Approval Poll Pt. 2: How do you feel, and how would you answer to, say, a Gallup poll?

[AlabamaLibrul]

Poll question: Approval Poll Pt. 2: How do you feel, and how would you answer to, say, a Gallup poll?

Revelations in the previous thread have led me to want some stats on this.

Apparently I went too long on the answers, so the ends of choices two and three should be "a 'real pollster' that."

[Jackpine Radical]

1. Do not approve, would not tell a pollster

because they always seem to assume that there is nothing to the left of Obama, so the only way to avoid having my opinion misinterpreted as dissatisfaction from the right is to go with Obama & his positions even when I'm further left. I not only would do that, but have done so in many instances. (For some reason we got a lot of polling before the 2010 midterms.)

[AlabamaLibrul]

2. Hmm....

Would you agree, then, that perhaps it is the polling that needs to be changed?

More of those "HCR: Too far / not far enough" type polls, and less "R/D/I: Approve/Disapprove" types of things?

[Jackpine Radical]

3. The polling would only need to be changed if

the point of it were accuracy in representing public opinion. I'm not at all convinced that is the goal of these questions.

[AlabamaLibrul]

4. That is certainly worth considering. I'm trying to get in idea of how everything lines up between

31% approval in my previous poll, vs. 39% approval among the entire nation, a much higher number among Democrats, but how many are answering truthfully?

Are people who use this website even being called in proper proportions? How do they choose people? Is the average person, hell, the average Democrat to the left or right of the average poster on here? More likely to approve or disapprove? More likely to say they do even if they don't?

These sorts of things.

[Jackpine Radical]

5. You are getting into the serious problems of sampling

that the pollsters face. In the old days, some people had phones and others didn't so they found ways to weight their phone results (e.g. using door-to-door subsamples), etc. But these days, society is changing so fast (adoption of unlisted cell phones, for example) that their adjustment research can't keep up. The only ultimate validation of poll results is in the elections. And for obvious reasons, that can lead to some pretty crazy situations. The polls might start getting weighted to take into account "black box voting" phenomena, etc. especially when exit polls are being thrown out in some places or adjusted to accommodate bogus ballot-stuffed results in others. For example, if a Kathy Nickolaus regularly throws in an extra 10% of R votes, the pollsters don't know real votes from Nickolaus votes & adjust their results to predict the addition of Nickolaus votes. Ane Wauk. Co. threw out the exit pollsters entirely in the Darling race this week, further sheltering their results from examination.

[Ozymanithrax]

7. You know, you are only sampling peole who see your poll and wnat to answer it...

Since you don't have a pie opions, which seems to be a thing we do here, it may not be valid.

[JDPriestly]

9. Exactly, Jackpine Radical.

[LoZoccolo]

6. This is like the birthers.

"Sure there's evidence, but it's fake!"

[AlabamaLibrul]

8. Is it?

To me, Democrats admitting (anonymously, on a Democratic website) that they would lie to a pollster about their approval of the President *does* call the "evidence" into question.

We'll have to let this run a while and see where the stats fall.