Warning

 

Close

Confirm Action

Are you sure you wish to do this?

Confirm Cancel
BCM
User Panel

Posted: 10/24/2016 9:46:54 PM EDT
I keep hearing, "Well, there are more registered Democrats" as the reason, but I don't really understand why that is. I am fairly good a teh maths, and have a layman's understanding of statistics, but I can't understand why you would alter the samples. If you poll 10,000 people, wouldn't you pretty much get the correct numbers of the various party supporters? i.e. if there are 30% Democrats and 25% Republicans in a group, my naive understanding of statistics has me thinking, in a decent sample, you would hit around 30% Democrats, 25% Republicans, etc.



Can someone with a better understanding of this explain the reasoning? Is it to compensate for using smaller sample groups? Compensating for selecting "likely voters" instead of just random people? Something else?


Link Posted: 10/24/2016 9:50:36 PM EDT
[#1]
Technically, It's not lying if you bias the poll to make you say what you want it to say.
Link Posted: 10/24/2016 9:54:23 PM EDT
[#2]
I took a political science class in college that largely focused on polling, and Lies, Bullshit, and Statistics came up a lot.

Remember when they sampled tons of random people for exit polls in 2004 and it showed Kerry destroyed Bush?

Turned out that Democrats were far more likely to respond to their polls far more than Republicans, so they oversampled the Dems. Stuff like that happens often when you go pure random.

The goal of the poll is to make the poll match your predicted voter turnout, and that's how they bull shit the polls. They're oversampling Democrats because they're pretending turnout will match 2012. Democrats are going to stay home for the crusty, corrupt, caucasian grandmother.
Link Posted: 10/24/2016 9:59:17 PM EDT
[#3]

Quoted:


I keep hearing, "Well, there are more registered Democrats" as the reason, but I don't really understand why that is. I am fairly good a teh maths, and have a layman's understanding of statistics, but I can't understand why you would alter the samples. If you poll 10,000 people, wouldn't you pretty much get the correct numbers of the various party supporters? i.e. if there are 30% Democrats and 25% Republicans in a group, my naive understanding of statistics has me thinking, in a decent sample, you would hit around 30% Democrats, 25% Republicans, etc.



Can someone with a better understanding of this explain the reasoning? Is it to compensate for using smaller sample groups? Compensating for selecting "likely voters" instead of just random people? Something else?

View Quote
There are a lot of independent voters that were republican at one time.  They over sample democrats to balance the polls.  You need to ask a statistician who specializes in political polling to get a confirmed answer.
Link Posted: 10/24/2016 9:59:35 PM EDT
[#4]
PSYOPS
Link Posted: 10/24/2016 10:00:04 PM EDT
[#5]
Discussion ForumsJump to Quoted PostQuote History
Quoted:
I took a political science class in college that largely focused on polling, and Lies, Bullshit, and Statistics came up a lot.

Remember when they sampled tons of random people for exit polls in 2004 and it showed Kerry destroyed Bush?

Turned out that Democrats were far more likely to respond to their polls far more than Republicans, so they oversampled the Dems. Stuff like that happens often when you go pure random.

The goal of the poll is to make the poll match your predicted voter turnout, and that's how they bull shit the polls. They're oversampling Democrats because they're pretending turnout will match 2012. Democrats are going to stay home for the crusty, corrupt, caucasian grandmother.
View Quote


I don't know, the biased polls are doing what they intended and disenfranchising voters. You can even see it here where the leaked emails have been posted about them intentionally biasing polls. The average voter isn't as well informed and some are sure to be swayed into not voting thinking it is a lost cause.
Link Posted: 10/24/2016 10:01:39 PM EDT
[#6]
Discussion ForumsJump to Quoted PostQuote History
Quoted:


I don't know, the biased polls are doing what they intended and disenfranchising voters. You can even see it here where the leaked emails have been posted about them intentionally biasing polls. The average voter isn't as well informed and some are sure to be swayed into not voting thinking it is a lost cause.
View Quote View All Quotes
View All Quotes
Discussion ForumsJump to Quoted PostQuote History
Quoted:
Quoted:
I took a political science class in college that largely focused on polling, and Lies, Bullshit, and Statistics came up a lot.

Remember when they sampled tons of random people for exit polls in 2004 and it showed Kerry destroyed Bush?

Turned out that Democrats were far more likely to respond to their polls far more than Republicans, so they oversampled the Dems. Stuff like that happens often when you go pure random.

The goal of the poll is to make the poll match your predicted voter turnout, and that's how they bull shit the polls. They're oversampling Democrats because they're pretending turnout will match 2012. Democrats are going to stay home for the crusty, corrupt, caucasian grandmother.


I don't know, the biased polls are doing what they intended and disenfranchising voters. You can even see it here where the leaked emails have been posted about them intentionally biasing polls. The average voter isn't as well informed and some are sure to be swayed into not voting thinking it is a lost cause.

Hopefully Trump calling BS will help offset a large percentage.
Link Posted: 10/24/2016 10:02:19 PM EDT
[#7]
I just read through the numbers behind a Pennsylvania poll that has her up 4% (margin of error was 3.2). Aside from a disproportionate number of Democrats polled the data was "weighted based on the 2012 Presidential Election." I think they're in for quite a surprise.
Link Posted: 10/24/2016 10:04:46 PM EDT
[#8]
Discussion ForumsJump to Quoted PostQuote History
Quoted:
There are a lot of independent voters that were republican at one time.  They over sample democrats to balance the polls.  You need to ask a statistician who specializes in political polling to get a confirmed answer.
View Quote View All Quotes
View All Quotes
Discussion ForumsJump to Quoted PostQuote History
Quoted:
Quoted:
I keep hearing, "Well, there are more registered Democrats" as the reason, but I don't really understand why that is. I am fairly good a teh maths, and have a layman's understanding of statistics, but I can't understand why you would alter the samples. If you poll 10,000 people, wouldn't you pretty much get the correct numbers of the various party supporters? i.e. if there are 30% Democrats and 25% Republicans in a group, my naive understanding of statistics has me thinking, in a decent sample, you would hit around 30% Democrats, 25% Republicans, etc.

Can someone with a better understanding of this explain the reasoning? Is it to compensate for using smaller sample groups? Compensating for selecting "likely voters" instead of just random people? Something else?
There are a lot of independent voters that were republican at one time.  They over sample democrats to balance the polls.  You need to ask a statistician who specializes in political polling to get a confirmed answer.


They aren't oversampling Democrats at rates double or triple what Democrats showed for Obama against Romney because they are trying to balance the poll and oversampling Democrats isn't the only trick they are using. That poll that showed Clinton 12 points ahead asked for the youngest in the house over 18 and asked 87 negative questions about Trump to get the numbers they did.
Link Posted: 10/24/2016 10:06:58 PM EDT
[#9]
Bias and oversampling are two different statistical terms. In short, oversampling is bad and bias is just a tool. From a math point, bias is used to manipulate your data to try and make it more accurate. If I take a sample from Seattle and try to extrapolate it to the rest of the state, I'll add a bias to it based on previous voting data (because we know Seattle is heavy Democrat but the whole state isn't). So bias isn't always bad, but it can be when deliberately used to skew the data to create the result that you want.
Link Posted: 10/24/2016 10:07:15 PM EDT
[#10]
Discussion ForumsJump to Quoted PostQuote History
Quoted:
I took a political science class in college that largely focused on polling, and Lies, Bullshit, and Statistics came up a lot.

Remember when they sampled tons of random people for exit polls in 2004 and it showed Kerry destroyed Bush?

Turned out that Democrats were far more likely to respond to their polls far more than Republicans, so they oversampled the Dems. Stuff like that happens often when you go pure random.

The goal of the poll is to make the poll match your predicted voter turnout, and that's how they bull shit the polls. They're oversampling Democrats because they're pretending turnout will match 2012. Democrats are going to stay home for the crusty, corrupt, caucasian grandmother.
View Quote

Bingo.
Link Posted: 10/24/2016 10:14:00 PM EDT
[#11]
Discussion ForumsJump to Quoted PostQuote History
Quoted:


They aren't oversampling Democrats at rates double or triple what Democrats showed for Obama against Romney because they are trying to balance the poll and oversampling Democrats isn't the only trick they are using. That poll that showed Clinton 12 points ahead asked for the youngest in the house over 18 and asked 87 negative questions about Trump to get the numbers they did.
View Quote View All Quotes
View All Quotes
Discussion ForumsJump to Quoted PostQuote History
Quoted:
Quoted:
Quoted:
I keep hearing, "Well, there are more registered Democrats" as the reason, but I don't really understand why that is. I am fairly good a teh maths, and have a layman's understanding of statistics, but I can't understand why you would alter the samples. If you poll 10,000 people, wouldn't you pretty much get the correct numbers of the various party supporters? i.e. if there are 30% Democrats and 25% Republicans in a group, my naive understanding of statistics has me thinking, in a decent sample, you would hit around 30% Democrats, 25% Republicans, etc.

Can someone with a better understanding of this explain the reasoning? Is it to compensate for using smaller sample groups? Compensating for selecting "likely voters" instead of just random people? Something else?
There are a lot of independent voters that were republican at one time.  They over sample democrats to balance the polls.  You need to ask a statistician who specializes in political polling to get a confirmed answer.


They aren't oversampling Democrats at rates double or triple what Democrats showed for Obama against Romney because they are trying to balance the poll and oversampling Democrats isn't the only trick they are using. That poll that showed Clinton 12 points ahead asked for the youngest in the house over 18 and asked 87 negative questions about Trump to get the numbers they did.


When you don't want to upset Komrade Klinton you bring her favorable results and numbers no matter what and find ways to do that. Ironically that will be there downfall because it's keeps them out of touch of the voters.
Link Posted: 10/24/2016 10:17:21 PM EDT
[#12]
Discussion ForumsJump to Quoted PostQuote History
Quoted:
Technically, It's not lying if you bias the poll to make you say what you want it to say.
View Quote


They will keep polling until they get their narratives to align.
Link Posted: 10/24/2016 10:21:35 PM EDT
[#13]
Another reason they do it is to make borderline voters feel like their vote won't matter for the other candidate since it's "game over".  Keep them home.
Link Posted: 10/24/2016 11:45:59 PM EDT
[#14]
Modern day polling is not done to measure voters choices, its done to sway voters choices.
Link Posted: 10/25/2016 4:44:36 PM EDT
[#15]
Thanks for the insight on this, guys. So the reason it's not simply dismissed as chicanery is that there ARE some oddities it might compensate for, but its far more likely to be simply to produce a desired result.


Link Posted: 10/25/2016 4:48:03 PM EDT
[#16]
Wag the dog.

Tell people what you want them to hear.
Link Posted: 10/25/2016 4:57:43 PM EDT
[#17]
To lend plausibility to the stolen election and the fraudulent voting machine returns.
Link Posted: 10/25/2016 5:18:25 PM EDT
[#18]
Theory is people back a winner. So if Hillary is thought to be the winner, people on the fence will vote for her.  So you squew the pols to get the result you want.
Close Join Our Mail List to Stay Up To Date! Win a FREE Membership!

Sign up for the ARFCOM weekly newsletter and be entered to win a free ARFCOM membership. One new winner* is announced every week!

You will receive an email every Friday morning featuring the latest chatter from the hottest topics, breaking news surrounding legislation, as well as exclusive deals only available to ARFCOM email subscribers.


By signing up you agree to our User Agreement. *Must have a registered ARFCOM account to win.
Top Top