• This is a political forum that is non-biased/non-partisan and treats every person's position on topics equally. This debate forum is not aligned to any political party. In today's politics, many ideas are split between and even within all the political parties. Often we find ourselves agreeing on one platform but some topics break our mold. We are here to discuss them in a civil political debate. If this is your first visit to our political forums, be sure to check out the RULES. Registering for debate politics is necessary before posting. Register today to participate - it's free!

Are polls very accurate anymore?

EnigmaO01

EnigmaO01

DP Veteran
Joined
Dec 5, 2009
Messages
23,489
Reaction score
17,887
Location
Indiana
Gender
Male
Political Leaning
Other
Can the pollsters get good sampling when a significant part of the electorate no longer uses a landline phone, therefore making it next to impossible to get cell phone numbers?

Thoughts?
 
EnigmaO01 said:
Can the pollsters get good sampling when a significant part of the electorate no longer uses a landline phone, therefore making it next to impossible to get cell phone numbers?

Thoughts?
Click to expand...

I don't think polls were ever very accurate.

I think statistics, at least when applied to something as individually diverse as people, have almost no validity.

Such use reminds me of Psychohistory*, expounded in Isaac Asimov's "Foundation" series.

* https://en.wikipedia.org/wiki/Psychohistory_(fictional)

I really find it amazing how we have bought into the "science" of taking a tiny sample and projecting the results onto the mass of humanity it comes from.
 
Last edited:
EnigmaO01 said:
Can the pollsters get good sampling when a significant part of the electorate no longer uses a landline phone, therefore making it next to impossible to get cell phone numbers?

Thoughts?
Click to expand...

Polls aren't done in a vacuum. There are zillions of variables used for adjusting results to reflect demographics. They have volumes of info on sex, age group, occupation, education, and so on. They adjust for who uses a land line rather than a cell phone. Volumes of studies on what age group answers most truthfully...who answers the phone at 6 pm as opposed to 8 pm. And on and on. Ethnic groups, church goers, union members, single mothers, etc.

If you polled people as to what beverage they were drinking you'd know not to equate a school room cafeteria at noon with a wine bar at night. You would have means of getting a more accurate cross section of the population.

Polling companies would be out of business long ago if they were not ballpark accurate.
 
Last edited:
Most reputable pollsters are substantially accurate. The problem hits whenever a single entity has the wrong framework in mind when drafting its sample. When there's an obviously different response (i.e. completely unexpected politician wins handedly on election day), then the pollsters need to rework their framework.
 
Captain Adverse said:
I don't think polls were ever very accurate.

I think statistics, at least when applied to something as individually diverse as people, have almost no validity.

Such use reminds me of Psychohistory*, expounded in Isaac Asimov's "Foundation" series.

* https://en.wikipedia.org/wiki/Psychohistory_(fictional)

I really find it amazing how we have bought into the "science" of taking a tiny sample and projecting the results onto the mass of humanity it comes from.
Click to expand...

How have the polls been generally accurate in the past then?
 
Anagram said:
How have the polls been generally accurate in the past then?
Click to expand...

I don't know. I don't follow them (although I don't mind quoting them back to people who like to use selected parts of them). Have they always been then?
 
Captain Adverse said:
I don't know. I don't follow them (although I don't mind quoting them back to people who like to use selected parts of them). Have they always been then?
Click to expand...

They've generally been pretty close. Almost always within the margin of error, especially when aggregated. For examples:

RealClearPolitics - Election 2008 National Head-to-Head Polls

RealClearPolitics - 2006 Senate RealClearPolitics Poll Averages

RealClear Politics - Polls

And I mean the math is solid behind polls, so long as they can get a representative sample. Something that may be becoming harder with cell phone use and Caller ID, but hasn't been a problem in the past.
 
Anagram said:
They've generally been pretty close. Almost always within the margin of error, especially when aggregated. For examples:

RealClearPolitics - Election 2008 National Head-to-Head Polls

RealClearPolitics - 2006 Senate RealClearPolitics Poll Averages

RealClear Politics - Polls

And I mean the math is solid behind polls, so long as they can get a representative sample. Something that may be becoming harder with cell phone use and Caller ID, but hasn't been a problem in the past.
Click to expand...

In the aggregate. Margin of error.

What's with all the variations? Why aren't they all the same? The polls range 5 to 8 points different between the lowest and highest of the poll services, and the margins for error are even more varied. So you aggregate them all and as long as they bracket the actual result that makes them accurate? :confused:

Seems to me that this allows people to cherry-pick to find whichever poll provides the greatest support for their point when dealing with issues that don't have a clear final result (like election's do).
 
Captain Adverse said:
In the aggregate. Margin of error.

What's with all the variations? Why aren't they all the same? The polls range 5 to 8 points different between the lowest and highest of the poll services, and the margins for error are even more varied. So you aggregate them all and as long as they bracket the actual result that makes them accurate? :confused:

Seems to me that this allows people to cherry-pick to find whichever poll provides the greatest support for their point when dealing with issues that don't have a clear final result (like election's do).
Click to expand...

I disagree strongly. Polls have margins of error because they only are picking out a sample. If you have a representative sample, the MOE is easily calculated, and is 100% correct. That doesn't mean the poll is worth it though. Aggregating polls can come up with an even better result by reducing the influence of out-liars on the farther ends of the MOE, because the polls will tend to clump around the actual number.
 
Anagram said:
I disagree strongly. Polls have margins of error because they only are picking out a sample. If you have a representative sample, the MOE is easily calculated, and is 100% correct. That doesn't mean the poll is worth it though. Aggregating polls can come up with an even better result by reducing the influence of out-liars on the farther ends of the MOE, because the polls will tend to clump around the actual number.
Click to expand...

Okay, sadly I must take the standard exit strategy and fall back on my atheist argument.

I simply don't believe the things you do about polling. It's a matter of personal dis-belief. ;)
 
Captain Adverse said:
Okay, sadly I must take the standard exit strategy and fall back on my atheist argument.

I simply don't believe the things you do about polling. It's a matter of personal dis-belief. ;)
Click to expand...

Suit yourself.
 
Captain Adverse said:
Okay, sadly I must take the standard exit strategy and fall back on my atheist argument.

I simply don't believe the things you do about polling. It's a matter of personal dis-belief. ;)
Click to expand...

Well listen, you said "I really find it amazing how we have bought into the "science" of taking a tiny sample and projecting the results onto the mass of humanity it comes from." But think about what must have occurred in order for, literally, trillions of economic dollars and thousands of man hours to be devoted to the notion of sampling and statistical predictions. In order for people to have spent that much money - you would need to see results for your input - in essence, proof that the science of statistics work.
 
MrT said:
Well listen, you said "I really find it amazing how we have bought into the "science" of taking a tiny sample and projecting the results onto the mass of humanity it comes from." But think about what must have occurred in order for, literally, trillions of economic dollars and thousands of man hours to be devoted to the notion of sampling and statistical predictions. In order for people to have spent that much money - you would need to see results for your input - in essence, proof that the science of statistics work.
Click to expand...

Once upon a time, when I was a very young man, I went to college and took Statistics in lieu of other math courses because...I hate math. ;)

I did learn some interesting things about Statistics which have colored my opinions ever since. The most important thing being; one can set up conditions and then use the same data to prove just about anything anyone wants to prove. It is all in how you poll and then present the results.

After that, I have looked askance at any polls projecting data to support a political idea.
 
Last edited:
EnigmaO01 said:
Can the pollsters get good sampling when a significant part of the electorate no longer uses a landline phone, therefore making it next to impossible to get cell phone numbers?

Thoughts?
Click to expand...

There's no evidence that this has occurred yet. Recent polls appear to be just as reliable as they have been in the past, if not more so. Logic tells me you're right and it's juts a matter of time before polls go downhill if they don't make changes in how they sample soon. But so far there's no evidence this has occurred yet, or at least none that I know of.
 
Captain Adverse said:
Okay, sadly I must take the standard exit strategy and fall back on my atheist argument.

I simply don't believe the things you do about polling. It's a matter of personal dis-belief. ;)
Click to expand...

You don't believe in math?
 
Captain Adverse said:
Once upon a time, when I was a very young man, I went to college and took Statistics in lieu of other math courses because...I hate math. ;)

I did learn some interesting things about Statistics which have colored my opinions ever since. The most important thing being; one can set up conditions and then use the same data to prove just about anything anyone wants to prove. It is all in how you poll and then present the results.

After that, I have looked askance at any polls projecting data to support a political idea.
Click to expand...

I think there's a German quote that says something like the following: Don't believe any statistic that you haven't falsified yourself. LOL
 
Polls are generally used to make a point. How far (honesty wise) the pollsters will go to make that point is debatable at best.

As a non-political example:
I read some time ago (so long ago that finding the article is probably a point of exasperation) that "the viewers of My Bloody Valentine 3D were mostly women of ages 20-30"
So. Did someone stand outside and count every person who went into the theater? id they guess at ages?
Did they poll one select group or ask people to participate in a survey? How did they ensure their control group represented the entire country?

I think if someone has a particular agenda to prove, "polls" are an easy tool to make "facts" sound "accurate".
Lots of quotes in there and for good reason.
I am an advocate of don't trust before verifying.

As an aside. I was hoping there would be a poll that went with this question. :D
 
Ntharotep said:
Polls are generally used to make a point. How far (honesty wise) the pollsters will go to make that point is debatable at best.

As a non-political example:
I read some time ago (so long ago that finding the article is probably a point of exasperation) that "the viewers of My Bloody Valentine 3D were mostly women of ages 20-30"
So. Did someone stand outside and count every person who went into the theater? id they guess at ages?
Did they poll one select group or ask people to participate in a survey? How did they ensure their control group represented the entire country?

I think if someone has a particular agenda to prove, "polls" are an easy tool to make "facts" sound "accurate".
Lots of quotes in there and for good reason.
I am an advocate of don't trust before verifying.

As an aside. I was hoping there would be a poll that went with this question. :D
Click to expand...

Polls are very rarely used to make a point. 99.9% of polling is used for market research and feedback on customer satisfaction, etc. Governments, corporations, media, advocacy groups, religious institutions, academia use polls and their expectation is reliable figures in to use a tool for some type of planning, marketing, policy making.
 
You must log in or register to reply here.
Back
Top Bottom