General Question

Bri_L's avatar

Can Political polls misdirect public opinion?

Asked by Bri_L (12191points) August 27th, 2008

I don’t really know how they work. Is it possible that they present an inaccurate picture because of who is willing to take time and answer them.

For instance if only a certain level of economically sound and politically active people answered the questions you would achieve a certain result.

Those results would then be broadcast. Then other demographics would see them and possibly be swayed by such polls.

Thank you in advance.

Observing members: 0 Composing members: 0

11 Answers

blastfamy's avatar

No doubt, absolutely!

RandomMrdan's avatar

Thats exactly what I was trying to get at with my question!! You stole my thoughts that I couldn’t project onto Fluther.

http://www.fluther.com/disc/21518/why-do-polls-show-that-the-political-race-between-mccain-and/

And thats what I mean when I don’t understand where they get these statistics from.

RandomMrdan's avatar

reminds me of my psychology course I took, and how the opinions of other can effect your own….example:

I don’t recall the name of the study or anything, I just remember the circumstances.

But, there is a row of individuals, and there are a select few who are in that row that are test subjects. They are presented a graph of numbers and data, and then would ask the group if they would agree with something they have drawn from this data. They would start at one end of the row, and see whether they would agree or not. Even if they shouldn’t agree, anyone apart of the test would agree, to see how the actual test subjects would react. And surprisingly a lot of the subjects would agree even though it was clear the conclusion drawn from the data was inaccurate. Maybe so they could fit in, I don’t remember exactly the reasoning.

Bri_L's avatar

That was what I was thinking back to as well, that and my statistics course.

wundayatta's avatar

Innumeracy is a big problem in the U.S. People have no idea how to assess the credibility of a poll. Even people who have taken stats may not have taken sampling.

The only way to know how to take the results of a poll is to understand how the poll was taken. The gold standard is a random sample of the appropriate population. If you have 1000 or more respondents, you’re in good shape, usually.

But, Bri L, you are right. Pollsters can, and do, mess with sampling technique. They will select a population that is biased towards them, say Southerners, and sample just that population. They represent the poll as a random sample. And it is. Except it’s just a random sample of Southerners. They are a good deal more sophisticated than this, so it is often difficult to understand the population they sampled, even if you can get them to tell you.

They also attempt to directly influence results through polling techniques, where you ask an apparantly unbiased set of questions, and then ask, did you know candidate A supports this? If you knew that, would that change your willingness to support Candidate A? People say yes, and then they go into the supporter column. The candidate releases their “internal” poll, and uses it to try to influence others.

Another thing that causes problems is the way people are forced to answer questions. Many people have not decided this far from the election. So pollsters often ask “If the election were held today, who would you vote for?” These people’s choices are very fluid, and yet the pollster is acting as if they are a solid supporter. Many people haven’t even thought about it yet. So there’s lots of wiggle room in responses.

Some pollsters have a vested interest in being correct. These tend to be the independent pollsters. They can’t sell their results unless they are seen as accurate and unbiased. Even known biased pollers (Celinda Lake for the dems, eg) can be accurate, because they don’t want to be seen as being sloppy methodologically.

So, lots to watch out for; lots to beware of; and yet, it all does even out. There are sites on the internet that report the results of all the polls, and show an average. Sometimes that gives you a better idea of what’s going on. Also, polls get more and more accurate as we get nearer the election. Right now, there’s a lot of bouncing around. So take everything now with a grain of salt.

Bri_L's avatar

@ daloon – thanks so much for that info. very much what I am looking for and suspect.

JackAdams's avatar

It’s done all the time.

August 27, 2008, 3:17 PM EDT

tinyfaery's avatar

I’m not sure if it’s the poll itself or the media coverage of it. Couple a skewed statistic with a derogatory story about candidate x and you’ve just convinced some people that candidate _ x_ is not the correct choice.

vectorul's avatar

How come the “error” in polling statistics no longer appers?

JackAdams's avatar

I think the “variance” is, where they state that it is accurate “to plus or minus 3%.”

August 27, 2008, 8:03 PM EDT

winblowzxp's avatar

yes, it’s within x points (usually percentage) usually.

Answer this question

Login

or

Join

to answer.

This question is in the General Section. Responses must be helpful and on-topic.

Your answer will be saved while you login or join.

Have a question? Ask Fluther!

What do you know more about?
or
Knowledge Networking @ Fluther