Polling, Will it ever be reliable again.

Does polling matter


  • Total voters
    60
#3
#3
It may have been earlier but the first major FURAR I remember was Florida exit polling in Bush Vs Gore which called Florida very early for Gore and eventually flipped back to Bush 2 months later.

now here we are again
 
  • Like
Reactions: headhunter15
#5
#5
So, I *think* the polling issue revolves around "shy voters".

To my knowledge, this hasn't been a probably historically, but in a nutshell there are some people who - when they receive calls / are interviewed by pollsters - don't want to admit that they are Trump supporters. These folks may not want to do so for a number of possible reasons:

(1) They are embarrassed about being Trump supporters, and don't want anyone to know;
(2) They are Trump supporters who are fearful of reciprocity in the form of threats, violence, etc. by non-Trump supporters, so don't want anyone to know;
(3) They are Trump supporters who knowingly and willfully want to mislead pollsters, thereby causing polls to underestimate their numbers.

Not sure how/if these issues could ever be addressed. I mean, this exact problem occurred in 2016, pollsters supposedly identified their issues, changed protocols, yet... here we are.

Polls still suck when it comes to measuring Trump's support. Doubt this will change, but hopefully, Donald's gone and never running for office again.
 
#8
#8
So, I *think* the polling issue revolves around "shy voters".

To my knowledge, this hasn't been a probably historically, but in a nutshell there are some people who - when they receive calls / are interviewed by pollsters - don't want to admit that they are Trump supporters. These folks may not want to do so for a number of possible reasons:

(1) They are embarrassed about being Trump supporters, and don't want anyone to know;
(2) They are Trump supporters who are fearful of reciprocity in the form of threats, violence, etc. by non-Trump supporters, so don't want anyone to know;
(3) They are Trump supporters who knowingly and willfully want to mislead pollsters, thereby causing polls to underestimate their numbers.

Not sure how/if these issues could ever be addressed. I mean, this exact problem occurred in 2016, pollsters supposedly identified their issues, changed protocols, yet... here we are.

Polls still suck when it comes to measuring Trump's support. Doubt this will change, but hopefully, Donald's gone and never running for office again.
He can run again in 24.
 
#9
#9
No. And it hasn’t been reliable in quite a while, since at least 2016 for the clear first big disconnect.

Read an interesting article on this. The current polling fiasco grew out of a frustration with people being frustrated with the MSM being out of touch with the American people. That’s where Silver and even Cohn came in. So they start their games and get lucky starting around 2008 I think. So the MSM co-ops their information and they get associated with the MSM... and people start not trusting them too preventing them from getting an accurate picture.

So no they aren’t trusted as they are associated with the MSM and people haven’t trusted them in a long time.
 
#10
#10
No and it was done to suppress votes.


Change my mind
I would concede no, but the suppress votes question seems irrational. Thinking that people will just fall in if the polls say one person or the other is a mistake. If you're candidate is losing in the polls you're much more motivated to turn out to support them, not just fall in. Skewed polls are I think a nonstarter for voter suppression because of this. There's no way to predict what effect the skew will have on voter turn out or who they vote for.
 
#14
#14
I would concede no, but the suppress votes question seems irrational. Thinking that people will just fall in if the polls say one person or the other is a mistake. If you're candidate is losing in the polls you're much more motivated to turn out to support them, not just fall in. Skewed polls are I think a nonstarter for voter suppression because of this. There's no way to predict what effect the skew will have on voter turn out or who they vote for.
Look at the third party vote argument. How many times do we hear it's a wasted vote? How many people actually prefer that third party over the two main ones but ends up voting for someone that " actually has a chance"?

That's a polling issue designed to suppress votes. We saw it in both 2016 and 2020 with third party candidates receiving enough votes to be included in the debates, but then the rules were changed so they couldnt.

People want to be on the winning side, it's why polls exist.
 
#15
#15
Honestly I’ve wondered if the polling error is largely driven by a lack of a reliable trustworthy input means for the pollsters.

I do believe these people are data driven. I’m sure they have their own biases but I’m going to give them the benefit of the doubt that they don’t want to be just shills and thus try to correctly model and analyze the data.

Thus if the polling results are **** and you consider the pollsters actually understand how to model data then the data must also be **** with regards to representing the actual ensemble of people it’s used to represent.

The primary means of collecting this data is the phone. But with the advent of the smart phone and proliferation of caller ID and other services like NoMoRobo you wind up with the pollsters unable to collect accurate data. And I don’t see how they can fix that unless people become willing to provide it. I don’t see that happening I know nobody in our house will talk to any of them and I don’t think we’re alone in that view. So unless they find a more reliable means to collect data theyre toast.

And it’s time to give Trafalgar and Rasmussen their due and recognize they have largely been successful in adapting to the current state of finding a means to collect properly correlated data. But even they still have some issues in the final numbers I think
 
#16
#16
Honestly I’ve wondered if the polling error is largely driven by a lack of a reliable trustworthy input means for the pollsters.

I do believe these people are data driven. I’m sure they have their own biases but I’m going to give them the benefit of the doubt that they don’t want to be just shills and thus try to correctly model and analyze the data.

Thus if the polling results are **** and you consider the pollsters actually understand how to model data then the data must also be **** with regards to representing the actual ensemble of people it’s used to represent.

The primary means of collecting this data is the phone. But with the advent of the smart phone and proliferation of caller ID and other services like NoMoRobo you wind up with the pollsters unable to collect accurate data. And I don’t see how they can fix that unless people become willing to provide it. I don’t see that happening I know nobody in our house will talk to any of them and I don’t think we’re alone in that view. So unless they find a more reliable means to collect data theyre toast.

And it’s time to give Trafalgar and Rasmussen their due and recognize they have largely been successful in adapting to the current state of finding a means to collect properly correlated data. But even they still have some issues in the final numbers I think
There is that but I do believe there is a good bit of confirmation bias as well from these guys. Instead of collecting evidence and creating a hypothesis from it, they are starting with a hypothesis and finding the evidence that fits it.

The land line thing is weak. It's the modern times, plenty of other ways to gather info. They arent leaving it because they were getting the info they wanted. They work for politicians and the MSM, not sure why you would assume they are neutral parties.
 
#18
#18
There is that but I do believe there is a good bit of confirmation bias as well from these guys. Instead of collecting evidence and creating a hypothesis from it, they are starting with a hypothesis and finding the evidence that fits it.

The land line thing is weak. It's the modern times, plenty of other ways to gather info. They arent leaving it because they were getting the info they wanted. They work for politicians and the MSM, not sure why you would assume they are neutral parties.
So my take away from your post is you aren’t willing to give them the benefit of the doubt in letting the data speak for itself. Basically a failure of critical thinking. And I think that’s probably a majority view just from observing this forum.

Everyday I’m paid to find a way to gather correlated data on an issue and try to put my own opinion on what is causing the problem aside and just let the data speak. It isn’t a natural human response I can assure you of that. But I have to make that attempt and if these pollsters are being professional then they are making that attempt also.
 
#20
#20
So my take away from your post is you aren’t willing to give them the benefit of the doubt in letting the data speak for itself. Basically a failure of critical thinking. And I think that’s probably a majority view just from observing this forum.

Everyday I’m paid to find a way to gather correlated data on an issue and try to put my own opinion on what is causing the problem aside and just let the data speak. It isn’t a natural human response I can assure you of that. But I have to make that attempt and if these pollsters are being professional then they are making that attempt also.
In my world there is any number of ways the same number can be looked at.

Square footage is a good for instance. There are two different ways to count it for building code, gross (everything) or net (occupiable). Then for various energy modeling there is occupied, unoccupied, full time use, part time use, activity, then you have heated & cooled vs built, then if you are dealing with a tenant that is a different "gross" and "net" that deals with leasable vs common. And then you get the funkiness of how do you count stairs? There is space below the stair you often cant use, how do you split the flights between floors. Is it all part of the level below, doubling that area of stair sf?

My point is you can gather honest information and present it any way you want based on who you are delivering it towards. Code officials get a very black white picture. Owners want some gray. Some owners want a very particular shade of gray.

I think in this case the people using the polls want a specific shade of gray. So it's not so much on the pollster. But how that info is used.
 
  • Like
Reactions: spyfish007
#22
#22
Polls matter, that’s statistics.

**** and biased polls are worthless. That’s what we’ve been getting
 
#23
#23
In my world there is any number of ways the same number can be looked at.

Square footage is a good for instance. There are two different ways to count it for building code, gross (everything) or net (occupiable). Then for various energy modeling there is occupied, unoccupied, full time use, part time use, activity, then you have heated & cooled vs built, then if you are dealing with a tenant that is a different "gross" and "net" that deals with leasable vs common. And then you get the funkiness of how do you count stairs? There is space below the stair you often cant use, how do you split the flights between floors. Is it all part of the level below, doubling that area of stair sf?

My point is you can gather honest information and present it any way you want based on who you are delivering it towards. Code officials get a very black white picture. Owners want some gray. Some owners want a very particular shade of gray.

I think in this case the people using the polls want a specific shade of gray. So it's not so much on the pollster. But how that info is used.
Ok so in your reply you’re saying the data is an absolute however the presentation of the results can and are manipulated? Agree 100%. In fact I make the joke quite often on “ok we’ve got the data. What do we want the answer to be” only some of that is tongue in cheek. But this isn’t necessarily a failing of the data set it’s rather the message and the messenger. That isn’t necessarily the pollsters in this example in fact it’s largely the MSM in my opinion.

Let’s look at some of the dialog that was going back and forth earlier this week.

A certain poster was touting this shiite data over and over with some made up “grade” of the poll and I still laugh at that idiocy logic. I reply cool story let’s see the MoE and sample size. And what extrapolations of the data sets to larger ensembles like a local state poll to a national poll. And while we’re at it let’s look at some correlation coefficients on their prior results taking poll questions and trying to infer a positive or negative view on the response they are trying to predict.

Most people aren’t going to even have a clue on the complicated math behind these polls that’s fine that’s not their job. But it isn’t the pollsters job either it’s their job to provide accurate polling estimates. The message and spin carried from that poll is what we are talking about now I think. And I blame the MSM largely for that.
 
Last edited:
#24
#24
Ok so in your reply you’re saying the data is an absolute however the presentation of the results can abs are manipulated? Agree 100%. In fact I make the joke quite often on “ok we’ve got the data. What do we want the answer to be” only some of that is tongue in cheek. But this isn’t necessarily a failing of the data set it’s rather the message and the messenger. That isn’t necessarily the pollsters in this example in fact it’s largely the MSM in my opinion.

Let’s look at some of the dialog that was going back and forth earlier this week.

A certain poster was touting this shiite data over and over with some made up “grade” of the poll and I still laugh at that idiocy logic. I reply cool story let’s see the MoE and sample size. And what extrapolations of the data sets to larger ensembles like a local state poll to a national poll. And while we’re at it let’s look at some correlation coefficients on their prior results taking poll questions and trying to infer a positive or negative view on the response they are trying to predict.

Most people aren’t going to even have a clue on the complicated math behind these polls that’s fine that’s not their job. But it isn’t the pollsters job either it’s their job to provide accurate polling estimates. The message and spin carried from that poll is what we are talking about now I think. And I blame the MSM largely for that.
I think for some of these polls there is little difference, now, between them and the MSM they have worked with for however long.
 
  • Like
Reactions: tbwhhs
#25
#25
Polls are extremely useful and an utterly reliable method of showing dumb people will pay good money for substandard services and continue throwing money away on it.

PT Barnum was absolutely correct.
 

VN Store



Back
Top