PDA

View Full Version : Why Are Presidential Polls Wrong?


Yar-El
10-12-2008, 04:50 PM
Artilce: Why Presidential Polls Are Wrong (http://www.livescience.com/history/080109-bad-polls.html)

The larger problem with many of today's political pollsters is that surveys are conducted in affiliation with media organizations, said Shawn Parry-Giles, a political communications professor at the University of Maryland who camped out in New Hampshire prior to the primary to make observations.

"Media aren't going to be self-reflexive about their poll," Parry-Giles said. "The journalists themselves just bought into the fact that [Obama] was so far ahead and it was inevitable. I was stunned by the coverage."

The media should stop treating polls as if they are factual information, she said.

"This is about what the voters say and do, and media has to be very careful about how they frame the polls," Parry-Giles told LiveScience.

One poll by CNN/WMUR/UNH on the anticipated results in New Hampshire had a relatively small sample size (which cripples a survey's accuracy) and a fairly large margin of error, but it was reported as accurate and went unquestioned, she said I thought we could talk about the credibility of polls. Recent history has shown the media has been wrong. Durring the primary races, media groups missed almost every polling mark. How credible are these polls?

Litofsky
10-12-2008, 05:04 PM
It all depends on the interviewees. I seem to recall that, in 1936, many people thought that Alf Landon was going to win: the polls all said that he was leading. But, as was later learned, the people interviewed were mostly Landon supporters.

I believe that the polls hold partial truth- I don't believe that California is going to suddenly turn Republican, or that Obama will be defeated in a landslide. Actually, I believe that this will be one of the closest elections in this history of the United States. I do believe that Obama has an edge, though: we're coming off one of the worst Presidents in our history (even if you don't view Bush as bad, most Americans blame him for the economic problems), and he's a Republican.

Therefore, I believe that Obama is winning, but that polls don't always tell the truth.

Achilles
10-12-2008, 05:26 PM
Ignore the polls for a moment and look at data that is a little less subjective.

Intrade.com (http://www.intrade.com/?request_operation=main&request_type=action&checkHomePage=true). Why is this better than a poll? Because people actually spend real money (not just free opinions) to indicate their support for a candidate.

Electoral Maps (http://www.realclearpolitics.com/epolls/maps/obama_vs_mccain/). Why are these better than polls? Because polls fluctuate and readers get caught up in 1-2 point shifts which sometimes have absolutely 0 effect on how the electoral votes will go down.

All these news outlets are vying for subscribers, viewers, etc, so they spend a lot of time crunching these numbers or those numbers, building fancy maps, or running silly simulations. Why? Because they need to keep you watching.

Most of the politicos that I listen to seem to agree that watching the horse race is little more than brain candy for the masses. Besides, I firmly believe that you need some education in survey techniques to be able to truly understand most of what's being said anyway. So and so says X, but unless you can look at polling methods, sample size, know how to analyze for bias, etc, it's difficult to understand what X is really telling you. And that's why reliable poll numbers are so elusive.

Yar-El
10-12-2008, 08:43 PM
Ignore the polls for a moment and look at data that is a little less subjective.

Intrade.com (http://www.intrade.com/?request_operation=main&request_type=action&checkHomePage=true). Why is this better than a poll? Because people actually spend real money (not just free opinions) to indicate their support for a candidate.

Electoral Maps (http://www.realclearpolitics.com/epolls/maps/obama_vs_mccain/). Why are these better than polls? Because polls fluctuate and readers get caught up in 1-2 point shifts which sometimes have absolutely 0 effect on how the electoral votes will go down.

All these news outlets are vying for subscribers, viewers, etc, so they spend a lot of time crunching these numbers or those numbers, building fancy maps, or running silly simulations. Why? Because they need to keep you watching.

Most of the politicos that I listen to seem to agree that watching the horse race is little more than brain candy for the masses. Besides, I firmly believe that you need some education in survey techniques to be able to truly understand most of what's being said anyway. So and so says X, but unless you can look at polling methods, sample size, know how to analyze for bias, etc, it's difficult to understand what X is really telling you. And that's why reliable poll numbers are so elusive. Poll mechanics? How a poll is developed through accumulating data. Demographics and such, right?

KinchyB
10-12-2008, 08:46 PM
Poll mechanics? How a poll is developed through accumulating data. Demographics and such, right?

Believe that's what he was referring to. For example (and please correct if i'm wrong) Rasmussen uses a completely automated phone polling system. As a result there is really no way to guarantee that you are talking to an actual voter, or even someone who is capable of voting. It could just be an 8 year old kid hitting buttons for fun. :xp:

Yar-El
10-12-2008, 08:52 PM
Believe that's what he was referring to. For example (and please correct if i'm wrong) Rasmussen uses a completely automated phone polling system. As a result there is really no way to guarantee that you are talking to an actual voter, or even someone who is capable of voting. It could just be an 8 year old kid hitting buttons for fun. :xp: :lol:
Recording, "Do you believe in world domination? Press#1"
Kid holding the phone, "Boop"
:lol:

Achilles
10-12-2008, 09:10 PM
Believe that's what he was referring to. For example (and please correct if i'm wrong) Rasmussen uses a completely automated phone polling system. As a result there is really no way to guarantee that you are talking to an actual voter, or even someone who is capable of voting. It could just be an 8 year old kid hitting buttons for fun. :xp:Demographics, sample size, appropriate margins of error, etc.

The Rasmussen example is the easiest to exemplify. Another well cited example is the fact that most telephone surveys do not include data from cell phone only households, which tend to be younger, more affluent, etc. So how reliable are your results if you have an entire demographic being under-represented? Most pollsters say that they account for this in their equations, however how they do that is unknown to us (not to mention probably incomprehensible to the vast majority of Americans anyway), so how can we gauge how well or how poorly they do this?

And this is just one of the more well known considerations. There are lots and lots of others. So yeah, polls are nice and the horse race is fun, but at the end of the day it's almost entirely useless for predicting outcomes.

Litofsky
10-12-2008, 09:36 PM
The Rasmussen example is the easiest to exemplify. Another well cited example is the fact that most telephone surveys do not include data from cell phone only households, which tend to be younger, more affluent, etc. So how reliable are your results if you have an entire demographic being under-represented? Most pollsters say that they account for this in their equations, however how they do that is unknown to us (not to mention probably incomprehensible to the vast majority of Americans anyway), so how can we gauge how well or how poorly they do this?
Results, as evidenced by this example, would be completely unreliable. Like you said, how polls account for this terrible skewing of information, most don't care and don't want to know. Therefore, polls are irrelevant and pointless (as you've stated).

And this is just one of the more well known considerations. There are lots and lots of others. So yeah, polls are nice and the horse race is fun, but at the end of the day it's almost entirely useless for predicting outcomes.
I'd say that polls are more 'educated' guesses as to what will happen, and not real evidence as to the status of the majority's mind. I direct the viewer back to my 'Presidential election of 1936' point.

Achilles
10-12-2008, 09:55 PM
I'd say that polls are more 'educated' guesses as to what will happen, and not real evidence as to the status of the majority's mind. I direct the viewer back to my 'Presidential election of 1936' point.I mostly agree with this. I do believe that certain types of polls certainly have their uses. Polls that try to gauge the public's reaction to something I think have some merit. Polls that try to predict the future would seem to be almost useless just as soon as they are published.

Yar-El
10-14-2008, 01:55 AM
I'm going to place this out there. Polls are meaningless. What really matters is who will walk away the winner. Votes will be the determining factor. Who really can tell the actual demographic ratio? Media groups are hoping they have the actual numbers. This election is going to be a tight race. We will have a winner on November 4th; therefore, all the polls don't have any value anyway.

Achilles,

You have given me stuff to think about. I thank you for making this a challange. I will see you around the forums. Your insight into how polls work has inspired me to think.

Thank you kindly,
Yar

mimartin
10-14-2008, 02:24 AM
I'm going to place this out there. Polls are meaningless. Polls are not meaningless. Polls are meaningless for determining the winners, but they are far from meaningless. Polls measure the public’s feelings at that moment and can be used by both campaigns to decide where they want to spend money and what message is working with possible voters at that time. As Achilles said, it all depends on the method used by the pollsters. What really matters is who will walk away the winner. Is that really all that matters? To me it matter if I am voting for someone or if I am only voting against the other person. For the first time in two elections, I’m voting for someone and not voting against the other guy.
Votes will be the determining factor. Oh, how I hope you are right.
Media groups are hoping they have the actual numbers. No, they only hope they get the number as in ratings, as long as they get their numbers the polling numbers are meaningless to them. This election is going to be a tight race. So is your prediction better that the media outlets, is it more scientific? It is along time until November 4th and I believe the closeness of the race will be depended on the economic news between now and November 4th.

We will have a winner on November 4th; therefore, all the polls don't have any value anyway. The polls will not determine the outcome, but polls do have value at least to both campaigns and parties.

Having taken a few statistics classes in college, I am not quite a cynical about polls as most. I do however, question the methods used in taking the poll because that will determine the reliability of the poll. Again, polls only measure the public’s opinion at that moment in time and should not be used in making predications about the public’s opinion in the future. They do however have value.