Non-Gallup/Rasmussen tracking polls thread

[J. J.]

All right.  Again, which pollsters do you want to throw out of the above list?

And again, none.  Let's look at the last week of state polling, from 10/28 onward.  The only ones (and I can only think of one of these) that shouldn't be counted would be state tracking polls.

I don't see why not?  State tracking polls are just three-day polls that roll.  In its last instance, Muhlenberg was just a three-day poll.  Why not use it?

I think it has a smaller sample size.  It's also, as a tracker, designed to do something slightly different that a standard state poll.

You must be logged in to read this quote.

Then we'll get to 'er.
[/quote]

The first thing I'd like to look at are the states where McCain underperformed on the final set of polls verses those where he over performed (and the same date on Obama). 

I'd then like the subset where McCain underperformed outside the MOE (and Obama underperformed outside of the MOE).

I'll look at margin later.

[J. J.]

So have we figured out yet why McCain did a couple of percent better than the polls in general, apparently?  You know what?  We might not really know. Oh the horror, the horror!  By the way, Obama will be close to a 7% lead before this is all over I suspect.

Historically, it's common for the candidate with a large lead to overpoll a small amount. Yes, even the whites ones, JJ.

This is true, and an argument to be had if Obama overpolled to statistical significance.

Actually, that wasn't true in a race I looked at a while back, Casey in 2006.  He underpolled slightly.  Both Swann and Rendell underpolled

[J. J.]

No, as a tracker, it's the same as a state poll.  It's still interviews being conducted over a three-day period.  Muhlenberg's three-day sample size was 618, bigger than Insider Advantage, ARG, Zogby, and only 7 fewer responses than Mason-Dixon.  No reason to throw it out.  It's a minor point, though, but unless you know something about Muhlenberg I don't, I have to include it for consistency if we're including all polls.

Other than being on the top of my list of appallingly bad polls, I would prefer to to compare trackers (in any state) to non trackers.

You must be logged in to read this quote.

I think that is it, from roughly 10/28 to the final; it would be state by state, of course.  I'll readily agree that it doesn't occur in every state.  After that, I'd like to see where Obama and McCain's final were outside of the MOE.  Plus or minus.

[J. J.]

But trackers are the same as a three-day poll.  Do you understand that?  There is no basis for considering them different entities.  They are the same as a three-day poll.  There's no difference.

Anyway, you didn't answer either of my questions, again.  Do you want all polls, regardless of pollster?  Is option B -- ([Obama minus final Obama]-[McCain versus final McCain]) -- acceptable?

I can even arbitrarily exclude Muhlenberg if you really really want.  Just realize that it makes no sense!

Acceptable but please exclude Muhlenberg.  Yes I want the polls regardless of pollsters, but by state and after 10/28.

[J. J.]

Alcon, my thought was a 1-2 point overestimation of Obama, nationally and that the states could give mixed results. 

Now, in what  states where Obama overpolled did McCain underpoll?

In what states did McCain overpolled did Obama underpoll?

[J. J.]

OK.  While I'm working on this, you can explain to everyone how the Muhlenberg tracking poll (612 phone interviews conducted over three days) is worth excluding while Insider Advantage (588 phone interviews conducted over three days) should be included.

go!

I think it's a different purpose of the poll.  Tracing polls are, well, intended to track.  A state poll is intended to be a snapshot of the electorate.

[J. J.]

OK.  While I'm working on this, you can explain to everyone how the Muhlenberg tracking poll (612 phone interviews conducted over three days) is worth excluding while Insider Advantage (588 phone interviews conducted over three days) should be included.

go!

I think it's a different purpose of the poll.  Tracing polls are, well, intended to track.  A state poll is intended to be a snapshot of the electorate.

A three-day tracking poll is not different from a three-day state poll.  Both are a poll conducted over three days.  Their methodology is totally identical.  They are both a "snapshot" of three days.  There are actually non-tracking polls conducted over more than three days.

In PA, SurveyUSA, Zogby and Strategic Vision all fielded polls that are less of a "snapshot" than the Muhlenburg tracker.

I've given you my reason.  They basically have a different purpose.  Also, part of the same sample is included in another tracking poll. 

[J. J.]

I've given you my reason.  They basically have a different purpose. 

I pointed out how that reason is totally invalid.  A three-day tracking poll is conducted the exact same way as a three-day poll.  You could not distinguish the two.  Everything about them is the same.  Unless you can explain how the specific methodology is different, or demonstrate that it is, you're just making stuff up.

Also, part of the same sample is included in another tracking poll. 

Huh?

A tracking poll that takes a three day sample will only have two days in the prior day's sample. 

Poll released on November 1 includes these samples:

Sample from 10/31
Sample from 10/30
Sample from 10/29

Poll released on November 2 includes these samples:

Sample from 11/1
Sample from 10/31
Sample from 10/30

Two of the samples counted on 11/1 will be in 11/2.

Two non tracking polls taken in the same period, even by the same firm, will not use the same sample.

[J. J.]

Yeah, but a tracking poll released on November 1st is the same thing as a poll conducted from October 29th to 31st.  So, since we're including polls from 10/28 otherwise, why throw a poll from the 29th-31st out?

Because there was also a tracking poll, with two days worth of the same sample, on 10/31.  If you include all those polls you are literally counting the same people over again.  I wouldn't have a problem comparing tracking poll to tracking poll, if we had a lot of tracking polls.

It's apples to oranges.

Now, will you answer my questions:

1.  Now, in what  states where Obama overpolled did McCain underpoll?

2.  In what states did McCain overpolled did Obama underpoll?

I'm interested in factoring out those states were both candidates either underpolled or overpolled.

[J. J.]

Because there was also a tracking poll, with two days worth of the same sample, on 10/31.  If you include all those polls you are literally counting the same people over again.  I wouldn't have a problem comparing tracking poll to tracking poll, if we had a lot of tracking polls.

It's apples to oranges.

Um, no, they don't call back the same people.  It's a random sample each time.  Read their methodology PDF.  It's basically like a continuous three-day poll that rolls over.  So, in the end, it's just a three-day poll.

Ah, yes.  The sample in in the poll on November 1 includes 2/3's of the respondents in the October 31 sample (in a three day sample).  It's not a fresh sample each day; that's where we got all the pro - ________ sample dropping off.

Now, will you answer my questions:

1.  Now, in what  states where Obama overpolled did McCain underpoll?

2.  In what states did McCain overpolled did Obama underpoll?

I'm interested in factoring out those states were both candidates either underpolled or overpolled.

I posted a table of that!  Dude.

Alcon, I'm asking where both candidates where both undercounted or both overcounted. 

[J. J.]

J. J.,

Had someone clarify what you meant.

I posted a Spreadsheet with the poll data.  If you want to compare this to finals, to see where Obama and McCain under- and over-performed, be my guest.  It's useless information, though; it relates more to how hard undecideds are pushed, and how third-parties are managed.  Either way, the table I presented is the closest useful version of that.  I mean, really, what does both McCain and Obama under-polling say?  Double Bradley Effect?  Really:  Probably university pollsters, Mason-Dixon, and other non-pushy pollsters.

I don't have spread sheet.  However, the purpose to look at where both candidates underperformed or overperformed.  This is more to rule out true undecides, really bad polling.  If Obama had 48% and McCain had 46% in a state poll, and the result was 52% to 48%, both candidates underpolled.  That has to be taken into account.

You must be logged in to read this quote.

I could deal with that, so long as there wasn't any double counting.  That was my concern. 

[J. J.]

I don't have spread sheet.  However, the purpose to look at where both candidates underperformed or overperformed.  This is more to rule out true undecides, really bad polling.  If Obama had 48% and McCain had 46% in a state poll, and the result was 52% to 48%, both candidates underpolled.  That has to be taken into account.

I can do that, but before I do, I want to know what methodology we're going to use to make conclusions from it.  I don't want to totally waste my time with what I suspect will amount to "where did low-push pollsters poll?"

What I'm looking at is a pattern.  If both candidates tended to underpoll in a state, it probably is more a function of the polling or of true undecided voters than anything else..  I've been looking more at cases where McCain underpolled and Obama didn't.  I'd like to knock out those states where both underpolled.

[J. J.]

What I'm looking at is a pattern.  If both candidates tended to underpoll in a state, it probably is more a function of the polling or of true undecided voters than anything else..  I've been looking more at cases where McCain underpolled and Obama didn't.  I'd like to knock out those states where both underpolled.

Even though you admit that's more of a function of polling techniques?  Be my guest, bud.  The Spreadsheet is there for you.  Tell me how it goes.

I'm trying to what states both candidates underpolled it, which would be an indication of polling techniques. 

Second, I couldn't get on to the raw data site.  It won't load.

You must be logged in to read this quote.

You must be logged in to read this quote.

So far, we have Obama overpolling, on average, in the tracking polls.  Three of those are out of the MOE (out of seven IIRC).  I did not expect that.

We have Obama overpolling in 21 out of 37 states.  I did not expect that, either, at least unless both candidates underpolled.  My question is, in these states, did McCain underpoll or overpoll?

Remember, I'm looking for a very tiny Bradly Effect.  I said I'd expect to be 1-2 points nationally, and that it would not be even state to state.

[J. J.]

Why would one candidate underpolling not also potentially be an indication of polling techniques?  Why not use the method I developed, wherein you subtract one candidate's underpolling from the other?

One think I noticed in 2006 (3 out 5 times) was that the Black candidate didn't overpoll as much as the white candidate underpolled

You must be logged in to read this quote.

I didn't say trackers were "better."  I said something occurred on trackers (and frankly something I was not expecting).

I would say 3 out of 7 outside the MOE was significant.

You must be logged in to read this quote.

I'm not ignoring anything, in fact, I'm asking a question about it.  I would expect that on some polls, both would underpoll; that could be due to undecides.  I'm interested in cases where both candidates did not underpoll.

Remember, I thought it would be very weak, but present.

[J. J.]

J. J., sincerely, here's why people don't especially take you seriously as an analyst.

3/5 races with black candidates in one year = You bring up multiple times

The 538 thing Lunar mentioned (and I have, several times) = You've totally ignored

Coincidentally, #1 justifies your theory and #2 rejects it.  #2 is, by any reasonable mathematical and scientific measure, more significant.  And yet, which one gets the air time?

Just a thought.

1.  Alcon, frankly, I don't care if you take me seriously as an analyst.

2.  If you recall correctly, I expected a very slight effect, 1-2 points.  It is higher than I expected in the tracking polls (on average), things that we really shouldn't have been seeing; the MOE's was 2 points on two of those polls.  I've also noted two things:

A.  It isn't even state to state.

B.  It tended to manifest itself by an undercounting of the white candidate.

3.  This is the second time, when you've claimed something, invoked "science," and I ask for data, that I will look at.  I never see the data.  I hear a lot of complaints, but no data.

I'm looking for something that I think is not a major effect, but is present; I said that going into this.  Can you provide that data?

[J. J.]

I provided you with the data.  You need to download a spreadsheet program and parse it for yourself.  I don't think your test is meaningful; I think it relies more on pollster methodology differences.  I already performed one test you asked me to do.  It came back with data that disputed your thesis, which you seem to have summarily ignored.  But I already spent an hour doing a meaningful test; you need to spend fifteen minutes getting the software, and then doing your own test.  Especially considering I dispute the meaningfulness of this test, and a test whose meaning we agreed on you've outright ignored.

First, I couldn't even get to the file.  I couldn't even connect to it.

Second, I quoted the results (which had a count of McCain underperforming in more states).  I wanted to see a "natural" undercounting, where both candidates udercounted. 

You must be logged in to read this quote.

Oh, the comparisons of the 2006 polling, which I've never seen you looking at.

You must be logged in to read this quote.

No, Alcon, polling results, compared to actual results are witchcraft. 

I see you god is back, science.  You know, this is the second time I've ask you for data; and the second time you've invoked the name of your god, "science" and not provided the data.

What I am looking for is a trend, like in all polling.  Not if something is statistically significant, because we are dealing with what I expected to be a very small variation.

[J. J.]

I'm trying to provide you with data.  I've linked to it, but you say your computer is not able to open it.  Why are you accusing me of not providing data, and mocking me, because of your technical issues?  You haven't even told me what format you can open.

That's not a rhetorical question, by the way.  Why are you mocking me because of your own technical issues?  I want that answered before I do any further data processing for you.

(P.S. I'm an agnostic.  calling 'science' my god is just stupid, not offensive to me.  it doesn't even really make sense as an insult.)

Alcon, I've asked you to provide the data.  I cannot get it from you, so I'm doing it myself.  And yes, so far there does seem to be a pattern that looks like a Bradley Effect on the state level in some states, I've just gotten across the Mississippi, so bear with me.

[J. J.]

Obama now leads nationally by 6.7%

Which is still out of the MOE.

You must be logged in to read this quote.

No, but it looks like thre was one in OH, MD, and MA in 2006, but not in PA or TN.  Wait until I finish, but a good clue is AZ.    And keep this in mind, the Bradley Effect is not a voting phenomenon, but a polling phenomenon.

[J. J.]

Here is what occurred:

I looked at the states where there was a strong undercounting trend for one candidate (only) with at least one poll undercounting out of the MOE:

McCain:  UT, AZ*, KY*, GA*, NH*, IA*, ND, AK*

Obama:  MA, VT, PA*, IN?*, NM*, CO*, NV*, MI*

For these, I have not included polls were both candidates were underpolled outside of the MOE on the same poll.

Note that the three in red had a high Hispanic population of Mexican origin.

Indiana had two for Obama, but it was Zogby and he might have added the samples from a previous poll. 

I looked at states that had some of these characteristics, either both candidates had a poll outside of the MOE but one strongly underpolled or there was not strong underpolling.

McCain:  MN, OH
Obama:  None

MN, Obama had one poll out of MOE, but McCain had two and underpolled, OH, McCain had two polls out of MOE, but Obama still tended to underpoll more.


Weak, but possible correlation:

McCain:  FL*, NY, DE,

In FL, Obama barely underpolled more than McCain, but one of McCain had one out of the MOE.

Obama:   VA, WA, ME*

VA, Obama  underpolled in one poll outside of MOE, but McCain underpolled more.  In WA, both underpolled but one of Obama's polls was outside of the MOE

*one poll only.

NJ was simply a mess, but the last poll was dead on.

Note this, all the strong undercounted states, McCains were in states that the Republicans won at least once in the last two cycles.  Obama's were split in that regard.  CO, NV, and NM have a large Hispanic segment and the "McCain Effect" might be that a white candidate will underpoll in a race against a black candidate with Hispanic voters of Mexican origins.

Even in the stronger correlation states, there tended to be more polls with McCain out of the MOE than Obama.

In short, Bradley is still there in some places, but it is weakening.  There may be an effect with Hispanic voters.

[J. J.]

Bottom line on the Bradley effect is even if we could prove it has existed for every black candidate (which of course we can't) the sample size would still be too small to prove it was because of race as opposed to the 50 million other reasons why polls can be wrong. We'll have to wait until there's been several hundred black candidates all with sufficient polling data on their elections before we can reach any definitive conclusions one way or another.

If I was cynical, I'd say that the people who invented the concept of a "Bradley effect" did so because they didn't want black candidates to be nominated for office by the political parties. Since I'm not, I won't say that.

Actually, that is not quite the case.  Wilder underpolled, but won.

[J. J.]

Bottom line on the Bradley effect is even if we could prove it has existed for every black candidate (which of course we can't) the sample size would still be too small to prove it was because of race as opposed to the 50 million other reasons why polls can be wrong. We'll have to wait until there's been several hundred black candidates all with sufficient polling data on their elections before we can reach any definitive conclusions one way or another.

If I was cynical, I'd say that the people who invented the concept of a "Bradley effect" did so because they didn't want black candidates to be nominated for office by the political parties. Since I'm not, I won't say that.

Actually, that is not quite the case.  Wilder underpolled, but won.

what does that have to do with the theory?

It's Nym's pseudo-conspiracy theory that the Bradley Effect was invented to discourage black people from running for office.  Wilder won, but underpolled.

The interesting point is what where a traditional Bradley Effect occurred, it looks it occurred with Republicans, people who generally be predisposed not to vote for a Democratic nominee in the first place.  In other words, it is a polling effect, but not a voting effect.  White Republicans voters might be less willing to be honest about race, but they still vote the same way (actually Obama tended to run better than Kerry in the several states I looked at).  In other words, race might be a factor in polling, but not in voting (at least for a Democratic nominee).

It's possible that the polling would have showed a white Doug Wilder winning by 1-2 points back in 1989, though it's impossible to know.

The other is the fairly strong underpolling for Obama in states with large population of Mexican origin.  A voter of Mexican origin (even way back) might be less willing to admit to voting for a black candidate.  There also could be a social issue/religion aspect.  The voter might be less willing to admit to voting for a pro-choice candidate (since most Mexican origin are Catholic, with an increasing sizable minority being Evangelical).

That effect will be the one to watch for in the future.  The traditional Bradley Effect may be decreasing, while this "Reverse Bradley Effect," "McCain Effect" "Mexican-American Effect" might be increasing.  And it looks like more of a voting effect than a polling effect.

[J. J.]

It's Nym's pseudo-conspiracy theory that the Bradley Effect was invented to discourage black people from running for office.  Wilder won, but underpolled.

Now I don't believe in the "conspiracy theory", but the fact that Wilder managed to win has nothing to do with the theory behind it. The mere fact the he underpolled...assuming that was due to race...could be enough to scare parties enough to refrain from nominating black candidates, in theory.

Nym's comment was this:

If I was cynical, I'd say that the people who invented the concept of a "Bradley effect" did so because they didn't want black candidates to be nominated for office by the political parties. Since I'm not, I won't say that.

No, the Bradley Effect was observed, not invented.  It was also observed decreasing in 2006 (and continuing to decrease in 2008).

You must be logged in to read this quote.

No, since I'm looking at polls that scored outside of the MOE, and states where the candidates both underpolled and overpolled.  Poll A was still bad, and that would put it well below the MOE, even if Smith won.

[J. J.]

It's Nym's pseudo-conspiracy theory that the Bradley Effect was invented to discourage black people from running for office.  Wilder won, but underpolled.

Now I don't believe in the "conspiracy theory", but the fact that Wilder managed to win has nothing to do with the theory behind it. The mere fact the he underpolled...assuming that was due to race...could be enough to scare parties enough to refrain from nominating black candidates, in theory.

Nym's comment was this:

If I was cynical, I'd say that the people who invented the concept of a "Bradley effect" did so because they didn't want black candidates to be nominated for office by the political parties. Since I'm not, I won't say that.

No, the Bradley Effect was observed, not invented.  It was also observed decreasing in 2006 (and continuing to decrease in 2008).

You must be logged in to read this quote.

No, since I'm looking at polls that scored outside of the MOE, and states where the candidates both underpolled and overpolled.  Poll A was still bad, and that would put it well below the MOE, even if Smith won.

Part of my point was that it isn't actually observed (in any scientific sense). The sample size is far too small to prove it exists in any scientific manner whatsoever, even if it could be proven that the relatively few black candidates with solid polling statistics to analyze in their races have consistently underpolled (of which the evidence is very mixed at best anyway).

Nym, until a few dozen black candidates run statewide, we won't adequate data.  At best he have this, and I'm looking for an effect that does not appear to be uniform nor large, which is what I expected.  I also expect the effect to change over time.

The underpolling of Mexican-American might be an under sampling but it is something that we should watch.