Release the Unweighted Exit Poll Data

Personally, I find researching this issue more than a little dizzying. One of the few times that my writing on this blog broke through to into the mainstream was over the issue of poll weighting by Party ID. The issue, first raised by Bush strategist Mathew Dowd in June, gained such force nationally that on three separate occasions, people I know asked me what I thought about "Gallup polling too many Republicans," or something to that effect. These were people who had never read a blog in their lives and had no idea I was personally playing a role in the controversy. Major polling organizations, such as Pew and Gallup, even issued press releases arguing against weighting by Party ID.

Further, what really sent traffic at MyDD skyrocketing, was when Jerome posted early exit results on Election Day. He told me over the phone that the exits were unweighted, but that didn't stop me from running around to the precincts in my neighborhood telling the volunteer I had met earlier in the day the good news. Apparently, as a final, painful irony, it was our own hopes that were raised and then dashed as a result of an unweighted poll... or were they?

After the election, exit polls and poll weighting remain highly controversial. The problem is simple: exit polls are rarely (if ever) divergent from the tabulated election results, but in this case they clearly were. However, those interested in investigating this discrepancy are unable to properly do so, because the public only has access to the post-election, re-weighted version of exit polls. These new exit polls are not divergent with the tabulated results, because they were re-weighted to match the final results.

Now, there is nothing wrong with re-weighting the exit polls after an election to match the final results. As Mark Blumenthal notes:

As regular readers of this site know, NEP weights (or adjusts) the exit polls so that their tabulations of vote preference match reality. This is a long-time standard practice for the national network exit polls. This is fine. The problem comes in when one wishes to examine the discrepancy between the unweighted Election Day exit polls and the final results. Not only is this the discrepancy that is at the source of the controversy, the unweighted exits are also not available to the public. Thus, interested parties have no means of examining this highly unusual discrepancy.

The two best articles I have found on this subject come from the already linked Blumenthal at Mystery Pollster and UPenn professor Steven Freeman. Both come to a similar conclusion. While charges of election fraud based on a discrepancy between exit polls and tabulated results are premature, the discrepancy is unusual enough and the subject matter important enough that the complete unweighted data and methodology must be released for further investigation. To again quote from Blumenthal:

Attention Keith Olbermann!: You want to "continue to cover [voting angst] with all prudent speed?" Excellent. Here is one piece of the puzzle you can help solve. The good news is, you don't need to find some "Deep Throat" informant or submit a Freedom of Information Act request. Just call up NBC's polling director and ask. OK, true, you may need to convince a few colleagues at the other networks to do the same. Nonetheless, the networks own and control the NEP exit poll data, so I'm sure they'll gladly help "debunk" this controversy by making the relevant data available. Right? Indeed. While we as a Party have a lot of important work to tackle, right now this work is certainly of importance. We must put pressure on the networks to release their data if there is any hope of removing the cloud of suspicion hanging over this issue. Be polite and contact them about this today. Remember, the networks are supposed to be serving the public interest, or else they would not have any rights to use public airwaves. Clearly, this is an issue that falls under the domain of the public interest, and as such this is information that should be available to everyone.

Tags: Activism (all tags)

Comments

13 Comments

Bootleg, or fundraise to buy?
Surely somebody has a bootleg copy, or can gain access to one.

Besides that, are we absolutely sure it's not possible to buy the data?  If so, I'm sure many people would be happy to fund it, if mydd puts up a "themometer" or "baseball bat" for that specific purpose.  See how quickly the Greens/Libertarians got their 150k.  A similar sentiment should prevail.

Two specific questions I hope someone can reply to:

 (1) Why has noone gotten hold of a bootleg copy of the data?

 (2) Are we absolutely 100% sure the data cannot be purchased?

by Winger 2004-11-16 12:58PM | 0 recs
I think the chances of y our finding out what
happened with the exit polling with out wide scale public interest is probably very slim. If they fucked up- they don't want to be reminded, if they didn't, then why would they have been so willing to claim that they did- either way- no incentive for them
by bruh21 2004-11-16 01:06PM | 0 recs
exit polling
The way it worked was that there were a number of release times of the polling data. So, the ones I released were from the 2 pm, then came a 4 pm, and finally, a 6 pm. I'd heard that the 2 pm sample was unweighted, not sure about the others, and who knows if what I was told was accurate.
by Jerome Armstrong 2004-11-16 01:17PM | 0 recs
Compare the 2 pm data
to the final results in the following states;
AZ, CO, LA, MI, IA

They're almost a perfect match!

by Coldblue 2004-11-16 01:21PM | 0 recs
Roper should provide data for public access
FYI,

I emailed Edison Research about releasing the raw data for public research.  I got the following response:

"Once the raw data is organized and data based it is turned over to the Roper Institute for academic and public research. We do not own this data, it is the property of the National Election Pool members. They will not have it on file until we finish the organization of the data for all 50 states and the national survey. Once they have it they allow access."

So maybe lobbying Roper might help.

by camilow 2004-11-16 01:30PM | 0 recs
Where is the Media Coverage? the Outrage?
Indeed, the subject matter is incredibly important. And extremely troubling. Yet, other than Obermann at MSNBC, there has been essentially zero coverage on any of the election day problems, the need for transparency in the tabulation of the votes, or -- worst of all -- the possibility that the Republicans have just stolen another one.

Even in Blog-land, there has been scant discussion of these issues. Is this due to the 3,510,000 popular vote differential? If so, that's not a good enough reason. Why? Bush won 286 to 252. Flip either Florida or Ohio, and there is a different President - elect today -- regardless of the popular vote count.

Of course, this is also the salient reason  why -- indisputably -- there was  such a concerted effort to disenfranchise voters, challenge voters, and make it more difficult to vote, particularly in Florida and Ohio.

Moreover, electioneering fraud is a historical fact of life. It has happened and continues to happen. And as we know, in 2000, it is highly likely the Republicans stole Florida and thereby stole the election, with overt help from the Florida Secretary of State and Governor, and 5 members of the U.S. Supreme Court. Then Bush -- for the next 4 years and particularly throughout the 2004 campaign -- lied incessantly and gave every indication of being obsessed with power.

So, there is clearly some good cause to Bush and his crew were prepared to say or do anything to be re-elected. Do we really think the dishonesty would have been VOLUNTARILY halted at the water's edge of overt fraud?

So approaching the issue in reverse fashion, I ask -- Could the no-paper-trail electronic voting system or the vote tabulation / certification process in Florida or Ohio have been subverted in a way that was essentially undetectable?

If so, why should we not affirmatively assume they stole it again?

And, if the Republican Party and its big corporate allies can get away with stealing the most important election of our lives, and the prevailing attitude of the general public and the press is: Oh well, stuff happens; What does this say about the rule of law? Or about the future of democracy in our nation?

My opinion: Definitely. No Doubt. If it is suspicious enough -- and I don't know, although it sure doesn't seem far-fetched  to me -- it should be fully investigated, at the bare minimum. How can we just say SO WHAT?

This whole thing literally reaks of a Cover-up, most foul, with at least passive complicity of an increasingly docile, bought-and-paid-for mass media.

by attydave 2004-11-16 01:39PM | 0 recs
contact info
Do we have a list of contact names, addresses etc.? That would be helpful, as well as a couple of sample letters with talking points.
by sarany 2004-11-16 02:02PM | 0 recs
representative
Chris, Freeman makes the assumption that those exit polls are random and representative of the voting population.  Why is this a valid assumption, if the precincts aren't randomly selected?

And, what about this posting by Ruy Teixeira?  He points out that the unweighted summaries had Dukakis beating Bush in '88.

Isn't it more true that the statewide summaries merely correlate strongly, but have no real predictive relationship?  

Why couldn't the explanation be that since the exit pollsters have an interest in visiting the more diverse precincts, they wouldn't visit as many of the strongly homogenous pro-Bush precincts where his turnout boost (compared to 2000) was most pronounced?

by tunesmith 2004-11-16 03:27PM | 0 recs
Exit Polls
The media are not going to help us - our only hope are the major bloggers and the Kerry people - so far their interest has been tepid - if all bloggers and the Kerry people got enthusiatic and aggressive, and put pressure to get exit polls data published, it would be done.
Until then, nada.
by Dorothy Ligon 2004-11-16 03:32PM | 0 recs
Poll data
I saw a new program on Discovery the other night. A group of Aussie scientists proved (using 21st technology) that Oswald(or someone)fired all three shots from the Texas school book depository. No other gunmen, no other shots...It was presented in a very convincing manner.
I guess my point is: Unless something can be done immediately to rectify any alleged voter fraud(is it possible?), are we all better off not going there? It's very difficult to contemplate the futility.I don't put anything past the involved parties. I don't trust the voting machine manufacturer nor any of the politicians.If this cannot be exposed and proven now, however, we will wind up with a television special in twenty or thirty years proving the sanctity of the Diebold voting machines.
by xpat 2004-11-16 03:49PM | 0 recs
Lets Look At The Bigger Picture
While I probably invite negative response from this comment; by spending such a large amount of time reading between the lines of polls, exit polls, voting data, and so on we risk missing the bigger picture. The election was close in 2004, no doubt about that, and it was far from being perfect. Election 2000 was even closer and there was even a legitimate argument there to be made that one candidate probably 'did' win except for a very flawed process. However with that being said the bigger question is why, in 2000 & 2004, the Democrats could NOT break out to a clear margin as Clinton did in 1992 and 1996? Perhaps we should spend less time being angry at what could have been and more time working on what might be.
by southerndemnut 2004-11-16 08:39PM | 0 recs
All Well And Good, But...
It's not a matter of merely being frustrated at what might have been, it's a problem which may be a fundamental flaw in the integrity of the voting system, which with Republicans in control of the apparatus, may render it impossible for the Democrats to win future elections, regardless of how much work is done to correct the definite problems the Democrats have in presenting their candidates or how much grassroots work is done on its behalf.    
by thurst 2004-11-17 05:06AM | 0 recs
recounts=good, exit polls=bad
I support the recount in Ohio, but I don't understand what we prove from the exit polls.

"exit polls are rarely (if ever) divergent from the tabulated election results,"

Is that actually true?  Was it true in 2000?  Was it true in the California recall?

I like having recounts, but trying to read something into the exit polls seems like a dead end.  So if they turn out different, what are we supposed to say, that people should believe an exit poll and not the actual results? Unless we have hard evidence of real electoral fraud, not circumstantial evidence or speculation, then we just look like bad losers.

Let's insist on recounts where we can, but then at a certain point we have to look forward and not backwards.  We spent four years talking about how 2000, an election in which we really got robbed, and see where it got us.  How well are we going to do if we spend another four years talking about an election in which we have no real evidence other than speculation that we got robbed?

by alhill 2004-11-17 06:34AM | 0 recs

Diaries

Advertise Blogads