How much weight...

[Tarfang_Trubasher]

Do y'all put in polls? Especially the "one week" polls conducted by Zogby and other such companies?

I was on the Barack Obama Test web site. A few questions, I thought anyway, were a bit too...biased (?) but, it was the "results" page I found the most interesting. There was a quoted source per each question that gave the full results, based on the quoted source.

Here is an excerpt:


So, my question is -- how serious do most folks take a 1,000 person poll? Do these help to solidify anyone's vote?

-TF

[Lurker]

A few questions, I thought anyway, were a bit too...biased

Only a few? That's one of the most extreme examples of winger propaganda I've seen in awhile. Most of those questions are loaded with disinformation.

Can you spot the lie in question # 2?

2. Proponents of drilling in the Arctic National Wildlife Refuge in Alaska point out that drilling could cover an area that covers 2,000 of the total 19 million acres that make up ANWR. Should the U.S. begin drilling in ANWR even if we won't realize the potential benefit for several more years?

If you can't pinpoint the lie you might be relying on the wrong news sources.

[Tarfang_Trubasher]

*sigh*

I give up.

You know, I can find a bunch of "winger" bullshit aimed at the other side too. I just want to know what people think of these polls, and how much weight they have. I didn't ask what you thought of the questions. Yes, fine, the questions are fucking one sided. This guy is selling a book. Great...the polls still exist in some form or fashion and can be used in dozens of other sources. It's not like these polls were created specifically for this site.

Zogby is supposed to be an independent polling group. I'm researching now to know if these questions were altered from their original state. If they were, fine, this site sucks. It was clearly biased to begin with, and I didn't question that. My problem is if these pollers are leaning with a certain bias, then I might as well add Zogby to the "ignore list" of my media resources.

As an additional side note, since I'm constantly accused of having poor media resources..some I've never even heard or seen, where do you "smart people" go that gives you such high enlightenment that allows your information to so greatly trump mine? I'd sure like to find this Mecca of uber info so I can stop being one of the drooling masses.

-TF

[Lurker]

Could you ask your question again without linking to a winger propaganda site?

[Tarfang_Trubasher]

Sure.

Do y'all put in polls? Especially the "one week" polls conducted by Zogby and other such companies?

I was on a web site. A few questions, I thought anyway, were a bit too...biased (?) but, it was the "results" page I found the most interesting. There was a quoted source per each question that gave the full results, based on the quoted source.

Here is an excerpt:


How serious do most folks take a 1,000 person poll? Do these help to solidify anyone's vote?

-TF

Is that better?

[Lurker]

I can assure you that none of those questions were from a Zogby poll. I'll see if I can find some of the actual questions.

[Lurker]

Well, color me shocked. Those are the actual poll questions that Zogby asked.

So, to answer your initial question...

Polls are supposed to guage public opinion, not solidify it.

If a poll question contains disinformation and bias to the degree some of those questions do then the poll is useless for determining public opinion. The poll becomes a push poll; an attempt to sway the uninformed.

[Embar Angylwrath]

Well Lurker, seems to me you think Zogby is pushing an agenda. Are you making that claim?

[Tarfang_Trubasher]

Could you link me the proof, Lurker? I have managed to fail in finding the questions.

Though, I've managed to find some other stuff -- Zogby is very much "in question" when it comes to polling. Apparently, they love to use "sample sets" from their databases. Many sites have questioned their polling methods, and several pointed to this web site (no, I don't think this is a "winger" jargon site ) .

I also get the impression you can "buy a poll" from them. Basically, you pay per question...then they do all the polling, and cough up your results. Then, in typical media fashion, the misinformation spreads. A growing issue I'm seeing is, people will gladly strike out at misinformation, without providing a credible link to the "correct" information. The Internet and global media is big...I'm sorry I don't have my own Ida to go pounding through it all at once.

Rasmussen seems to have a decent reputation for polling -- maybe I'll check out some of their stuff.

-TF

[Harlowe]

I don't give any weight to push-polling. In fact, I pretty much ignore political polling other than to see how candidates are doing in general and that is just out of curiousity.

[Lurker]

Could you link me the proof, Lurker? I have managed to fail in finding the questions.

I can't find the complete poll but there are sections of it reported on by enough valid news agencies that I'm sure those are the actual questions. I think you are right that ATI contracted Zogby to run the poll. I think they formulated the questions as well.

Well Lurker, seems to me you think Zogby is pushing an agenda. Are you making that claim?

The questions contain significant bias and disinformation. Draw whatever conclusion you want from that.

[Lurker]

Looks like ATI News is just a conservative front organization. Not that I'm accusing the author of Top 200 Reasons Not to Vote for Bill Clinton of having an agenda or anything.

[Klast Brell]

Believe it or not I used to work in political polling way way back in the early 90's Did a lot of market research as well. Do you prefer crunchy or creamy peanut butter. that sort of thing.

Anyway. The firm I worked for Conducted polls for the local paper. The Minneapolis Star Tribune. You would not believe the extent to which we went to make sure we were getting a representative, yet random sample. The typical respondent base was a little over 800. We had randomly generated phone numbers for the whole state. These numbers were divided by county. Each county had a quota for how many completes (jargon for properly executed survey) needed to be done. This number differed by county and was proportional to the county population. If a county with 100,000 people required 20 completes a county with 30,000 people required 6. For each call we would ask for the person in the household 18 or over with the most recent birthday. If that person was not home we would try to obtain the best time to call back to reach that person. This was to eliminate any self selection bias. So the results were not weighted towards those who were most likely to answer the family phone, or towards those who most enjoyed doing surveys.

The random phone numbers generated per county were also proportional to that counties population. In the calls disconnected numbers were disposed of. Likewise any number that reached a business, resulted in a completed survey, or where the resident refused to participate was discarded. All the remaining No Answer, Answering Machine and Busy Signal numbers were kept and called again and again until we either reached our quota of completes for that county, or just plain ran out of numbers eligible to be tried again. Only if all the numbers were exhausted would a new set of random numbers be generated for that county. And then they would only generate a set proportionate to the number of completes left to do in that county. This was done to prevent weighting towards those who were most often home. If we called an endless stream of numbers and dialed them only once we would disproportionately represent homebodies and under represent people who liked to go out.

Calls were made 10 AM to 10 PM 7 days a week. making sure that the number of calls during the day, evening and weekends was evenly distributed as well. Again to remove weighting against people who worked nights etc.

and that's just the tip of the iceberg. We were required to use fake names. No names that could be associated with a nationality or religion etc. No Yitzhak Israel, No Juan Mendoza, No laquifa Jackson. All names had to be generic sounding. we turned in to a room full of Dave Johnsons and Jenny Smiths. I used Peter Anderson for my phone name. Again to avoiding biasing the sample by encouraging or discouraging people who had an affinity or prejudice toward any group fro participating.

It went on and on. Questions were extensively studied to make sure their phrasing and the possible answers given were not biased. Once a question was decided on it would be used and reused poll after poll for decades. this allowed for comparisons over time. Ever heard the phrase "on the wrong track" used in a news piece about a political survey. The specific question "Would you say that the country is generally headed in the right direction or on the wrong track" Has been used for decades by most political poling firms. unchanging text like this allows tracking of public opinion of presidents congress and the economy over the years knowing that the results can be apples to apples compared.

Do I trust polls? Ones sponsored by a major news organization who make the text of their questions and the answer choices are available for review I trust.
Others. Unless they are associated with a major university I don't trust them as far as I can comfortably spit out a rat.

[Lurker]

I used Peter Anderson for my phone name.

Hmmm... I would have figured you as a "Rod Johnson".

[Ddrak]

So, my question is -- how serious do most folks take a 1,000 person poll? Do these help to solidify anyone's vote?

If conducted properly, a 1000 person poll should be within 3% of the true value 95% of the time.

No one should vote based on polls. That's putting the cart before the horse.

Dd

[Kulaf]

I'd like to point out that the question in the sample given in the OP is rather ambiguous. Is it search and seizure of his person and personal effects? In that situation I would say no......even the police do not need a warrent to search your person if they have probable cause to suspect you were involved in a crime. Certainly Homeland Security should seek a warrent to search anyones home citizen or not.

[Tarfang_Trubasher]

The question I linked was arbitrary -- though, I think the key statement there is "non-citizen" to my eyes.

-TF

[Klast Brell]

Kulaf points out one of the challenges to polling. Once a question is designed you can do nothing to influence the responses.
That especially means that you can not explain a question or define it's parameters. About the most you could go to clarify somethig like that would be to say. "Overall" or "In general"

And honestly anyone can pick any question apart. Consider the approval rating for the president. Everyone has somethings they approve of him for and some things they do not. But this is not an essay test. And neither is your ballot on election day. You will have to choose a candidate and live with your choice. Likewise your candidate will be signing a law or vetoing it. There may be negotiations along the way. But that is where the rubber meets the road. Executive, Legislative, Judicial. The reasons for their choices may be richly nuanced. They may have reservations. But they have to reach a blank and white decision. Yea or Nay. Sign or Veto. Plaintiff or defendant.