PM PDT by Intolerant in NJ
[ Post Reply | Private Reply | To 1 | View Replies]

To: Intolerant in NJ

This is my point. You explained it. Corrupt pollsters


31 posted on 08/14/2016 2:45:38 PM PDT by BigEdLB (Take it Easy, Chuck. I'm Not Taking it Back -- Donald Trump)
[ Post Reply | Private Reply | To 30 | View Replies]

To: BigEdLB

First, this is a lot of work, so thanks for your efforts.

Second, in your opinion as someone who understands these things - do you think that most polling companies oversample Dems for the reason that people on FR would say - i.e. they want to skew the outcome to favor the Democrat as a way of influencing public opinion?

Or is there a statistically-valid reason for doing so?

In other words, does your chart help educate us on the inner workings of fair and honest polls, or does it expose manipulation?


32 posted on 08/14/2016 2:49:48 PM PDT by bigbob (The Hillary indictment will have to come from us.)
[ Post Reply | Private Reply | To 1 | View Replies]

Bottom line, any poll winds up with an assumption...either created or by happenstance of response....on the weight of an election. IF Dem turnout is TEN percentage points more than GOP, THEN perhaps Hillary is way ahead. But that has no basis in reality. As the OP stated, Gallup has party Id EVEN. Now pollsters could use other methods, but it is reasonable to state they Party ID question is even.

In 2012, If I recall, The Dems only had a 3 point advantage in Party ID polling at this time of the year. So from that score, the GOP is already polling 3 points better than 2012.

It is true this election is going to be bit odder than any since the Perot elections, because some of the particulars are all over the map...like Trump getting big Latino and Black support in some polls, and not in others. Like Dems voting Trump and GOP voting Hillary.

Then you have the Sanders voters, and Johnson polling above what I personally think his name recognition is.

So this election could be a real watershed true shift election...meaning....none of the usual models are going to work, and if they do, it would probably be dumb luck.

I see this as globalism vs nationalism, and Hillary’s corruption vs Trump’s perceived personality issues. That has the potential to provide a very unusual set of voting patterns in Nov.

And we haven’t even made it to the debates yet...and Trump has yet to even start a traditional ad/local office campaign.

Throw in the leaks, and any pollster who is telling you they have this nailed is lying.


33 posted on 08/14/2016 2:49:52 PM PDT by Crimson Elephant
[ Post Reply | Private Reply | To 27 | View Replies]

To: Intolerant in NJ

The two-tailed probability of 3 or fewer out of 29 is: 0.000043 ie 43 in one million (Vassarstats.net).

Not very likely!


34 posted on 08/14/2016 2:50:28 PM PDT by ScaniaBoy (Part of the Right Wing Research & Attack Machine)
[ Post Reply | Private Reply | To 30 | View Replies]

To: All

This is, I think what longroom was trying to say before he was mysteriously silenced. I use Gallup as my backup and compare their numbers to these other pollsters. No secret forula sauce. At some point, if it hasn’t already started, I expect some of these folks to start to make up stuff as they go along, if the surveys do match the narrative. Being a Democrat means never having to tell the truth.


35 posted on 08/14/2016 2:52:28 PM PDT by BigEdLB (Take it Easy, Chuck. I'm Not Taking it Back -- Donald Trump)
[ Post Reply | Private Reply | To 31 | View Replies]

To: bigbob

I am in line with those here at FR who believe the Dem oversample is a way to create public opinion. I also think it is starting to backlash. Marist is th worst of the bunch, but PPP isn’t far behind.


36 posted on 08/14/2016 2:55:45 PM PDT by BigEdLB (Take it Easy, Chuck. I'm Not Taking it Back -- Donald Trump)
[ Post Reply | Private Reply | To 32 | View Replies]

To: BigEdLB

“what longroom was trying to say before he was mysteriously silenced.”

You’re right. Longroom.com is gone.

You’re also right in that we have no guarantee that the polls we read about are truthfully reported. At all.

H3llary is losing.


37 posted on 08/14/2016 2:56:15 PM PDT by combat_boots (MSM: We lie to you sheep at the slaughterhouse to keep you calm during slaughter)
[ Post Reply | Private Reply | To 35 | View Replies]

To: Crimson Elephant

Good post. I think you got almost all of it. One thing that one should add is that the response rate to any poll has declined over the years and is now (according to PEW) below 10%. (NB. Very rarely do the polling companies tell you how many interviews they have started, or how many phone numbers they have called).

As I have written elsewhere some research articles indicate that the errors in the polling have not increased despite the reduced response rate. That may well be true, but one has to be careful. Any election where there is more reason for voters for one candidate to not divulge their choice can cause very large discrepancies in the polling results.

Given the media’s relentless pursuit of Trump we may (cannot prove it, but we may) have such an election now. In that case the polls are not worth anything.


38 posted on 08/14/2016 2:57:25 PM PDT by ScaniaBoy (Part of the Right Wing Research & Attack Machine)
[ Post Reply | Private Reply | To 33 | View Replies]

To: GilGil

In November D+ will be 2+ or 3+ - it certainly won’t be 10+ or 16+

That’s wildly improbable. And even with the funny math, pollsters still can’t get Hillary over 50%.

Old rule of thumb an incumbent polling below 50% is usually going to lose.

CW though predicts a Hillary win in November so no pollster is going to go out on a limb and say the opposite.

So here we are.


39 posted on 08/14/2016 2:59:05 PM PDT by goldstategop ((In Memory Of A Dearly Beloved Friend Who Lives In My Heart Forever))
[ Post Reply | Private Reply | To 2 | View Replies]

To: BigEdLB

Interesting to break it down by state like that. For instance, Florida shows polling data is 4.8% oversampled for Dems based on State party affiliation. That’s significant, unless by some deeper measure one assumes registered Dems are almost 5% more likely to actually vote in Florida, which would seem unlikely.


40 posted on 08/14/2016 3:02:02 PM PDT by Flick Lives (TRIGGER WARNING - Posts may require application of sarcasm filter)
[ Post Reply | Private Reply | To 1 | View Replies]


Navigation: use the links below to view more comments.
first previous 1-2021-4041-6061-76 next last

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.

Free Republic
Browse · Search
News/Activism
Topics · Post Article

FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson