There was a bit of controversy last night over the huge difference in Florida between the exit poll data and the raw returns. What I’m wondering is, do exit polls generally work and this was just an anomaly or are they generally inaccurate? If they are generally inaccurate then why do we keep using them? If they are generally accurate then what happened yesterday/last night?
-XT
It was only late last night that I realized something I have long ago. Exit polls simply can’t be good predictors of vote outcomes.
Whether exit polls are meant to be predictive or not I don’t know.
Here’s one of many problems. If the poll is stopped before the voting, then one can never know what follows and if one waits until after the actual voting ends, the vote itself is a much better measure of the outcome.
FWIW, I heard someone on a talk show this evening as I was driving home – I THINK it was Dick Morris, but I’m not sure – who said that exit polls up until this election have generally been dead-on accurate. In fact, he said, the exact method has been used in other countries to monitor elections and ascertain that they have not been fixed or stolen. Really. When the real results came in so very different from the EPs, he was astonished and could not believe or understand it. He noticed that many of them were off by about the same 3%. The question was raised as to whether somehow the EPs themselves had been fudged, or whether there was some fundamental flaw in methodology that has just now turned up.
Now, my memory may be faulty, but it seems to me that in previous elections the EPs were pretty close to the ultimate outcome. I don’t know how many exit pollers there are, but it has to be a large number. I can’t see how anyone could have induced most of them to lie. It’s also hard for me to imagine that the organization that they report their findings to was intentionally lying about the results in order to discourage people from voting. I don’t know enough about the mechanics of the EP reporting, or the process by which they are tabulated and projected to know if there could be a simpler explanation tracing back to a central processing or calculating error.
Or maybe the exit polls gave a more accurate picture than the official vote tally – which was produced, many places, by Diebold’s electronic voting machines. :dubious:
I have seen several people saying the Exit Poll/Actual Return mismatch was biggest in Florida and Ohio, (the two most nerve-wrackingly swingy states) and that the mismatch is of an unprecedented width.
And I have heard that the electronic voting machines in Ohio were designed by a company CEO’d by a known big-time GOP fundraiser.
Yeesh…
But I haven’t seen any hard numbers on exit polling. Actually, I take it back. I’ve seen one set, on CNN, which seemed to have exit polls matching returns in Ohio, while I’ve seen another set, on MSNBC, showing the wide disparity I’ve heard mentioned. I don’t know which is to be thought of as the exit polls everyone is talking about.
Any help here would be appreciated.
-FrL-
I thought exit polls in the 2000 election and in several local elections throughout the country had been shown also to have been vastly in error. I don’t have a cite for that, but that was my understanding. Anyone have any proof that, except for this election, election polls have ‘generally been dead-on accurate’??
Not this again. If you have any proof of this I’m all ears BG.
-XT
Actually, I believe Ohio didn’t use electronic voting machines, because the Diebold ones were shown to be insecure.
Florida did, though.
It depends. If the race is close, no, because they have especially large error margins. If the race is not close, sure, but then you don’t need them.
Both Florida and Ohio used Diebold voting machines.
Here in FL, there was a lawsuit filed to force the machines to produce a paper trail. Surprisingly, Jeb Bush was against it…I believe courts ruled that the machines HAD to give a paper trail, but on appeal a higher court said that there was not enough time to get it done.
Where am I supposed to get proof?
Why continue to rant that assertion if there isn’t any then?? Perhaps proof will come out in the future (if its in fact true)…until that time why not give it a rest? It didn’t make a major difference unless it was MASSIVE cheating with those machines…and if it was, then it will come out in the wash.
-XT
I have a theory.
NBC is reporting that it was only the early exit poll that showed Kerry in the lead, and that the final exit polls at the end of the day accurately reflected the outcome.
Two factors at work (IMO):
The people who were more likely to be lined up outside polling places early in the morning were Democrats, highly motivated and eager to oust Bush. Thus they showed up disproportionately in the early exit polls.
Word of the early exit polls leaked onto blogs and message boards (as we saw here yesterday) which may have helped motivate Republican voters to get out and stand in line.
As much as anyone, I would like to see a paper trail on voting machines. (We don’t have that in Georgia.) However, I think what we saw yesterday was not fraud at work, but an unusual and misleading release of preliminary (not final) exit polls.
Actually, it won’t. IF there was cheating, there will be no proof of it. It will always be speculation, because those machines have no paper trail. It is odd that the only exit polls that were widely divergent from the official vote tally were in Ohio and Florida, but I’m open to the possibility the polls were in error.
But the potential for data manipulation must also be conceded. I’m curious to know if anyone acted on this information in order to prevent vote tampering.
It’s not a rant, it’s a perfectly relevant alternative hypothesis to explain why the exit poll results in this election did not match the official vote tally. In most countries, such a mismatch is usually taken as grounds to question the validity of the official tally – that’s why exit polls are done, not to give the TV networks some early results to broadcast. Remember what happened in (Caucasian) Georgia last year?
True, it’s not reasonable to rant (at this point) that there is cheating going on with the electronic machines.
However, it IS reasonable at this point to strongly object to the fact that we’ve been put in the position of not being able to tell for sure whether there is or isn’t. It’s reasonable to ask *why * we don’t have a verifiable system of casting votes. I took this to be the thrust of the original comment.
And if the some of the numbers seem funny, particularly in a way that suggests that they might be fudged in favor of precisely the people who put us in such a position - well, the implementers of the system just brought suspicion upon themselves, didn’t they?
It IS a rant…and one we’ve been hearing here for months now. Hell, its a rant we’ve been hearing since the 2000 election if we take the electronic machines out of the picture…i.e. Republican voter fraud conspiricy.
And its the perfect rant too. No proof? No problem because there is no way to prove it…therefore there is no way to DISPROVE it either. The pink dragon is in my garage…and it flys!!
What I’d still like to see is how accurate other exit polls have been through history (as well as how accurate they were in other places in this election) before automatically jumping on the ‘its got to be Diebold!!’ bandwagon. In fact, I’m not jumping anywhere near said bandwagon until I see some hard proof that this was the case…and something this big is going to be hard to cover up IMO, even without a paper trail.
-XT
Sorry, my last post was addressed to BG.
-XT
Actually, having just come back from Cleveland, I can say that not all of Ohio uses the Diebold voting machines…In fact, I think it may be a majority that did not. What they used in Cleveland were the infamous punchcard machines. I saw one woman checking for chads so carefully…You could tell she was going to be damn sure that her vote counted!
Oh yeah…In regards to Florida, what I would like to see is a comparison between vote totals county-by-county or perhaps even at the precinct level in some areas between 2000 and 2004. (That can probably be found, at least at the county level, on the Florida State website…I haven’t looked.) That to me seems like something that would help us understand not only why the vote differed from exit polls but also from the 2000 vote.
I don’t know, personally I consider it to be the main problem that you can’t prove whether or not the reported vote totals are accurate. Back when we had those crazy paper ballots, you could hold a recount and each side had to either put up or shut up.
On the electronic voting machines: give up. The time to deal with this ended when the polls opened. The results now only exist in an unverifiable form. That’s why the nerds were screaming bloody murder about the whole thing years ago. Now the people who want to believe they were rigged will, the people who don’t won’t.
Or, to wax poetic, “Diebold is god.”
The only point of this debate is to try and move to a verifiable vote for the next election cycle. I think that’s unlikely, because people will accept these results and ignore the debate from now on.