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

Foxfyre wrote:

Can too. Even a clock that is stopped is right twice a day.

Come on Fox, be rational. You're talking about an exit poll that in most all races had both candidates within a percentage point or two of what they actually got. Thats not exactly "Even a clock that is stopped is right twice a day", is it. See my longer answer here.

0 Replies

The whole point Nimh is somebody was using exit polling to indicate Kerry was doing much better than he was most likely actually doing. In my opinion, it was too much across the board to be pure coincidence. The actual spread of the polls is immaterial to me. I acknowledge that most of the results were within the margin of error, and I also know that Bush did better than most projected even tho still within the margin of error.

The exit pollsters, however, I believe operated unethically and should not be hired for another election.

0 Replies

I knew this guy who bought an Edsel and was arrested for drunk driving, his neighbor blamed Edsel and they went out of business. So foxfyre I guess I get your drift.

0 Replies

Foxfyre wrote:

The exit pollsters, however, I believe operated unethically and should not be hired for another election.

The irony is, they already tried that. After the fiasco in 2000, they did fire the exit pollsters, and created a new system.

(Kurtz: "They [..] apologized before Congress, dissolved their Voter News Service consortium, hired two new polling firms, tested the upgraded equipment all year and abandoned their practice of "calling" a state after a majority of polls had closed there, which stoked so much controversy when they initially awarded Florida to Al Gore.")

For sure, they'll be revising their methodologies and trying to find out who they missed. Exit polls work by polling at selected precincts that historically have been representative of the overall state. But in any election, new patterns might emerge, and those precincts might turn out to no longer be representative. For example, if Bush managed to mobilise a bunch of new voters in specific kinds of areas that normally don't produce as high a share of the state's electorate, the system will fail. This year, with its huge numbers of new voters showing up in perhaps irregular geographic patterns was therefore bound to be volatile, polling wise.

The truth is, the exit polls are always going to be off a little bit. What needs to be done more (even more) is to educate people and journalists about how polls work and what they can be taken to mean - and what not. To be fair though, the broadcasters already learned a lot of that lesson this year, that's why we had to wait so long before they called states.

(To any other possibly interested poster (??), I tried to summarize the whole exit poll argument in 10 points in an answer to Fox in the other thread EDIT: And fbaezer added some additional explanation in the next post there.)

0 Replies

sozobe wrote:

You would not BELIEVE how superstitious I am right now. I read that Kerry keeps a four-leaf clover in one pocket, that he got right before things turned around in the primaries, and a buckeye in the other pocket, and I approve wholeheartedly.

Turns out that superstitious guy committed the cardinal superstitious sin -- he started writing his victory speech early yesterday evening, around 7:30 I think it was. Right about when things started to go bad...

0 Replies

And then - what I should have done in this thread immediately after the elections, as a responsible thread master.

Remember that last rundown of polls I listed here before the elections?

So - which polls got it right?

The actual result: Bush 51.1, Kerry 48.0 <gnashes teeth>
Margin: Bush +3,1 <yeah yeah yeah, get on with it>

Who predicted it?

Pew did. The by some conservatives much aligned poll of "The Research Center for the People & the Press" published a vote projection that had said, exactly: Bush 51, Kery 48. Margin: Bush +3

The honour is shared though. In my list, I had the last GWU/Battleground poll, which had said, Bush 50, Kerry 46. But what I'd missed was that both the Dem and the Rep pollster that work together on that Battleground poll also still published their own vote projections (listed by pollingreport.com).

And the Republican pollster Tarrance Group had it right: Bush 51.2, Kerry 47.8 Margin: Bush +3,4

Finally, one runner-up: TIPP, which had projected Bush 50.1, Kerry 48.0, Margin: Bush +2,1.

Just in case you clicked that Pollingreport link above, you'll also see CBS mentioned as polling Bush+2, but they updated that still with a last poll saying Bush+1, so no cigar; and they list Battleground (Bush +4) and Gallup (Bush +2), but those were overridden by their subsequent vote projections, which was a good thing for Tarrance/Battleground (see above) but a bad thing for Gallup (which projected a tie). So no further runner-up statuses there either.

Here it is, then:

1. Pew, 2. Tarrance, 3. TIPP

0 Replies

Oh yeah, and who did worst?

Top three worst:

1. The Economist/YouGov
Its last poll had Kerry up 3%, 50% to 47% for Bush.

2. Harris online
The last Harris telephone survey out had Bush up by 1%. That's pretty average. But its last online survey had Kerry up by 3%: 50%/47%.

3. e.a. Fox News
The last Fox News tracking poll out had Kerry up 2 points at 48% to 46% for Bush.

3. e.a. Lake (D)
Lake was the Democratic pollster that took part, alongside Tarrance, in the GWU Battleground poll. The vote projection it published foresaw a Kerry win by 2,1%.

0 Replies

But you're giving the campaign polling groups. None of these did the exit polls.

0 Replies

The exit polls - complete data now online - have the race correct practically to the percentage point. I already gave you that link a couple of times.

EDIT: Sorry, that was a bit short. It is my understanding that the exit polls are done by one consortium, commissioned by the broadcasting networks together, so no competing numbers there. See my next post.

0 Replies

Just to entertain Cyc, if your list had a 4th place runner-up, it would be Rasmussen's call of Bush (Now Bush The Greater Mr. Green

) 50.2, Kerry (Now The Footnote) 48.5, while Zogby and Harris are among the backmarkers in the pack.

Laughing

0 Replies

Fox, here's two links for the exit polls again: the MSNBC overview of the presidential election exit polls and that of CNN. Thus far I gave the MSNBC link, but the CNN one is actually probably easier in use.

The numbers are all the same though - as far as I understand, the networks work together on the exit polls, collectively commissioning a group of pollsters to deliver them the data. So if you've come across different sets of numbers, those were probably different updates compiled as the data were processed in the course of the day/evening.

It was then up to the networks individually to lay the exit poll data next to the actual results coming in, and draw their own conclusions on when to call states and for whom.

0 Replies

Okay Nimh, there isn't enough stuff out there to glean from the links on exit polling for this election yet. I was constantly scanning all the USA alphabet channels, CNN, Fox, and Cspan during the early election returns and all had Kerry well ahead in the exit polls and were reporting grave concern from the Bush campaign, cautious euphoria in the Kerry campaign. Then they began to identify the early problems - disproportionate number of female samples, etc.

But I honestly don't care enough to continue the effort here. There were problems early on and apparently later the pollsters cleaned up their reports.

The problem was that the premature skewed reporting might have influenced the vote.

Some stuff I did find:

Quote:

"It was on Air Force One on election day that strategist Karl Rove started calling around to get the results of early exit polls. But the line kept breaking down. The only information that came through as the plane descended was a BlackBerry message from an aide that simply read: "Not good." Not long afterward, Rove got a more detailed picture and told the President and senior aides the bad news. Florida Governor Jeb Bush had been saying the state was looking good, and the Bush team had expected to be ahead in Ohio. But Kerry was leading everywhere. "I wanted to throw up," said an aide onboard. Bush was more philosophical: "Well, it is what it is," he told adviser Karen Hughes."
http://www.realclearpolitics.com/

http://news.yahoo.com/news?tmpl=story&cid=694&u=/ap/20041103/ap_on_el_pr/eln_tv_exit_polls_1&printer=1

http://www.nypost.com/postopinion/opedcolumnists/31590.htm

http://www.exit-poll.net/

0 Replies

Foxfyre wrote:

But I honestly don't care enough to continue the effort here.

Thank god. You were starting to remind me of some high school football coach who, after his team wins, keeps complaining because the ref made a bad call in the first quarter.

0 Replies

Well I certainly wouldn't want to upset you Kicky. I just don't like to be told I don't know what I'm talking about when I witnessed it with my own eyes.

0 Replies

Don't worry, you didn't. Sometimes a good analogy comes to me, and even though I know I shouldn't do it, I have to throw it out there. I'm cursed that way.

0 Replies

Thx for the links Fox. From your Yahoo link:

Quote:

Some of these sites cautioned readers not to make too much of the information. The Command Post delivered the news under the headline "Grain of Salt." [..] And Slate warned: "these early exit poll numbers do not divine the name of the winner."

[..]"Doing an early poll is like reporting the results of the game at halftime," Lenski said. "You only have about a third of the information."

[..] Pollsters anticipate a postmortem to find out why that happened. Some possibilities: Democrats were more eager to speak to pollsters than Republicans, or Kerry supporters tended to go to the polls earlier in the day than Bush voters.

[..] None it was used to wrongly call a state, due to extra care after the disastrous election night of four years ago when networks twiced prematurely declared Bush the winner in Florida.

Pretty much my point. If you're gonna pick up on preliminary, raw unprocessed data from midway through the day, of course its not gonna predict the winner right to the percentage point. A reporter would be foolish to take it as such.

Once the exit poll data was complete and processed however, the exit polls became very accurate indeed.

In fact, they became more accurate as time passed - see where I quoted what further preliminary data looked like.

So I dont really see where the problem or the scandal is or the pretty near 100% incompetency comes in.

Anyone foolish enough to take mid-day data and base his decision on that ... I'm sorry. Like I said, if anything needs to be changed, its that people need to be educated better. Its not like Command Post and Slate pretended it was more than it was ... and how their "grain of salt"-type announcements came down to "screaming" that Kerry was ahead, like you were claiming in the other thread, I dunno.

0 Replies

But my whole issue is that the reporters DID report the early raw data almost gleefully which was my original complaint. The final figures, after all the vote was in, of course will be accurate.

0 Replies

Foxfyre wrote:

But my whole issue is that the reporters DID report the early raw data almost gleefully which was my original complaint. The final figures, after all the vote was in, of course will be accurate.

What do you mean?

Exit polls Not Equal

actual election results coming in.

So if the exit polls had really been executed with 100% or 90% incompetency, as you implied, then they would have turned out pretty far off-target even in the end as well, compared with the actual election results, wouldn't they?

But as it happens, they were apparently done well enough to eventually turn out pretty near-accurate numbers.

So the question is not of the exit polls themselves. Its more of people basing judgements on early preliminary data that only refers to, say, the first half of the day. Thats just stupid. The exit pollsters cant help that.

But then again, who actually did?

You've been saying the media was "screaming" that Kerry was winning. But when you give the links, all we see is the most veiled hints by news journalists (""appeared subdued", "a rare sense of doubt", that kind of thing) - with the most explicit hint coming from Fox, which can hardly be said to have been "gleeful" about it. And there's online publications that explicitly headlined all the caveats that applied to the numbers in question: "grain of salt" and all that. Not a single state was called.

<shrugs> The liberal media conspiracy you've been hinting at just doesnt seem to want to emerge from all of this. I mean, we're now down from your original insinuation of the numbers having been tampered with and journalists "screaming" about them to affect the vote, to noting that journalists dropped some vague hints on the basis of data that, due to being incomplete, mid-day data, was still a few percent off. Big deal.

0 Replies

I have been quite consistent in my comments I believe. That you choose to make something different from it is your prerogative. But I'm going to bed.

0 Replies

Fox, ya gotta remember that first reports on just about anything are almost invariably wrong. If any blame is to be laid, it is dual:

1) Irresponsible media types breathlessly relaying incomplete, essentially conjectural information as "breaking news" in the absence of any actual relevant news, attempting thereby to engage and retain the target audience in the interest of nothng more substantial than ratings

and

2) Uncritical acceptance and eager pursuit of such information on the part of folks who really ought to know better than to put much value on "early reports"

0 Replies

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

Related Topics

Quick Links

My Account

able2know