Gallup Tracking Poll Thread [Obama vs McCain]

[minionofmidas]

Trying to figure out the day by day totals of the tracking polls is nothing but guesswork.

You are basing assumptions off of assumptions. You are assuming that you know what the day rolling off is, in order to determine what the new day coming on is. After days upon days of guesswork, the margin of error becomes very large, thus putting those numbers in high doubt.

That doesn't make any mathematical sense at all.  It's a three-day rolling average.  It's the margins of the days divided by three.  You can stagger it (I'm sure, because I've done it before, I just don't remember how) and determine the approximate daily average within a small range.

Margin of error?  what?  Margin of error is a sampling issue; it has nothing to do with calculating poll averages like this.  And there may be errors, but they will average out -- otherwise the rolling average would be significantly off.  Since it's based on rolling average, that isn't possible.

It's not exact (there are rounding issues) but it isn't "guesswork."

Breaking it down here:
The average for the last three days is of course no guesswork at all.
Individual days include a portion of guesswork o/c, although the longer the time frame you use, the better it should get.
The difference between one day and the day three days previously is no guesswork at all.
Very large bounces are of course going to be readily apparent with no major guesswork involved.

ie, if you assume that the 23rd and 24th were actually both McCain +1, that gives
25th McCain +4, 26th Obama +8, 27th Obama +14, 28th Obama +2, 29th Obama +8, etc. Stretch it across a long timeline, compare with events and assumptions about a trend for Obama to poll better on weekends, and you should be able to come up with a highly reasonable assumption at what's most likely.

[minionofmidas]

Wow. Is this election over? (No, I'm seriously asking. Isn't the recent pattern basically one of no movement between a few days from now and two weeks before the election, when it will once again look like everything's in flux - but the results then show the answer to that as "not really"?)

[minionofmidas]

Wow. Is this election over?

We haven't seen the "Whitey" tape yet...

[minionofmidas]

I furthermore concede this election to the Democrats.

NOOO! Don't jinx it!

[minionofmidas]

In other words, J.J. got the year right, he just got the wrong name.

[minionofmidas]

In other words, J.J. got the year right, he just got the wrong name.

I think the last candidate to lead by this much and lose was Gore, as the chart shows. 

The chart is using the media color scheme, not the sane color scheme. Look again. (Besides, Gore didn't lose the election anyhow. )

[minionofmidas]

...And it is all 6's.

Battlground is usally decent.

Not when they're weighting and non-weighting like that they aren't. So maybe they were once - but that's not stupidity anymore (which all the pointed-out minor strangenesses in the DKos thingy seem to come down to), that's quite obviously an agenda.

[minionofmidas]

Sarah Palin - the last, best hope of earth?

[minionofmidas]

But the weekend/weekday bounces are just a figment of your imagination.

Except other people noticed them on Gallup.  He's one from an Obama supporter on DU.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=132x5393055#5393230

They sort of existed. For a while. They've been gone since the Conventions, and will not return.

[minionofmidas]

Or they both do and will soon be regressing to the mean.

Actually, the odds are 1 in 400 that this will happen.

Based on what?

1 in 20 polls is an outlier.  For each poll to be an outlier, that's 1 x 1 to 20 x 20 to 1 to 400.

One appears to be an outlier, one isn't.  For both to be an outlier, the odds are 1 to 400.

(Ignoring Alcon's pertinent remarks about perfect samples - which as a rule do not exist in opinion polling - for the moment.) Except that that's a random cutoff point. Not every point outside the MoE is as unlikely as every other point outside the MoE, and not every point within the MoE is as likely as every other point within the MoE.

We're not talking a field divided into a 95% field that is white and a 5% field that is black. We're talking a field that is, almost in its entirety, a very whitish shade of grey, moving into downright white along much of the one half, and into darker shades towards the other end, theoretically reaching black at the very extreme edge.

As to the math...

Yes, two out of perfect samples to be of the 5% share of our whitish-grey field that we whimsically declare to be "black" is a 1-in-400 share. But, as Jas points out, we're actually talking at least one of three and at least one of another three.

If there's a 5% chance that any particular poll is an outlier, then the odds that 1 of a set of 3 polls is an outlier is 3(0.05*0.95*0.95) = 0.135 (that is, 13.5%).

Close, but not quite. That is actually the probability that exactly one of the three is a "bad" sample. The probability that at least one is "bad" is actually 1-(0.95*0.95*0.95) = 0.142625 (the part in the bracket is the probability that all three are "good".)

You must be logged in to read this quote.

Correct.
The actual figure then is approximately 2.0342%, or more than one in fifty.

But of course, all of this really presupposes that pollsters are getting an entirely random sample from the population, are not lied to (or even erroneously given wrong information), and are not performing any mathematical tricks on their raw data to make the random sample's demographics fit their notions of statistical reality.

[minionofmidas]

Two polls showing different results, let's say 8%.  Both polls are reliable and have the same SS.  It is NOT more likely for one poll to be 8% off than for both to be 4% off.  End of story.

The odds of being inside/outside the confidence interval is 95% / 5% by construction of these polls. It doesn't matter how much one of the polls is off by or how they compare to each other.

No, it does. It's a Gaussian bell curve (theoretically). 5% is a random cutoff.

To provide the maths to back up Lunar's thesis would require me remembering maths stuff I learned twelve years ago and haven't used for nine, so I won't try.



   
 
 

[minionofmidas]

Two polls showing different results, let's say 8%.  Both polls are reliable and have the same SS.  It is NOT more likely for one poll to be 8% off than for both to be 4% off.  End of story.

The odds of being inside/outside the confidence interval is 95% / 5% by construction of these polls. It doesn't matter how much one of the polls is off by or how they compare to each other.

No, it does. It's a Gaussian bell curve (theoretically). 5% is a random cutoff.

The MOE is constructed on the basis of the "random" cutoff. If one chose a different cutoff, there would be a different MOE. We're saying that a poll is correct if the actual result is within the MOE. That's a 95% chance by construction.

Yes.

You must be logged in to read this quote.

For the question of whether it's within MoE it doesn't matter. For the exact probability of an event, it matters as a matter of course. This isn't about the construction "Margin of Error" - this is about the stuff "Margin of Error" is constructed out of.

[minionofmidas]

Bottom line of my argument:

I agree with Dan, ie it's the model not the poll (or the model and the poll). There's no way this difference is attributable to "a bad sample".
My maths excourse was mostly aimed at showing how absurd it is to try and solve these things through maths alone.

[minionofmidas]

My guess is that the 2nd likely voter model has the party identification as 37-37 D-R like in 2004.

And that is based on?

Not exactly that, necessarily. But he's got the basic point right - the traditional turnout model is more of a Republican Best Case Scenario than anything else. Basically, we know it's wrong, we know it's too Republican, we're just in the dark as to by how much (we don't even know whether the new model is better. Then again, we can't rule out that the new model is still too Republican )

As to all the idiot misinterpretations of the relationship between the pre-04 hackery and what actually happened in 04 (not that the relationship was particularly straightforward): There was unprecented (well. Within the past thirty years) new and youth turnout in 2004. And it did lean Democratic nationally, and it's what kept the EC result respectable.
But it didn't lean Democratic by sufficiently much to swing the election (or to throw the pollsters' traditional turnout models too far off course), and in sizable parts of the country - the South, mostly - it didn't lean Democratic at all. It leant Republican.
Now... these New Voters aren't exactly a monolithic block. There are parallels between the Dems' and the Reps' new voters, but there are differences too (I won't go into detailed stereotyping here).
And now ask yourself: Are the Dems' new voters going to vote Democratic again? You betcha.
Are the Reps' new voters going to vote Republican again? Are they going to vote again? I don't really think so. I still think turnout will probably be somewhat lower than four years ago.

[minionofmidas]

A reminder: https://uselectionatlas.org/POLLS/PRESIDENT/2008D/polls.php?fips=55

PPP claimed the reason for the discrepancy was all other posters (excluding ARG for obvious reasons) were using a traditional turnout model while they were using a new one boosting black and youth turnout.

It turns out PPP actually underestimated Obama.

You just violated J. J.'s Second Rule, again.

I think it might be between the two turnout numbers, but any candidate that will count on it is usually in trouble.

He did not, because he is not a candidate.

[minionofmidas]

A reminder: https://uselectionatlas.org/POLLS/PRESIDENT/2008D/polls.php?fips=55

PPP claimed the reason for the discrepancy was all other posters (excluding ARG for obvious reasons) were using a traditional turnout model while they were using a new one boosting black and youth turnout.

It turns out PPP actually underestimated Obama.

You just violated J. J.'s Second Rule, again.

I think it might be between the two turnout numbers, but any candidate that will count on it is usually in trouble.

He did not, because he is not a candidate.

The rule doesn't just apply to the candidate. 

J. J.'s Second Rule of Elections:  "When a politician or activist talks about a large group of voters that, a. aren't being polled, or b. really going to turn out and swing the election, there is no such group."

The candidate isn't necessarily involved.

They are being polled - just sampled down.

[minionofmidas]

Looking at the fact that McCain can pass away any time because of his old age, seriously can we afford Sarah Palin as our President after 8 years of Bush presidency ?

Seeing as you survived Bush, if just barely, I suppose Palin can't be too bad either.

[minionofmidas]

Alright... is Gallup smoking crack, or is it everybody else?