Polling Methodology

[Luvnit2]

https://poll.qu.edu/methodology/

I hear polling is not accurate as they only call landlines. That is not correct, all legit polling places call cell phones and landlines. They also call likely voters as elections get closer.


We do call cell phones. This is increasingly important as more than half of the nation only have a cell phone and no landline telephone.

Established in 1988, the Quinnipiac University Poll is a nationally and internationally recognized polling organization that adheres to rigorous methodological practices.

[Arturo Bandini]

The best poll is the one coming out of the ballot boxes.

[KillerGoose]

(11-27-2023, 04:22 PM)Luvnit2 Wrote: https://poll.qu.edu/methodology/

I hear polling is not accurate as they only call landlines. That is not correct, all legit polling places call cell phones and landlines. They also call likely voters as elections get closer.


We do call cell phones. This is increasingly important as more than half of the nation only have a cell phone and no landline telephone.

Established in 1988, the Quinnipiac University Poll is a nationally and internationally recognized polling organization that adheres to rigorous methodological practices.

I myself have been polled, albeit not for a presidential election. It was via text but I have also been called on my cell phone. 


Polling is fairly accurate if the person understands margins for error. For instance, in 2016 538 had Hillary as the favorite but she was only polling as roughly a four point favorite. The standard margin for error is five points. Due to this, she was only given a 71% chance of victory and of course, she lost. That isn't that surprising. On the other end of the spectrum, Biden was around a eight point favorite. That is the polls telling you of something much more certain, and he was given a 90% chance of victory. 

Right now, the polls are very bad for Biden. I think Trump would be the favorite if the election were today. However, there is still quite a bit of time left. I think one thing that could end up helping Biden is the fact that right now Trump is in the background. As 2024 progresses, he is going to have to jump more into the public eye which could potentially hurt him. If anything is certain, it is that most Americans strongly dislike Donald Trump. He is not a popular man by any means, and most Americans don't believe he was a good President. It is possible that him jumping back into the spotlight galvanizes the liberal base again, just enough to give Biden a second term. It doesn't look great right now, though. 

[Luvnit2]

(11-27-2023, 04:42 PM)KillerGoose Wrote: I myself have been polled, albeit not for a presidential election. It was via text but I have also been called on my cell phone. 


Polling is fairly accurate if the person understands margins for error. For instance, in 2016 538 had Hillary as the favorite but she was only polling as roughly a four point favorite. The standard margin for error is five points. Due to this, she was only given a 71% chance of victory and of course, she lost. That isn't that surprising. On the other end of the spectrum, Biden was around a eight point favorite. That is the polls telling you of something much more certain, and he was given a 90% chance of victory. 

Right now, the polls are very bad for Biden. I think Trump would be the favorite if the election were today. However, there is still quite a bit of time left. I think one thing that could end up helping Biden is the fact that right now Trump is in the background. As 2024 progresses, he is going to have to jump more into the public eye which could potentially hurt him. If anything is certain, it is that most Americans strongly dislike Donald Trump. He is not a popular man by any means, and most Americans don't believe he was a good President. It is possible that him jumping back into the spotlight galvanizes the liberal base again, just enough to give Biden a second term. It doesn't look great right now, though. 

Each poll posts own margin of error, normally in range of 3 to 4%.

I agree it is very early. I agree we do not know the impact of the trials will have on Trump or the impeachment (it is coming in 2024) of Biden.

I just wanted to clarify the false comments made in this forum on stating polls are not accurate and only use land lines to gather info.

As for current state, if I were a Democrat, I would be very concerned Haley is beating Biden way outside the margins of errors and the past 10 polls, Trump is beating Biden some within margin of error, others outside margin of error. 

I also agree, the only poll that matters in the end is a vote as mentioned by anther poster.

[*pally*]

And it is not just the methodology of how they call, though I would point out that people increasingly ignore calls and texts from phone numbers they don't recognize. The other often ignored part of polling is how they weigh the responses to take the limited number of responses per category of respondents into a "representative" percentage nationwide. For example, if a polling company decided that 10% of 18-25-year-olds actually vote so that is how they weigh their poll interpretation that number can prove to be totally bogus if the actual voting percentage of young voters is 20%. The other way polls get skewed is by how the questions are worded or by questions not asked for clarification of results.

[*pally*]

(11-27-2023, 04:51 PM)Luvnit2 Wrote: Each poll posts own margin of error, normally in range of 3 to 4%.

I agree it is very early. I agree we do not know the impact of the trials will have on Trump or the impeachment (it is coming in 2024) of Biden.

I just wanted to clarify the false comments made in this forum on stating polls are not accurate and only use land lines to gather info.

As for current state, if I were a Democrat, I would be very concerned Haley is beating Biden way outside the margins of errors and the past 10 polls, Trump is beating Biden some within margin of error, others outside margin of error. 

I also agree, the only poll that matters in the end is a vote as mentioned by anther poster.

why would we worry about Haley when right now Trump is clearly going to be the Republican candidate?

[Belsnickel]

I harp on this topic a lot. Polling methodology varies from pollster to pollster. Some are better than others. Where polling breaks down is how they weight things, their sample sizes, how they sample, etc. Not all polls are created equal. In addition, how we interpret polling information is also important. The timing of polls in relation to an election is important. National polls for POTUS a year out from the election, for instance, are worthless beyond seeking to understand the issues the electorate is concerned about since there is a lot that can happen in a year and, much to my dismay, POTUS is not a national popular election.

One of the biggest things to understand about polling is to never rely on popular media reporting on polling information. They are terrible at it. Even looking at the toplines from the pollster can result in a distorted understanding, but if you don't understand the details of polls and statistics, looking deeper can be disorienting.

[SunsetBengal]

(11-28-2023, 09:53 AM)Belsnickel Wrote: I harp on this topic a lot. Polling methodology varies from pollster to pollster. Some are better than others. Where polling breaks down is how they weight things, their sample sizes, how they sample, etc. Not all polls are created equal. In addition, how we interpret polling information is also important. The timing of polls in relation to an election is important. National polls for POTUS a year out from the election, for instance, are worthless beyond seeking to understand the issues the electorate is concerned about since there is a lot that can happen in a year and, much to my dismay, POTUS is not a national popular election.

One of the biggest things to understand about polling is to never rely on popular media reporting on polling information. They are terrible at it. Even looking at the toplines from the pollster can result in a distorted understanding, but if you don't understand the details of polls and statistics, looking deeper can be disorienting.

You've done a great job of breaking down the minutia of polling and how their methodology can be skewed to obtain a somewhat desired result.  Likely the biggest thing that I have personally taken from your advice is to reference the 538 when I read or hear about a poll with 'surprising' results, as the average of the trusted polls is likely closest to the reality. 

[Dill]

(11-27-2023, 04:30 PM)Arturo Bandini Wrote: The best poll is the one coming out of the ballot boxes.

According to Trump and the GOP, 

that one is LEAST accurate. 

[Luvnit2]

(11-27-2023, 07:04 PM)pally Wrote: why would we worry about Haley when right now Trump is clearly going to be the Republican candidate?

Trump's VP at a minimum and if for some reason Trump's numbers go south, then she will be the nominee.

I still say Biden gets replaced at the DNC convention, unless his numbers go up significantly. If he is polling like this in June, he is gone.

[Luvnit2]

(11-28-2023, 10:03 AM)SunsetBengal Wrote: You've done a great job of breaking down the minutia of polling and how their methodology can be skewed to obtain a somewhat desired result.  Likely the biggest thing that I have personally taken from your advice is to reference the 538 when I read or hear about a poll with 'surprising' results, as the average of the trusted polls is likely closest to the reality. 

You keep mentioning this, but show us one time the Rea Clear Politics averages has missed the margin of error.

What is the 538 polling results for 2020 Presidential race? Did they beat RCP numbers for POTUS? Are there any shared polling sites between RCP and 538? My guess is yes.