RoadBikeReview.com's Forum Archives - Non-Cycling Discussions


Archive Home >> Non-Cycling Discussions(1 2 3 4 )


Why I hate polls....(24 posts)

Why I hate polls....ClydeTri
Aug 15, 2003 12:39 PM
Saw this on the front page of CNN.com right now:
"When you first learned of the blackout, did you fear it was a terror attack?"

They give you a Yes or No to answer. What is wrong with this question? Think about it.

I am talking about the use of the word "fear". Fear is an emotion. Should the sentence have read:
"When you first learned of the blackout, did you think it was a terror attack?"
There is a difference between the two sentences. One can both think and fear, but one can think but not fear. Did the author of that sentence mean the word "think"? When I read it my initial response was no, I did not fear it, but I thought it. Where do they get these people? Is this what journalism schools are producing? I am not a journalist, or a writer, or author, but, I do understand words have distinct meanings and one can pick and choose them to make whatever point one wishes.
My point? Polls are meaningless in most cases. They give you one or a couple words choices for answers that might take hundreds of words. Off my soapbox.
There is a difference between the sentences,TJeanloz
Aug 15, 2003 12:46 PM
There is, as you point out, a difference between the sentances.

Did I think it was a terrorist attack? No.

Did I fear that it was? Yes, because it would have much greater ramifications on the national psyche. As it stands, it was an inconvenience, if it were a terrorist act, we'd have a problem.

My gut feeling though, is that the terrorists have won. Everytime something goes a little unexpectedly, everybody jumps up and says: "IT WASN'T TERRORISTS". Apparently we are living in fear that it would be, which seems to be the goal of the terrorists in the first place. In 1965, nobody probably even thought of the possibility of terrorism.
Then your answer should be "No".czardonic
Aug 15, 2003 12:49 PM
They picked the workd "fear" because they wanted to know who was afraid that this was a terrorist attack. Is that not a valid question for them to ask?
Then your answer should be "No".ClydeTri
Aug 15, 2003 1:04 PM
My "gut" feeling is that they wanted to know how many people "thought" that it was a terror attack. The use of a word such as "fear" puts the concept of emotions into the polling question without most people even realizing it. If I was asking the question clearly wanting to know if people "feared" it, I would have asked: "when you learned of the blackout, did you experience the emotion of fear in relation to this blackout?" Or something similiar, then there would be a definitive question on emotions and the blackout. I would bet a sizeable percentage of the public who have participated in the CNN poll did not consider that fear is an emotion, but rather subconsciously substituted the word "think" in their mind when casting their vote.

No, Czardonic, I am not making some big liberal conspiracy out of this. Just making the point that polls can be used to get any answer you want and are usually very poorly devised.
What if your gut isn't right?TJeanloz
Aug 15, 2003 1:13 PM
I think the "fear" question is perfectly reasonable. Actually, more reasonable, because we know what our own emotions are - I know I FEAR it was terrorism. But if you ask me what I think caused it, I have no idea, I don't work for the power company, and I don't know how these things work.

I think in the context of polling the general population, it makes more sense to ask if they were worried that it was a terrorist act than what they thought the cause was.
I agree with TJeanloz.czardonic
Aug 15, 2003 1:26 PM
What the public "thought" it was is not news. Who would know? Emotion is the only thing the avereage respondent could speak to with any authority, and it is something that has real, non-emotional reprecussions.

Anyway, fear seems like a perfectly rational response to the hypothetical possibility of terrorists unleashing a veil of darkness over 10s of millions of people. I doubt that many people unwittingly mischaracterized their response by answering yes to "fear" of terrorism when they actually had a mere dispassionate recognition of the possibility.
They used the word "fear" because they wanted hits..nmMR_GRUMPY
Aug 15, 2003 1:05 PM
i fear you are making too big a deal of thisrufus
Aug 15, 2003 2:46 PM
the word fear in the english language has often been used as a synonym for "think" or "suspect". i don't see the big deal about it.
Choice of a word can make 80% in the results.dr hoo
Aug 16, 2003 4:58 AM
Not that this would look like a case with that magnitude of variance.

Self selecting web polls are worth very little in general. My guess is that IF they thought about it, they probably went with fear to bias the response to that category.

Also, those who are fearful, and emotionally charged, are going to take the time to answer the poll in higher numbers. This biases the case selection towards one of the choices.

I have no confidence in this kind of poll. However, surveys can be very useful devices, if the survey design is stringent and pre-tested, and analyzed appropriately.
The real issue is:Spoke Wrench
Aug 16, 2003 5:48 AM
How are they going to use the results?

As dr hoo suggested, if you let me word the question, I can pretty much assure you of whatever answer you want to get. Polls regarding abortion have proven that fact with both sides claiming hugh majorities in opinion polling results.

Then the people from the other side of the aisle cite them when they write their stupid little letters to the editor.
Nobody takes bogus internet 'polls' like thisOldEdScott
Aug 18, 2003 6:05 AM
seriously. It's what we call J.O. polling. Makes someone feel good, but nothing's really accomplished.
I dislike them for other reasons, tooDougSloan
Aug 18, 2003 8:16 AM
I agree that the wording can vastly affect the results.

However, I dislike them, particularly with elections, because they are used to *shape* the elections, not just report them. If someone is shown to be leading in an early poll, that person likely will obtain more support and contributions. People like to side with winners. If not for the polls, we'd not know who was "winning" until the primary votes or general election.

This is done on many other issues, as well. People carefully craft polls to prove a point, rather than objectively reporting opinions. You don't prove facts through polls; you don't prove guilt of defendants through polls; you don't determine foreign policy through polls. Yet, many people seem to think so, even if subconsciously.

Doug
Ummm, well, sorta ...OldEdScott
Aug 18, 2003 8:43 AM
In my own work, there are two kinds of polls: Polls where we are VERY intent on neutral wording, so we can really find out what folks think about something. This is (usually)for internal campaign use. Why run a poll if we don't want good, clean answers that tell us true stuff?

There's also what we call 'push polls,' where the wording is loaded (in part) to give us a result we want. Those aren't used much anymore, because everyone (including the press) is hip to them and won't report them. I know the reactionary crowd on this board believes the press swallows every liberal or Dumocrat push poll hook, line and sinker, and reports the treasonously slanted results as gospel, but the fact of the matter is it almost never happens. By that I mean I don't even remember the last time it happened. Sigh.

Go ahead and believe that they do, though, if it makes you (generic Fox conservative you) feel better. Wouldn't want to rock the foundations of your comfortable and comforting belief system with facts.
no coffee this morning?DougSloan
Aug 18, 2003 8:54 AM
Sounding a little abrasive this morning.

I didn't accuse anyone in particular of "push polls." You really want to dislike all us on the other side of the aisle, huh?

I really am not that concerned about push polls. Those are so transparent that I'd imagine most people are on to them. My concern is about legitimate, objective polls that are misused in a way. They are used by reporting their results, and then people thinking that the mere fact that the polls show that a large group of people think a certain way, then maybe I should, too? In other words, the results (not the questions) influence people. Or, the poll is used in a non-sensical way. Why poll people on whether OJ is guilty? It's stupid. People don't know the detailed facts of the case nor what the applicable law is (in advance of the trial, is my concern). Or, having polls determine foreign policy -- we elect people to make these decisions; this is not a pure democracy, subject to the whims of the day, as they might be influenced by the lastest photos from the middle east or some other poll. We don't have mob rule.

That's what I meant.

Doug
Just got some bad weekend poll numbers. nmOldEdScott
Aug 18, 2003 8:57 AM
a Fox News poll? ;-) nmDougSloan
Aug 18, 2003 9:00 AM
political humor to brighten your dayDougSloan
Aug 18, 2003 9:07 AM
A Good Laugh:

Jay Leno: "And welcome to Day 119 of the Bush vacation. Isn't it 119?"

Jay Leno: "You know, one thing about President Bush, I don't understand, and I like the guy. I like the man. But I see him on TV. He's clearing brush all day. He's chopping down -- he runs in 100 degree heat. He has a club for guys that run in 100 degree heat. Doesn't bother him. Takes a bite out of a pretzel, boom, he's down like that. Falls off a scooter. I don't understand."

Jay Leno: "Boy, it was hot today. ... 102. So hot, Satan today entered the governor's race."

Jay Leno: "And an NBC News poll has found that if the election were held today, 31% of California voters would vote for Arnold Schwarzenegger and 26% were not sure. Today, Gray Davis announced he is changing his name to 'not sure.'"

Jay Leno: "Yesterday, Governor Davis stopped at a gas station over here in Brentwood. But bad news, they weren't hiring."

Jay Leno: "Because of the equal time rule, all of the candidates get equal time. If they're on TV, the others have to get the same amount of time. Arnold Schwarzenegger movies will be kept off the air for the next few months because you'd have to give the same amount of time to all these [other candidates]. You know, forget Arnold. Let's see if we can get Carrot Top to run and an end to those stupid 1-800-Collect commercials."

Jay Leno: "Arnold Schwarzenegger vowed today to get the budget under control. See, how bad are things in Sacramento when you need someone from Hollywood to show you how not to go over budget?"

Jay Leno: "Hey, have you heard about this porn star named Mary Carey who's running for president? Running for president. Running for governor. Porn star and president just seems to come together. I don't know why. I don't know why I make that connection."

Jay Leno: "She's running on a platform of taxing breast implants. Which, of course, is California's largest natural resource. Did you know that?"

Jay Leno: "And election officials here in California are concerned that having 247 candidates will require a ballot so long, it will difficult to count. Today in Florida, they said, 'What? You count the ballots?'"

Jay Leno: "And as if this whole thing isn't confusing enough, I don't understand this. And I'm trying to follow it. Election officials announced this week - I guess this makes it more fair. They said the alphabet on the ballot will begin with the letter 'R,' then 'W,' then 'Q.' You know, even Sesame Street is laughing at California."

Jay Leno: "And good news in Liberia. The reign of terror has finally ended. Liberian President Charles Taylor has stepped down. He stepped down this week. And President Bush, a little sorry to see Charles Taylor go because this is one African leader whose name Bush can actually pronounce."

Jay Leno: "Today President Bush said the economy is getting better. He says sales of surface-to-air missiles is up 10%."

Jay Leno: "Yesterday in New Jersey, the FBI arrested three men suspected of trying to smuggle a surface-to-air missile into the US. And did you hear their excuse? I can't believe this. They claimed they were going to Newark and didn't want to be outgunned."

Jay Leno: "You know, it's amazing. We spend all this time looking for weapons of mass destruction. Turns out, they're in New Jersey!"

Conan O'Brien: "Arnold Schwarzenegger has hired billionaire Warren Buffett as his senior economic adviser. His latest move, yeah. Yeah. Yeah, and not to be outdone, Gary Coleman has announced that his senior economic adviser will be Thurston Howell III."

Conan O'Brien: "Also in California, one of the candidates running for governor -- 'cause I think there's a lot of them -- one of the candidates running for governor is a 100-year-old woman. That's the latest. Yeah, good for her. Yeah, the 100-year-old says she'd like to recall Governor Gray Davis, but more importantly she'd like to recall where she left her teeth."

Conan O'Brien: "A company has come out with a George Bush action figure that commemorates his landing on the aircraft carrier earlier this year. Yeah. Yeah, the company says by posing Bush and making him talk, kids can get an idea what it's like to be Dick Cheney."
The last one's my fave, obviously! nmOldEdScott
Aug 18, 2003 9:12 AM
Isn't there another use for "push" pollsTJeanloz
Aug 18, 2003 8:56 AM
My understanding, which may well be flawed, is that "push" polls are used, in and of themselves, as advertising. They "randomly" call people to poll (I say "random", because my home in New Hampshire is bombarded with poll telephone calls, while my house in Massachusetts gets almost none - so it's clearly not random), and ask them leading questions, that they don't really care about the answer to - the question is the advertisement:

"Do you think Bill Clinton should have orgies in the Oval Office" etc.

Or: "Do you think George Bush, who has vetoed every Medicare bill he's seen, cares about New Hampshire seniors?"
Oh, those aren't polls at all, obviously.OldEdScott
Aug 18, 2003 9:10 AM
It's a bunch of yahoos sitting at a phone bank. No data's really collected, as you say. It's a huge inside joke.

I'm talking about pseudo-legitimate polls that are very cleverly designed to ALMOST be legit. Those are the ones that are (or used to be) useful in campaigns, especially for underdogs. "New poll shows Yail Bloor rapidly gaining ground on 70-term Repub incumbent." You could create a perception of momentum. Gary Hart used that technique to a fare-the-well in the 'where's the beef' primary.

Nowadays, you release poll results and the media says: We gotta see the methodology, gotta see the questions, gotta see this this and this ... It's really hard to slip a skewed poll past them.

We've lost our co-conspirators in the Vast Left Wing Conspiracy.
Here's an exampleOldEdScott
Aug 18, 2003 9:41 AM
An example: In this year's governor's race, the Repub candidate is beating us by something like 8 points, according to independent polls. We just ran a very, very straight poll that shows we've closed the gap to 1 point, well within the margin of error.

Here's the kicker, buried deep within the methodology: We weighted the sample according to historical turnout over the past six elections. That's all the methodology says.

But the fact is, Something like 40 percent of Dems in this state historically vote in gov's races, to only 30 percent Repubs. So the sample in our poll is weighted toward Dems, exactly reflecting those numbers. Is that legit? Arguably, yes. Also arguably, no, since this is the first year in almost 30 years that the Repubs have a strong candidate with a real possibility to win, so they're pretty jacked up and more than usually likely to vote.

Fair and balanced? You decide.

Anyway, the political reporter for the state's largest paper sniffed out the consequence of the weighting, and is so far refusing to accept this as a legitimate poll, although he admits we have a point. We say, what do we have to go on other than historical figures? Should we just add a bunch of Republicans because they seem excited for now? That's our spin anyway.

Bottom line: The questions in this poll are straight arrows, legit as hell. No loaded language or terminology. But we're still having problems with it. You really have to go to great lengths to get a biased poll accepted these days.
that's the improper purpose I was talking aboutDougSloan
Aug 18, 2003 9:55 AM
If you want to do a poll to find out information useful to your organization, that's fine. However, it seems you also want to use it as publicity -- to say to the state, "Hey, look! We are popular! Other people are with us, so should you. It's going to be close, so get out the vote."

Polls as tools for assessing where you are are fine. I object to using them as tools to persuade. I doubt any member of the public would understand the nuance of your poll methodology, even *if* it were explained.

Nonetheless, Free Speech and all that, you can say what you want, if you can get it reported.

Doug
that's the improper purpose I was talking aboutJon Billheimer
Aug 18, 2003 10:34 AM
I've been following this discussion, and as an uninformed voter I think the point Doug makes is valid. Polls are used more often for propaganda purposes than pure information gathering, at least in political campaigns. But also with respect to public affairs issues where there are usually opposing political agendas.
Don't necessarily disagree, althoughOldEdScott
Aug 19, 2003 5:23 AM
politics IS in many ways propaganda, with two sides making their case using the tools at hand, and the system works to the extent that voters know the game, sort it out, and make a judgment. It's like a free market. There's an invisible hand at work.

Still, the point I was trying to make above is that misleading political polls are far less common than you might think these days. My example -- a very good poll, relentlessly unbiased -- goes unreported because a perfectly defensible assumption in the sample gives just a WHIFF of skewing the results.

The liberal media does a pretty good job of policing these things nowadays.