Saturday, July 15, 2006

Another Bogus AP Poll

"Republicans are in jeopardy of losing their grip on Congress in November" reported the AP. Don't worry. It's just another slanted poll.

"The latest poll from the AP tells us that “Most Americans Plan to Vote for Democrats”. According to the poll, registered voters are tired of a Republican Congress and are willing to reverse control by a margin of 3 to 1.

The AP polled 789 registered voters which side they would vote for and of those questioned, Democrats were chosen 51 to 40 percent." - NewsBusters (via: RightOn)

Besides being an outrageous attempt to manipulate public opinion, these were not "likely voters" that were polled. Registered voters don't always vote. Especially during mid-term elections. So if the Democrats want to hang their chances on bogus polls, please - by all means, go ahead.
From EH

9 comments:

Jim said...

Faux Dynamics Opinion Polls are the truth handed down from God meant only to inform, but AP polls are bogus and meant only to maniputlate public opinion, not report it.

Hypocrisy, hypocrisy, hypocrisy.

The Game said...

you had it right in the beginning...
you are having trouble with those little things called facts again...
There are facts to show how this poll is completely invalid...
I have never seen data to show a Fox Poll is invalid...probably because they are fair and balanced polls...
you are not even trying anymore, are you Jim...you saying the word hypocrisy with nothing to back it up is so damn pathetic...

Jim said...

I have often pointed out how questions in the Faux Dynamics Opinion poll are written in such a way that their results are misleading if not meaningless. I've cited the actual questions for you. Questions like (and this is not exact but similar) Do you believe it was right that that the US attacked Iraq and Saddam is no longer in power. This is a loaded question that pushes the responder to indicate support for the attack even if he doesn't because he's glad Saddam is not in power.

It's a bad question and Faux news does it all the time. So don't tell me about how Faux polls are so righteous. Your NewsBusters "revalation" is crap because the "facts" presented are insignificant. Polls are often done using "registered" voters and "likely" voters and the defined sample is stated as one or the other. Most polling organizations have the expertise to know how to poll a relatively small sample with valid results. Thats what "margin of error" is all about.

And the only poll from 2004 that I recall giving Kerry a comfortable lead was the Exit Polls which is of ACTUAL VOTERS, not simply registered voters. So that NewsBusters "fact" is bogus.

The Game said...

You make no sense here...atleast at the end...
I don't notice fox loading poll questions...but it is clear that is done all the time...
Where you don't make sense is at the end...likely voters is different from registered voters...What newsmax says makes perfect sense...it is a perfect valid criticism that would be made in any stats or research class...
You have to look at how your population is affected by outside sources, by things happening in their environment..
Such as....what is the trend and prediction to how people vote in mid-term elections?
If voter turnout is lower...and it seems pollsters always know which type of people are less likely to vote....having people in a poll that are actually going to vote would make it more accurate...
And you bring up exit polls showing Kerry way ahead shows how worthless polls can be...since he lost by over 5 million votes...
Jim, I have a masters degree in this area...political nonsense will not work on this one...this was a half hearted answer on the issue...I usually don't have the energy to give full, page after page answers, and usualy I just don't care to explain it at all since it won't matter anyway

Jim said...

If you watch any of the news programs, you will see that polls are done many ways--registered, likely voters, registered independents. These polls are done on any of these samples. As long as the pollster makes clear what sample is being used, what is the big deal?

NewsBusters seems to be touting that they have found something secret and sinister in the poll. But they only know the sample parameters because AP makes it clear what they are.

Anyone can read the poll numbers and read the sample definition, and read the margin of error and decide for themselves how meaningful the results are. Characterizing this poll "an outrageous attempt to manipulate public opinion" seems like a gross exaggeration to me.

Jay Bullock said...

Okay, then, Game, how about a Fox/ OD poll from last week showing the generic congresional ballot as 42% D and 34% R? And that poll, too, was of registered voters . . .

You can see it here, along with the AP poll and polls by a dozen other organizations going back months, so you can see the trend. Going back to September, there is not a single poll on there (including several of likely voters) showing a Republican lead in the generic ballot question.

The reason why I think this poll--and all on this question--is worthless is not some kind of bias or the use of RV vs. LV, but rather that these polls don't take into account gerrymandering. Not enough districts are actually competitive for the overall mood of the country to make as much difference as it did in, say, 1994, when the Rs took over the House.

The Game said...

no matter how you word it, any poll is going to show republicans in trouble and that might be true...jay did bring up a great point...congressional races are more local races...senate races can be more national...these polls do not really show what is going to happen in November...
and I guess my problem with polls like the AP poll is not even the registered vs likely...its the population...skewed way to far to the Left..and don't try and tell me 51 percent of the country is democrat and 40 percent is republican...

Jim said...

Game, on what basis do you claim that this poll is "skewed way to far to the Left"? Because the result shows more support for Democrats?

jhbowden said...

The polls are correct.

Dems will still be in trouble though, because Republicans still beat them on security, which trumps all other issues for many since it is the precondition for them. Moreover, if you can't trust terrorist-loving Democrats with your security, why on earth would you trust their government bureaucrats with you health care?