ad info

Editions | myCNN | Video | Audio | Headline News Brief | Feedback  





Bush signs order opening 'faith-based' charity office for business

Rescues continue 4 days after devastating India earthquake

DaimlerChrysler employees join rapidly swelling ranks of laid-off U.S. workers

Disney's is a goner


4:30pm ET, 4/16









CNN Websites
Networks image

Special Event

Seminole County Absentee Ballots Case Continues

Aired December 6, 2000 - 4:26 p.m. ET


MARTIN SAVIDGE, CNN CORRESPONDENT: I'm Martin Savidge in Tallahassee as we take a look inside the courtroom here, Leon County Circuit Court, the courtroom of Judge Nikki Clark. A case coming from Seminole County, and it is a dispute from the presidential election specifically dealing with the issue of the validity of absentee ballots.

Quite a few absentee ballot, in fact, enough of them that could swing the outcome of the trial -- or the election in favor of Vice President Al Gore. And as hard as it may be to be believed, it is possible that it all could boil down to basically an error in printing as to how these ballots may or may not have been properly applied for and how Republicans dealt with that error that was discovered.

David, essentially, as much as anybody would hate to think that the presidential election could be decided on that, that is sort of getting to the crux of it if it went that way?

DAVID CARDWELL, CNN ELECTION LAW ANALYST: Well, it was hanging by a chad for a long time. So now we come down to one line wasn't put on a post card.

SAVIDGE: And it could make all the difference. On the stand right now is James DeLong. He is a statistician for the plaintiff, in this case a Democratic voter from Seminole County. And he is essentially trying to come up with a very intricate equation here basically saying that if the judge finds in favor of the Democrats in this case, how do we deal with the issue of relief? Do we throw out all the absentee ballots from Seminole County or something less? And that's what he's proposing there.

JAMES DELONG, STATISTICIAN: And of these 10,006 being votes for Governor Bush and Secretary Cheney, and 5,209 votes for Vice President Gore and Senator Lieberman.

UNIDENTIFIED MALE: Just so we're clear, those were votes that were cast absentee in Seminole County?

DELONG: Those are the whole universe of absentee ballot votes cast in Seminole County. And if you knew nothing about the 1,932 counted ballots that you were going to discount, other than that they were absentee votes from Seminole County, then it would be reasonable to say since that since we know nothing other than this, and since that is the break of votes, that is kind of the breakdown -- the proportional breakdown of votes from Seminole County, that they should be discounted in proportioned.

But in reality we know much, much more than that these ballots are from absentee voters in Seminole County. We know the proportion of registered party members in these 1,932. We know that this is a set of ballots that comes from -- that 1,833 divided by 1,932 is about 94 percent registered Republicans. And about 3 percent registered Democrats and 3 percent others. And you don't want to throw away that information. That's valuable information that you can use in making a much better estimate.

UNIDENTIFIED MALE: Sorry to interrupt. Do we have information about the ratio of registered Democrats and registered Republicans amongst the entire group that voted by absentee ballot in Seminole County?

DELONG: Stipulation 31.

UNIDENTIFIED MALE: And what did that information tell you?

DELONG: 9,858 registered Republicans and 4,292 registered Democrats.

UNIDENTIFIED MALE: And how did that ratio compare to the way the votes were actually cast between Governor Bush and Vice President Gore in the population of absentee ballots in Seminole County?

DELONG: Well, it's about 2 1/2-1 for Republicans, in versus Democrats, among those who submitted, and it's a little less than two- to-one in votes cast which means that the kind of independents and others who voted absentee looked like they had a propensity to vote for Gore and Lieberman. But it's quite close.

UNIDENTIFIED MALE: Did you actually calculate what your point estimate would have been had you used the proportional reduction alternative rather than the technique that you employed?

DELONG: I don't want to call it an estimate, because it isn't. It throws away a lot of information. I did calculate that if you discounted ballots in proportion to the overall absentee vote, you would then discount 1,240 kind of ballots for Governor Bush and would discount 646 ballots for Vice President Gore, leaving 594 ballots as the margin.

UNIDENTIFIED MALE: So, even if you did it the other way by looking at the votes among the whole pool, the net reduction for Governor Bush greater than for Vice President Gore would be 594?

DELONG: Would be 594, but I don't see any reason that you should think that you should take that as an estimate.

UNIDENTIFIED MALE: And I know you touched on this already, but could you tell the court why you chose the technique that you used between 1,504 and 1,779 as opposed to the alternate -- the reduction method against the 574.

DELONG: Well, just that you should use all the important pieces of information that you have and everything in kind of the voting literature, everything in the exit polls suggests that party affiliation is an extremely powerful piece of information that you should use and that in addition there are the other kind of fuzzier things about the choice of model, like the fact that these voters were in the sample of those who were sent -- who were on the Republican mailing list and sent Republican postcards.

And it's trying to take account of that piece of information, while at same time guarding against model error by trying to form an estimate that you're sure is low and an estimate that you're sure is high. That led me to the 1,504 to 1,779 range.

KEN WRIGHT (ph), ATTORNEY: Can I have just a moment, Your Honor?

CLARK: Of course.

SAVIDGE: On the stand, we have been listening to the testimony of James Delong. He is a statistician that was hired by the plaintiff, essentially talking about, when it comes to the issue of relief here -- if it goes, the ruling, in favor of the Democrats -- some have speculated that perhaps the judge would be very reticent to throw out all of the absentee ballot because of errors that were committed on a few.

And essentially here, what is being presented by the plaintiffs is: That is not necessarily the case, that the judge could have another option, and that is to statistically figure out what is the correct amount of votes that should be accepted or thrown out mathematically -- perhaps not them all. Let's listen.

DELONG: I was not consistently watching it.

WRIGHT: The numbers that you were discussing on the chart, where you were trying to predict the propensity of the voters to vote according to their party, that's all based on exit polls, is that right?

DELONG: It's based on what exit polls from November 7 tell us about what registered Republicans registered Democrats say they voted for, yes.

WRIGHT: Right. And exit polls can be faulty, we learned on election night, isn't that right?

DELONG: Can be faulty to a degree of one or two or three percentage points when they are not adjusted for turnout. But they cannot be -- or they are highly unlikely to be significantly faulty when you ask about the proportion of people who belong to one political party who voted for that party's candidate.

And even a kind of three percentage point error in, say, estimating from exit polls the Republican propensity to vote for Governor Bush -- which was an extraordinarily large error, much bigger than I think any statistician would think, you know...

WRIGHT: Professor... DELONG: ... possible, that that would produce only a 50-vote swing in my totals.

WRIGHT: Professor Delong, how much exit-poll data do you know of that deals with the absentee-voting?

DELONG: There is no exit-poll data that deals with absentee

(CROSSTALK) WRIGHT: And all of the ballots in this case are absentee ballots.


WRIGHT: Isn't that true?


WRIGHT: OK. So exit polls don't tell us anything about that, do you?

DELONG: Are you suggesting that Republicans casting absentee ballots do not vote for Governor Bush?

WRIGHT: No, all I am suggesting is that exit polls don't tell us how absentees votes are cast. Isn't that true?

DELONG: Exit polls tell us a great deal about how Americans vote, unless absentee ballot are cast by people who have very little relationship to the rest of America. They tell us quite a bit. And, in fact, if you look at the stipulation...

SAVIDGE: You're listening to the cross-examination right now of James Delong. He is a statistician. You are watching a court case here in Leon County Circuit Court here in Tallahassee that deals with Seminole County -- and at stake: thousands of absentee ballot, enough ballots that essentially could alter the outcome of the election, not only in the state of Florida, but the presidential election and the race for the White House.

We'll take a break and come back for more testimony after this.



Back to the top  © 2001 Cable News Network. All Rights Reserved.
Terms under which this service is provided to you.
Read our privacy guidelines.