If you were a bookie... Polls and bets on the 2004 elections

One more observation. About that "trendlet" concerning the Nader vote. Its solidifying (the trend, I mean). Very curious. So Kerry supporters please note: it might be a bad idea to mentally add Nader's numbers to Kerry's when looking at three-way race polls. Might not happen.

The Gallup poll is an example. If you take out Nader, 1% goes to Kerry, 1% to Bush and 3% to Other/No opinion/Neither. Same with the CSM one. Take out Nader and his 7% dissolves equally three ways: 2% to Bush, 3% to Kerry and 2% to Not Sure. Zogby has it even starker: take Nader out of the equation and all of his 3% goes to Other/Unsure.

This might mean two things, imho. First, on the thing Zogby's showing: Nader voters just dropping out of the two-way race altogether if he's not in. This suggests that the Nader leftists are so disaffected from Kerry that they'd rather stay out altogether than choose the lesser evil. I.e., bad news for Kerry.

Two, though, on what the other two polls are showing (and I've already observed earlier examples of): almost equal proportions of Nader voters reverting back to Dem and Rep preferences if he's not available. This suggests that Nader is right when he claims he's hurting Bush, too - that he's attracting disaffected rightists as well. That argument is also borne out by the detail in the CSM poll: the region-by-region data. According to this poll, Nader actually does best in the Midwest (9%) and South (8%). Compared to only 5% in the West and Northeast. Hardly yer typical liberals, then.

In both cases, a Democrat's "Nader is a spoiler" argument is starting to sound less convincing. The first thing suggests that the Dems' choice of Kerry as candidate has chased Nader voters beyond the pale, in any case. (Of course you could counter that this might partly be so because Nader is succesfully convincing them that Kerry's that bad). The second thing suggests Nader actually weakens Bush almost as much, just like he says he does.

Anyway - just polls.

IMHO, btw, Mr. Powell has been humiliated so thoroughly by his "colleagues" in the Bush administration that they should be glad if he even votes for Bush in the privacy of the voting booth - let alone run as his sidekick.

That'd be great. (Nader might not be a spoiler.) I had noticed a tiny trendlet here on A2K in that direction, as evidenced on the thread by CerealKiller I think about the thought processes of those who plan to vote for 3rd-party candidates.

Fingers crossed.

Agreed re: Powell. I have this little fantasy that at the 11th hour he will become so fed up with the Bush administration and so dismayed by the idea of them getting another 4 years in power that he does something drastic. Makes the talk-show circuit airing some very dirty laundry. Whatever.

A girl can hope.

fraid, army man, too much ingrained loyalty. but yeh, if only ... a call of conscience to stop the madness ... thats also in the line of duty, isnt it?

perhaps he wont because he thinks there wont be as bad a madness in the second term anyway ... the messianistic project obviously bumped into its own limits in Iraq, and second terms tend to be milder / less ambitious, in any case (reagan, clinton, even fdr)

Re Nader:



I haven't checked the data, but my impression thus far is that the polls will, on average, show that Nader is siphoning.

Like I said, I haven't checked but am pretty sure of what would be the result of checking.

I have the opposite impression.

Hmmm ... well, that graph is of polls up to May 14. And of course, when I first signalled a "trendlet" of things being different on May 17, the point was that it was indeed a break with what I'd previously seen. And the last few polls seem to confirm that observation.

But yeh, havent looked at it systematically - just an impression I got. Will look up more details.

The Iowa Electronic Markets '04 "Winner Takes All" Trading Board does not indicate a lot of investor faith in Kerry. Over the past few days, in fact, the odds are breaking against him.

Pres04_WTA
2004 US Presidential Election Winner Takes All Market


Market Trading Boards, a relatively new phenomonon with little more than a 12 year track record, have displayed statistically significant superior Accuracy (Download Note: 12 Page PDF File) compared to traditional polls, exhibiting a strikingly smaller mean average error. Interestingly, one of the graphs illustrating this seems to show that when facing an incumbent, a challenger polls rather better than the actual outcome of the election. I gotta go back and do some more looking into that.

<thinks again>

One thing that could make it look different in the graph Craven posted is that it only records net totals. Of course, that is the only thing that counts in the end - with how much will a potential Kerry lead be cut or reversed? But it also skips over a large part of the "Nader voter" picture. Lemme explain.

Say, a poll has Nader at 7%. Take out Nader, and Bush gets 3% more, Kerry 4% more. Thats a pretty even spread, which would be what I'm talking about here. But in the graph it would just show up as a 1% swing to Bush. It's both true - just a difference of what you focus on.

Or say, a poll has Nader at 7%, and if you take him out Bush gets 1% more, Kerry gets 2% more, and None/Other gets 4% more. Again, all the graph will show is a 1% swing to Bush. It will not show that Kerry would apparently, at the moment, not win over the preference of almost 3/4s of these Nader voters in any case - even if the guy didn't take part.

Again, of course only the margin of victory counts in the end, that's true. But the suggestion that Nader merely (or mostly) siphons off otherwise safely Democratic votes would in both examples be demonstrably incorrect. It all just depends on what you focus on!

Lies, damn lies and ... ;-)

One sunny day in 2005 an old man approached the White House from across Pennsylvania Avenue, where he'd been sitting on a park bench. He spoke to the U.S. Marine standing guard and said, "I would like to go in and meet with President Bush."

The Marine looked at the man and said, "Sir, Mr. Bush is no longer President and no longer resides here."

The old man said, "Okay" and walked away.

The following day, the same man approached the White House and said to the same Marine, "I would like to go in and meet with President Bush."

The Marine again told the man, "Sir, as I said yesterday, Mr. Bush is no longer president and no longer resides here."

The man thanked him and, again, just walked away.

The third day, the same man approached the White House and spoke to the very same U. S. Marine, saying "I would like to go in and meet with President Bush."

The Marine, understandably agitated at this point, looked at the man and said, "Sir, this is the third day in a row you have been here asking to speak to Mr. Bush. I've told you already that Mr. Bush is no longer the president and no longer resides here. Don't you understand?"

The old man looked at the Marine and said, "Oh, I understand. I just love hearing it."

The Marine snapped to attention, saluted, and said, "See you tomorrow."

Anyone give me homework? <big grin>

(Yes, I am certifiably crazy. No need to tell me.)



Yes, I do detect a trend - be it in a very sorta, kinda, I dunno kind of way ;-)

nimh, you should come and work with me

1. Evidently there's a "pollster slant". Much of the answer depends on how the wuestion is presented. This explains why Zogby has all the Naderists leaning to no one in particular.

2. The percentage for Nader is so small, the average sampling error is difficult to detect. His constituency ranges from 3% to 7%, according to different pollsters.
If this is the case referred to the total of Nader would-be voters, then you can imagine the average sampling error on the potential shift in case Nader doesn't run is even greater.

3. To add it up, all the data says is that not all Nader votes would become democrat votes in case he didn't run.

----

Now, if I were a bookie, I'd try to look for state, not national, polls.
If the polls don't open, we can safely tell that Bush will carry again the states in which he had a lead of over 10% against Gore, and that Kerry will carry those won by Gore by the same margin.
This would assure Bush 115 electoral votes, and Kerry 145 electoral votes.
I just don't see the Republicans winning New York (only with Giuliani in the ticket they stand a slim chance), Massachussets, California or Hawaii. Neither can I see the Democrats prevailing in Texas, Idaho, Alabama or North Carolina.

What I would like to know is how are polls doing is key states such as Ohio, Tennessee, Michigan, Illinois, Missouri and, of course, Florida.

You got a job for me? I'll be looking for one soon ... :wink:


Your overall relativation of the statistical (in)significance of the "trendlet" is right, of course. But this one inference is incorrect - simply because earlier Zogby polls did not have all the Naderists leaning to no one in particular.

I.e., in the Zogby poll from early April, the Nader vote split up into Kerry (2%), Bush (1%) and Other (1%). And in the one from mid-March, they split up into Kerry (2%) and Bush (1%).

So the recent results could just be statistical aberration, but not something special about the way Zogby presents the Q, because then you would've found that showing up in the earlier polls as well. Right?



Here you go:

Dales' Electoral College Breakdown

The Hedgehog Report

Note that they both use the state-by-state polls to compile an overview of what the EC breakdown would be if elections were held now - and that they arrive at quite different outcomes ;-).

Dales now has Bush in the lead with 247/231 (with tossups), while Wissing has Bush ahead 307/231. Both use some personal judgement when deciding whether the latest poll is cause enough to move a state from one category (e.g. lean Kerry) to another (e.g. toss-up).

Dales also has composite maps that go purely on the last poll that came in, without any "editorialising": the one that counts the controversial online Zogby state polls currently has Kerry in the lead by 296/242, the one that doesnt has Bush in the lead 280/231.

And then there's further predictions on the EC breakdown by

Federal Review (now: Bush 274 Kerry 264)

Election Projection (now: Bush 201 Kerry 337)

But they dont just register the latest polls, they use them to make a projection towards the future (as in, if the current trends continue towards ... etc). Sounds like one more step towards coffeeleaves, but anyway.

Quirky thing about it is that all of these are maintained by Bush supporters.

In his book 13 Keys To The Presidency, historian Alan Lichtman proposes a set of parameters, or "Keys", by which one may assess the prospects of a Presidential contest. It is Lichtman's contention, based on his research into the body of American Electoral History, that possession of eight or more of these "Keys" all but assures the holder's victory, while a contestent holding fewer than five of the "Keys" has virtually no chance, not one such ever having been elected. Below is my assessment of the current standings.


KEY 1 (Party Mandate): After the midterm elections, the incumbent party holds more seats in the U.S. House of Representatives than it did after the previous midterm elections.
Advantage:
BUSH

KEY 2 (Contest): There is no serious contest for the incumbent-party nomination.
Advantage:
BUSH

KEY 3 (Incumbency): The incumbent-party candidate is the sitting president.
Advantage:
BUSH

KEY 4 (Third party): There is no significant third-party or independent campaign.
Advantage:
BUSH

KEY 5 (Short-term economy): The economy is not in recession during the election campaign.
Advantage:
BUSH

KEY 6 (Long-term economy): Real per-capita economic growth during the term equals or exceeds mean growth during the previous two terms.
Advantage:
Kerry

KEY 7 (Policy change): The incumbent administration effects major changes in national policy.
Advantage:
BUSH

KEY 8 (Social unrest): There is no sustained social unrest during the term.
Advantage:
BUSH

KEY 9 (Scandal): The incumbent administration is untainted by major scandal.
Advantage:
Disputable

KEY 10 (Foreign/military failure): The incumbent administration suffers no major failure in foreign or military affairs.
Advantage:
Disputable

KEY 11 (Foreign/military success): The incumbent administration achieves a major success in foreign or military affairs.
Advantage:
Disputable

KEY 12 (Incumbent charisma): The incumbent-party candidate is charismatic or a national hero.
Advantage:
Kerry

KEY 13 (Challenger charisma): The challenging-party candidate is not charismatic or a national hero.
Advantage:
BUSH

That gives us:

BUSH: 8

Disputable: 3

Kerry: 2


So, by my assessment, Kerry has essentially no chance to pass the Lichtman Test; only three "Keys" are in play, and to stand a reasonable chance of winning, Kerry requires six added to his tally. Should the Middle East situation contine to trend favorably, particularly as regards Iraq, Kerry is irretrievably doomed, while should foriegn affairs turn against The Incumbent, The Challenger still comes up well short of comfort.

You have some odd math Timber. For example Key 4 goes both ways, you gave it to Bush.

On others you just seem to want to call what I'd call Kerry keys "disputable" and hand Bush all his.

Anywho since I agree with your conclusion no biggie.

edited post above to include more detail on the various websites with state-by-state polling overviews and/or EC breakdowns ...

Great links, nimh.

It all comes down to a dozen states.

Well, CdK, its just my assessment. I suppose #4 could go both ways. On the other hand, I feel categorizing as "Disputable" number 9 really gave me a bit of difficulty; no major scandal has YET attached to the Whitehose ... though there's plenty of effort to that end. My real feeling is that one goes to The Imcumbent as well, at least so far. Now, number 10 and number 11 are, to my oerception, at least marginally more favorable to Bush; I don't see any "Major Failure" in either foriegn or military affairs, and while some don't, I happen to think both Afghanistan and Iraq fall under the description of Military Successes. I see no way any of those three favor Kerry, nor do I see that they mitigate against Bush. But then, like I said, its my assessment. To be fair, I should mention Lichter himself didn't fare too well in 2000. Others using his parameters called at best a very close race, while Lichter hinself called for a handy Gore win.

It all comes down to a dozen states IF the numbers remain constant and IF there is no major national disaster. Also, if we have another 9/11 between now and November, that could change everything.

Michigan:

Kerry 47%
Bush 45%

Kerry 45%
Bush 43%
Nader 3%

EPIC/MRA poll. 600 likely voters. MOE 4%.

All of these numbers are from Rasmussen:

Alabama. MoE 5% (no trend)

Bush 57
Kerry 36

Maine. MoE 5% (no trend)

Kerry 54
Bush 35

Minnesota. MoE 5% (March)

Kerry 48 (47)
Bush 43 (44)

New Jersey. MoE 5% (April)

Kerry 51 (51)
Bush 39 (39)

South Carolina. MoE 5% (no trend)

Bush 49
Kerry 39

See this post at Daily Kosfor CA, GA, IL, MO, NY, NC, OH, OR, PA, TX and VA.