Who else here thinks the polls are inaccurate?

Does anyone think that maybe the sampling sz is too small or maybe only certain people are asked? I've been a registered Democrat a long time and have NEVER been polled!!:patriot:

It's all in how the questions are asked, the order of the words used, where the polls are taken, etc.

There are a lot of variables, and I think we're being told what they want us to hear right now.

Closer to elections, when more people are polled, it's more difficult to control the group being asked, so the results are more accurate. JMO...

poll that only gave Clinton and Obama as choices....makes you wonder.

:spank:

:)

:shrug:

:patriot:

Just wait until you hear all the excuses for why a candidate isn't doing well

It's because they only call landlines. I, for example, haven't had a landline in 8 years...I'll bet I won't ever get called for a poll.

and actually do attempt to adjust for that, as they do for any difficult-to-reach demographic.

Interesting article on Cell phone use from last year

http://people-press.org/reports/display.php3?ReportID=276

But truth be told that demographic skews towards subgroups of potential voters who are not in the likely voter category. They tend to be under 30, single and renting.

An article on it.

http://people-press.org/reports/display.php3?ReportID=276

I took one of those on line tests to see who my perfect candidate would be and guess who came out on top. Hillary! Just below her was Dennis Kucinich, I guess he was my second favorite. I support Kucinich and I find it hard to believe that a Hillary supporter would answer even remotely close to the way I would on most of the questions. It was a joke as far as I could see. Peace, Kim

making her the sixth-most-loyal-to-party Senator in either party during this Congress.

Secondly, on-line tests have nothing to do with telephone polls.

Thirdly, all polls are showing roughly the same thing.

Secondly, glad to meet an expert.

If there were extreme outliers, I would question those polls. But the results seem to be consistent from poll-to-poll, so I think they're fine.

can possibly be accurate. Also, even with widespread sampling you'll get inaccurate results, i.e., if
you sample one person for each of 2500 counties across the US, the sample will come up far more conservative than the general populace, because it will be heavily weighted toward rural populations. The one vote from a county with 832 residents in western nebraska will equal a single vote from downtown manhattan with its 5 million residents.

Supposedly, the pollsters balance the polling but their methodologies are their own and they know full well that polling not only reflects the public but also directs the undecided, particularly in the past few cycles -- we KNOW the MSM is biased, and think how many polls are conducted by MSM entities. How can we possibly trust them?

they sample one person per county? 2500 randomly-chosen people is more than enough to ensure an accurate poll.

Do they have a computer cranking out numbers with a random number generator? Then why doesn't it ring up cell phones as well as land lines? "In a random poll of likely voters..." If they are already identified as "likely voters" how is THAT random?

We only know that it is random because they SAY it is random. Maybe 30 years ago, when the population was far less transient, and there were no cell phones, and the media conglomerate did not own the polls, there were accurate polls. These days, polls are designed to shape opinions, not to reflect them.

from demographic studies, as well as questioning in the beginning of the poll. They fully disclose their methodologies, including the exact questions being asked, in their full polling reports.

If you believe polls do not reflect opinion, why are all polls showing the same thing, regardless of company? And why did the polls do such a great job of predicting the near-tie in 2000, Bush's narrow popular-vote victory in 2004, and the Democratic sweep in 2006?

so I guess even Hillarys opponents want her to win

:crazy:

If you combine the polls that provide fairly consuistent results, more than 2500 people are polled every week.

I think the polls reflect the current state of the electorate:

Hate Bush.
Hate Congress.

Clinton and Giuliani in the lead.

their candidate the edge. But since that candidate has dropped so low in the polls, they now don't like it.

And the sample size is never large enough.

A margin of error of 4 to 5% is fine.

the stats guys said they were pushed to use the smallest sample size possible and the math made some questionable assumption. They used 5,000 to represent the nation. Today the population is even greater and the math and questionable assumptions haven't changed.

to save money and work faster. They've realized that they don't need a 2% MoE for primary polls; 5% works fine.

Unfortunately, way too many Americans are easily manipulated.

:evilgrin:



IMPEACH INVESTIGATE INDICT IMPRECATE INCARCERATE :banghead:

eom

Only when I dislike the results...

Man I am so snarky lately.

As long as the polling follows the statistical rules & guidelines, they will be accurate.

If many polls all disagreed with each other, then that would be a red flag. If many different polls taken by many different pollsters line up, you can have great confidence in the results.

doubt the polls.

But they are almost always accurate within their margin of error, regardless of what you think of the results.

I question the group that is being asked. I don't have any problem with the sample sizes - they seem to be statistically acceptable. But since our nominating process consists of caucuses and primaries in the early states, I do not think that these polls can accurately predict who will win Iowa, New Hampshire, etc.

Most of the national polls that I have seen are asking registered democrats - which all sounds well and good until you stop and think - how many registered democrats will actually caucus or vote in their state's primary? There are many registered democrats who ONLY vote in one election every 4 years and that is the national presidential election. Even in the early states, some will not bother or be able to leave their jobs to vote.

Now, in Iowa, I have seen polls that consist of "likely" caucus goers. A lot of people will say they are likely to vote in the caucus - but will they? On a frigid cold day or night?

The poll I would like to see is one that asks PREVIOUS caucus goers who they intend to caucus for - or one that asks PREVIOUS primary voters who they will be voting for. Of course this leaves out first time voters, who could be statistically significant - so no poll is perfect. But I have not seen a poll such as this. Do they exist? I don't know. Candidates right now are not relying on the polls as a predictor of what WILL happen - they are using them to build up their candidacy and to push the idea that there are a certain number of "frontrunners".

These polls are not really useful in predicting a nominee. Once the nominees are set and of course there are only two (or maybe three) to choose from, then the polls are extremely accurate. Historically, the polling has been very poor at predicting who will be the nominee. So all candidates supporters take heart - most likely your guy (or gal) does have a chance - maybe a small one - but there is a chance. Remember, not one vote has been cast yet. DO NOT BELIEVE THE POLLS!

Or just the ones you disagree with?

First, polls are generally accurate if they are done correctly. That doesn't mean perfect nor does it mean that a poll today will accurately reflect an election taking place in Jan. of 08.

Second, sample size is an easily determined number when the sample is picked correctly and can be rather small. As long as one is willing to have a margin of error of say 5% instead of 1%, then one can have a small sample size.

Third, pollsters no longer use random dialers but lists of voters, often previous voters, and choose the correct proportions of various subgroups. If 18% of the electorate is AA then 18% of the sample is etc. This is called a stratified random sample.

Fourth, online polls are voodoo.

You are currently in:

Stage 1: Denial.

Patient is in critical condition; patient has made no advances in the Grief Process: there remain just over three months to Super-Duper-Hyper-Spectatcular Tuesday (February 5th), at or around which point patient's condition will reach fatality status, and stays will be enjoyed. Development into Hillhating Nimrodhood is strongly recommended, with an eye towards swift metamorphosis into Bargaining-Stage Bidentia. If necessary, shock therapy should be prescribed: to begin with...

She's called "Madame Inevitable" for a reason, ya know? Notice how this is the second week running that the Hillmeister is cracking the 50%-plus range in the major polls? Obama??? Edwards???-- Give it up, no-hope losers!

"She's with Me, She's the Nominee - Get Used To It!"

She's a juggernaut baby!

Booyah!

You'll be a Biden Bargainer before you know it!