Obama's Long Ride Down - The Numbers

[Alcon]

J. J., with all respect, it's virtually impossible to justify saying that Wisconsin was "tightening."

You might as well just admit you were wrong so that BRTD will shut up for a while and we can all go on with our lives.

[Alcon]

J. J., with all respect, it's virtually impossible to justify saying that Wisconsin was "tightening."

You might as well just admit you were wrong so that BRTD will shut up for a while and we can all go on with our lives.

You mean:

02-16   ARG   C +6   
02-14   Research 2000   O +5   
02-13   Rasmussen   O +4   
02-12   Public Policy Polling   O +11

I'm sorry, but a +6 Clinton is a tighter race that +11 Obama.

You neglected to include the following:

02-18   ARG   O +10
02-17   Public Policy Polling   O +13

Not only that, but he made that claim explicitly saying he was ignoring ARG.  And there are a number of other flaws in his analysis, but let's keep this to the other topic.  BRTD's messed up enough topics with this as it is.

[Alcon]

When I was posting the tightening, those four polls were the ones out.  PPP looked like an outrider at the time (2/12).  We did also have the ARG numbers all over the place.

No, you posted about the tightening on the evening of Feb 18th.  At that time, the PPP poll from 02-17 (which was consistent with their result from 02-12) had already been added to the forum database.  Even if we exclude ARG and give you the benefit of the doubt and exclude the 02-12 PPP poll as an outlier, we're left with a sum total of three polls for 2008; all of which had Obama holding steady with a 4 point and then 5 point lead.

On what evidence, therefore, were you basing the assertion that the race was tightening?

Exactly.  And even if he wasn't, you'd have to put a huge amount of trust in the ARG poll.  And ignore the fact that all other pollsters were showing a trend that ARG wasn't.  And ignore the fact that J.J.'s original comment explicitly stated he wasn't referring to ARG.

[Alcon]

J. J.,

So, in other words, it's irrelevant who does polls, and:

=THEORETICAL=
Pollster A, 1/1: D+6
Pollster B, 1/1: R+3
Pollster A, 1/8: D+9
Pollster B, 1/8: R+1
=END THEORETICAL==

Is a trend toward the GOP?  Especially if pollster B is crap?  That's a silly interpretation.

And you still are ignoring the fact that you said you were disregarding Zogby and ARG polls in the analysis, which makes for the following pattern (margin, followed by pollster in parentheses):

C+15 (A)
C+13 (A)
C+16 (A)
C+22 (A)
C+7 (A)
O+4 (A)
O+11 (B)
O+4 (C)
O+5 (D)
O+13 (B)

So, we have Obama's highest showing from pollster A, two new polls from pollster B showing a slight Obama trend, and polls C and D showing a closer race than pollster B.  But, after these polls were released, pollster B releases a poll corroborating the pro-Obama trend continuing.  Pollsters C and D's tighter showing is therefore not much of an indicator of trend, so much as pollster differences.

Confusing?  In other words:

1. Pollster A showed a strong, gradual trend toward Obama.  Their last poll showed Obama at his highest level ever.

2. Pollster B found Obama +11.  Afterwards, two pollsters found Obama +4 and Obama +5.  Then Pollster B, clearly finding more pro-Obama results as a whole, found Obama +13.

The only way we can find a pro-Clinton trend is by ignoring the second result from pollster B, and then comparing their first results to results from other firms.  That doesn't make a lick of sense! 

[Alcon]

First of all, let's go back:

02-12   Public Policy Polling   O +11

I don't recall PPP, but let's assume that it's a good poll.

02-13   Rasmussen   O +4 

Good company, did well in 2004

02-14   Research 2000   O +5

Not particularly good, but close to prior results.

02-16   ARG   C +6   

Bad polling in general.  I'm not saying, "Clinton's winning, because ARG is showing she's up six."

I'm also looking at ARG and saying, "It probably isn't 10-12 points off. Obama is probably around +3 to +6."  If the first PPP is correct, and these other polls are correct, Obama has dropped from +11 to +6, perhaps to +3.

Then we get the second PPP; it is Obama +13.  Is this bad methodology, an outrider, or it it right and the bots, ARG, Research 2000 all off?  It turned out to be the third option, which surprised me.  Or, in other words, the opposite of what happened in NH happened in WI.

You are rewriting history.  You said (emphasis mine):

I'm not crazy about ARG or Zogby, but some of the other polls have been showing a tightening of the race.

So, you cannot keep using the final ARG poll as an excuse for your being wrong.  You were ignoring that at the time.  Now we also know you were ignoring PPP.  Either you were ignoring them entirely, or you wrongly failed to check the poll database before making your claim.

Removing PPP and ARG, we have:

Strategic Vision (2/10): Obama +4
Rasmussen (2/13): Obama +4
Research 2000 (2/14): Obama +5

In other words, no trend.

So, that leaves us with three options:

1. You were wrong for including the first PPP poll but not the second; or,
2. You were citing poll results that never actually existed.

Either way, you made a mistake, and the fact that you keep trying to defend it, is perplexing to me.

[Alcon]

Again, you specifically said that you were seeing the trend without ARG, Zogby and - apparently now - PPP.  That leaves three polls that came out Obama +4, Obama +4 and Obama +5.  Please explain how this constitutes a trend.

[Alcon]

In the heart of Pennsylvania.  Obama is no longer able to carry white, middle class and white blue collar voters.  He won 90%+ of the black vote in Philly but lost big in the demographics he was winning before Rev. Jerimiah. 

Obama was always losing working-class whites badly, outside of Wisconsin, which has very different working-class whites.  With which demographics is he doing so badly?

[Alcon]

You're arbitrarily choosing Obama's best showing among the working-class and then comparing his current standing against it...and in a state with a remarkably different working-class culture than Pennsylvania.

Demographic groups don't exist in a vacuum.  Obama hasn't fallen nationally since then, which negates your argument too...