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 Circuit Court Hears Democratic Challenge to Absentee Ballots

Aired December 6, 2000 - 4:00 p.m. ET


MARTIN SAVIDGE, CNN CORRESPONDENT: This is Martin Savidge in Tallahassee, Florida, where I'm joined by, well, with the legal assistance, David Cardwell, as we take a look into one of the courtrooms. There are actually two court contests going on today in Leon County Circuit Court. At stake, potentially tens of thousands of absentee ballots, literally enough that could swing the election in favor of Vice President Al Gore, depending on how the courts could rule.

But the subject at debate that is really at debate are absentee ballot applications, and we have been here as Gerry Richman -- he's the lead attorney for the plaintiffs in this case -- was beginning to present some documents, essentially, David, getting into some real evidence here at the trial.

DAVID CARDWELL, CNN ELECTION LAW ANALYST: Right. He's introducing some letters that were sent by the division of elections to the supervisors about the absentee voting law changes -- memoranda, internal memoranda to the supervisor's office on the procedures to be followed -- basically showing that the changes that were made in the legislation, that there was a lot of notice that they needed to be strictly followed.

What he's showing now are photographs of the supervisor of elections office. What Mr. Richman has said in his opening statement, he wanted to show by that, was the physical surroundings, so that the court can see when Mr. Leach (ph) and the other person with him were in the room allegedly putting those numbers on the request forms, Mr. Richman wants to show the physical surroundings where of that and where they and could not be seen.

SAVIDGE: Now, we want to point out that the crux of this is not so much dealing and talking about impropriety with absentee ballots but with the application forms that voters would use in order to request a ballot.

CARDWELL: Right. There's not a question that the ballot itself is illegal, except in the sense that Mr. Richman has made the statement that if the -- if the request for the absentee ballot is illegal, then it flows logically that the ballot itself is illegal.

SAVIDGE: So that the process itself may have been tainted.

CARDWELL: Correct.

SAVIDGE: If these ballots had not had the voter registration numbers on them, then in theory -- or the applications, then the voter would not have been sent the ballot?

CARDWELL: That's correct. The procedure that Sandy Goard's office was following was that if the request came in and it did not have all of the required information, it was put over into a stack or a box of the requests that a ballot would not be mailed to that person.

SAVIDGE: So the Republicans were notified of the error, and as a result, election workers came in and corrected it.

CARDWELL: Well, apparently, they identified the error on their own and contacted the supervisor's office, and that's why she said since it was a request she thought it was reasonable. They're saying that, you know, she let them come in and correct the request forms.

SAVIDGE: All right. Well, there is more testimony about to begin here. Let's take a listen.

UNIDENTIFIED MALE: Your honor, I would offer it into evidence?

JUDGE NIKKI CLARK, LEON COUNTY CIRCUIT COURT: There is no objection is there? Admitted, admitted.


UNIDENTIFIED FEMALE: Counsel, please identify yourself.


UNIDENTIFIED MALE: Professor DeLong, could you briefly tell the court your areas of specialty?

JAMES DELONG, STATISTICIAN: My principal areas of specialty are economic history and macroeconomics in the course of which I use the normal quantitative statistical tool kit of a quantitative economist. (OFF-MIKE) Ph.D. in economics comes from Harvard in 1997, and my two fields of specialization for my Ph.D. degree were economic history and econometrics: That is the application of statistics to economic questions.

UNIDENTIFIED MALE: And Your Honor, I believe that we have an understanding that there's no objection to Professor DeLong's qualifications to participate as an expert here?

CLARK: (OFF-MIKE) proceed.

UNIDENTIFIED MALE: I just have a question (UNINTELLIGIBLE) -- expert in what specific area?


UNIDENTIFIED MALE: Econometrics? OK. Statistics applied to economics.

UNIDENTIFIED MALE: Statistics applied to questions of macroeconomic phenomena, including the matter that you're aware he's going to testify right here.

UNIDENTIFIED MALE: I have, Your Honor, with all due respect, I haven't heard any qualifications with regard to this witness in evaluating absentee ballot statistical data.

CLARK: Well, then you need to qualify your witness, counsel.

UNIDENTIFIED MALE: Your Honor, I'd be happy to do that, but I do believe I had an understanding with Mr. Levy (ph), telphonically achieved last night, that we would have no objection to his being qualified to testify in this proceeding, and I...

CLARK: OK. Well, we don't have to take a break, but if you guys want to confer.

UNIDENTIFIED MALE: Your Honor, he don't object to his testifying and you can give him whatever weight you think is appropriate.

CLARK: Mr. Young?

UNIDENTIFIED MALE: I'll join in that stipulation.

CLARK: Proceed, please.

UNIDENTIFIED MALE: Thank you, your honor. Professor DeLong, could you just tell us when you were retained in this matter?

DELONG: A week ago, a little more than a week ago. I think a week ago yesterday.

UNIDENTIFIED MALE: And do you know who retained you?

DELONG: You did.

UNIDENTIFIED MALE: And do you know who your client is?

DELONG: Harry Jacobs of Seminole County.

UNIDENTIFIED MALE: And are you being compensated today in connection with your testimony?

DELONG: I certainly am.

UNIDENTIFIED MALE: And at what rate?

DELONG: $500 an hour.

UNIDENTIFIED MALE: And how did you determine that rate?

DELONG: It's the normal rate that I've charged in those few cases where I've done the legal work in the past on an hourly basis.

UNIDENTIFIED MALE: And so therefore you've been hired in connection with litigation previously?


UNIDENTIFIED MALE: Have you ever testified before?

DELONG: No, it's never reached that stage.

UNIDENTIFIED MALE: Now, were you retained to examine any particular question in connection with this litigation?

DELONG: Yes, I was.

UNIDENTIFIED MALE: And could you tell the court what that question was?

DELONG: The question was that should the court decide in this case that the appropriate remedy is for it to discount absentee ballots cast in the November 7th general election to the number of 1,932. How should those discounted ballots be apportioned between those that contain votes for Governor Bush and those that contain votes for president -- for Vice President Gore?

UNIDENTIFIED MALE: And in considering that question, was there any information you were interested in learning?

DELONG: I was interested in learning as much as we could possibly learn about the voters who cast the ballots at issue. I was interested in learning their party affiliations. I was interested in learning as much as I could about the propensity of voters in Florida of particular political parties to cast votes for Governor Bush, Vice President Gore and other candidates in the 2000 general election. I was interested in general statistics, the turnout, voter participation, of why people undertake to cast absentee ballots.

I think that sums it up.

UNIDENTIFIED MALE: And did you succeed in obtaining any of that information?

DELONG: I succeeded in attaining a good deal of that information, mostly late last week, through a draft of what I understand has become the final stipulation in this case.

UNIDENTIFIED MALE: Your honor, could I give the witness a copy of the stipulation?


UNIDENTIFIED MALE: Would this assist you in your testimony?

DELONG: It would assist me greatly.

UNIDENTIFIED MALE: Without the exhibit (OFF-MIKE)...

(CROSSTALK) UNIDENTIFIED MALE: And if I could, your honor, just to help along the testimony, I have a demonstrative chart that has some of the numbers from the stipulation on it.

CLARK: Can you place that here so all the attorneys can see it?

UNIDENTIFIED MALE: So many attorneys. But I want Your Honor to see it as well.

CLARK: I can see.

UNIDENTIFIED MALE: Is that any good?

UNIDENTIFIED MALE: I can't see it.

CLARK: Tell you what: Make yourself comfortable where you can see it.

UNIDENTIFIED MALE: (OFF-MIKE) no extra chairs.

CLARK: Welcome to Tallahassee.


If you want to pull up a chair right over here, there's probably room.


Can everybody who needs to see that chart now see it?

OK, counsel, resume.

UNIDENTIFIED MALE: Professor DeLong, can I call your attention to page 6 of the stipulation, items 34, 35, 36 and 37? Do you see those?


UNIDENTIFIED MALE: And could you just -- and if it helps you to use the chart that I've put up, could you tell me if any of the information you were looking to rely on is contained in the stipulation?

DELONG: Indeed, a substantial part of it is.

UNIDENTIFIED MALE: And could you just tell the court what information that is that you relied upon?

DELONG: The number of absentee ballot request forms that were modified and thus become the subject of this case, the number of absentee ballots that were voted and returned pursuant to those particular absentee ballot request forms, the party registration of the voters who returned these absentee ballots.

UNIDENTIFIED MALE: And just so the record is clear, could you attach numbers to those categories?

DELONG: We have 1,833 registered Republicans and we believe 54 registered Democrats out of the 1,932 people who returned their ballots, and these 1,932 returned ballots is out of 2,126 altered ballot request forms.

UNIDENTIFIED MALE: Now, having obtained that information, could you describe generally for the court the analytical techniques you employed in trying to address the question and form an opinion?

DELONG: (UNINTELLIGIBLE) time to move to the next chart?

UNIDENTIFIED MALE: Before you tell me what the opinion is, I think you should tell the court how you did the work for the opinion.

DELONG: Well, the important thing is to get the best point estimate you can of what the propensity of these 1,932 voters to vote for each of the candidates -- Governor Bush and Vice President Gore -- was, and to use what information we have about the characteristics of the voters in order to arrive at that best point estimate. And then to say what standard statistical tools tell us about the confidence with which we can make statistical estimates in this case.

UNIDENTIFIED ATTORNEY: And so did you apply those analytical techniques to the data?

DELONG: Yes, I did.

UNIDENTIFIED ATTORNEY: And did you actually form an opinion on the question you were asked to address?

DELONG: I did form an opinion on the question.

UNIDENTIFIED ATTORNEY: Would it help you to have the chart up?

DELONG: The next chart, yes, that's good.

UNIDENTIFIED ATTORNEY: Before you tell us the opinion, I would ask you to describe the chart.

DELONG: This chart has four bars. The leftmost bar gives the number of absentee ballot requests that had been rejected and that were then altered by operatives of the Republican Party. The second, then, there are 2,126 such ballots, so -- though, as I understand it, there might have been more; that it's not clear the count is completely complete.

Then there are 1,932 absentee ballots returned from absentee ballots sent out pursuant to these rejected absentee ballot request forms. Then there are the 1,833 ballots returned from registered Republicans. And the last column...

UNIDENTIFIED ATTORNEY: Before you get to the last column...

CLARK: Actually, let me ask you a question: Do you have this chart duplicated so that you can pass out copies to myself and to counsel?

UNIDENTIFIED ATTORNEY: Your honor, I have to ask someone else that question. (OFF-MIKE) do we have copies of this chart that we could hand out to people?


UNIDENTIFIED ATTORNEY: The answer is no, your honor. I can try and get them, but I don't...

UNIDENTIFIED MALE: We don't need it.

CLARK: I do.


CLARK: All right, go on.

UNIDENTIFIED ATTORNEY: Before we get to the fourth column, Professor DeLong, is it correct that the first three columns on my second chart are the same as the first?

DELONG: Yes, it is.

UNIDENTIFIED ATTORNEY: Now, if you would like to -- let me ask the question this way: Could you tell me what opinion you reached in connection with the question you were considering?

DELONG: My opinion is that, if the court decides to discount 1,932 absentee ballots in accordance with how it is most likely that those absentee ballots were, in fact, cast in the general election for president, the court should discount ballots in such a way as to discount somewhere between 1,504 and 1,779 more ballots cast for Governor Bush than ballots cast for Vice President Gore.

UNIDENTIFIED ATTORNEY: Let me ask you to pause there -- when you said more ballots cast for Governor Bush than Vice President Gore, do you know the numbers that you would urge the court to discount for each of them that result in that net change?

DELONG: At the bottom of the range -- all right, say, with 1,504, my opinion would be that 1,717 ballots cast for Governor Bush should be discounted and 213 ballots cast for Vice President Gore should be discounted. At the high --

UNIDENTIFIED ATTORNEY: Pause there. I note, in adding those up, the number is slightly less than 1,932. Is there a reason for that?

DELONG: That 15,904 of, kind of, all absentee ballots kind of returned had valid counted votes for president. But there were 15,994 total ballots -- absentee ballots cast. So there are 90 that don't have a vote for president. And, presumably, some of those 90 are among these at issue, and so you should correct for that.

UNIDENTIFIED ATTORNEY: And at the top end of the range, could you explain to me how you calculated the 1,779 net change against Governor Bush?

DELONG: That was simply equal to 1,833 counted Republican ballots and 54 democratic ballots.

UNIDENTIFIED ATTORNEY: And what does that represent -- that calculation?

DELONG: And that is simply a calculation made, assuming that people are Republicans or Democrats for a reason. That we should presume that Republicans are highly likely to cast ballots for the Republican candidate and Democrats highly likely to cast ballots for the democratic candidate. Hence, because we know how many registered voters of each party are in this pool, we can use the difference between party registrations as an estimate of the difference in actual votes cast.

UNIDENTIFIED ATTORNEY: So am I correct, the net change at the high end of your estimated range of 1,779 is just the difference between the votes that were registered Republican votes -- which is 1,833 -- and the registered Democrat votes, 54.


UNIDENTIFIED ATTORNEY: Can you explain to the court, then, how you developed the number at the bottom end of your range, the 1,504 -- not just the calculation but the thinking behind the calculation?

DELONG: Well, it's reasonable -- the probability that a Republican is going to vote for Governor Bush in the last general election is high. The probability that a Democrat is going to vote for Vice President Gore is also high; but this is not a country of strict party discipline. Crossover voting is remarkably common in America and especially...

SAVIDGE: On the stand right now is James DeLong. He is a statistician for the plaintiff. So we're doing a lot of number crunching here.

I'm going to ask David Cardwell to, sort of, tell us what this is all about.

This is actually getting into the stage of relief, actually?

CARDWELL: That's correct. What the statistician, James DeLong is telling the court is that he's gone in and done an analysis of the absentee voting patterns in Seminole County and has taken the number of absentee ballot requests that they know were handled by the Republican Party and had the numbers added -- how many of those were ballots that were then returned.

And he's applying his analysis to those ballots to determine, sort of, the distribution of the vote: How many of those absentee ballots were probably votes for Gore, how many of them were for Bush. And he's telling the court how many could be discounted, for example, if the court wanted to determine exactly how many votes were cast for Gore or for Bush from these so-called questionable absentee ballots. SAVIDGE: And why is that important at this particular time in the trial -- to be going over this issue?

CARDWELL: Well, it appears we're getting near the end of the presentation of their case. In fact, this morning Gerry Richman had said the statistician might be his last live witness.

What they're trying to do here is to give the court an alternative, because -- looking at the facts here and looking at a remedy of knocking out all 15,000 ballots, the court may say, that's very severe, I don't want to disenfranchise others because a few may have had some questions about them.

So what they're doing is giving her an alternative to where she could find in their favor, find there was wrongdoing and throw out a certain number of votes based upon the statistical analysis that would be much lower than the total absentee ballots -- something like, I think he was saying approximately 1,500 ballots may have been cast for George Bush.

So that gives her an alternative. She may not think that's so severe, but it would still have a big impact because then that would still shift the statewide total to be in favor of Vice President Gore, regardless of which of the two statewide certifications you use.

SAVIDGE: But this is, sort of, statistical extrapolation, here. How much weight is it really going to have for the judge, and how much weight has it had in the past, in the case of Judge Sauls, when we've heard testimony like this?

CARDWELL: Well, it didn't convince Judge Sauls when we had the dueling statisticians in his courtroom. We're going to have dueling statisticians here, because Mr. DeLong is being put on the stand now by the plaintiffs; but the defendants indicated they have their statistician. He's going to come on and will refute this.

SAVIDGE: All right, let's go back to the testimony of James DeLong.

DELONG: ... only if the Republican mailing list was completely ineffective. In its, kind of, goal of trying to target voters more likely than average to vote for Republican.

Similarly, the 1,779 number -- it's probably too high, as I said before, because it assumes a degree of party loyalty that you don't find in America.

UNIDENTIFIED ATTORNEY: Let me ask you this question: do you -- did you do any testing or provide any kind of analytics to answer the question as to how likely it is that the right answer is not in the range of 1,779 to 1,504, but either higher than 1,779 or lower than 1,505.

DELONG: Well, so far --

UNIDENTIFIED ATTORNEY: Did you do any of that? DELONG: Yes.

UNIDENTIFIED ATTORNEY: And what did you do?

DELONG: Well, so far we've been talking about -- I guess one way is to call it model error -- that is, the fact that your simplified view of the process and motivations generating the data you're looking at may not be completely correct, may not be completely complete, may be off some way or another.

And, indeed, I've tried to bound the model error by picking both assumptions that produce a number that I regard as almost surely too large. I have a hard time seeing how party discipline or other factors could be strong enough to make our point estimate our best kind of estimate -- bigger than 1,779.

And another way of proceeding that I'm sure produces a kind of most likely point estimate that's too small. Again, its assumption that the Republican Party's that it's targeting efforts were not effective.

In addition to modeling which is almost always the big thing, I mean, any statistical analysis, there is also sampling variability. The simple fact that the luck by sheer kind of chance, it sometimes happens that the sample that you happen to draw and look at is not representative or is a bit off from the population that you want to assess.

Statisticians are best at bounding sampling error. That's what the most important bountiful pieces of their tool coot are for and use standard statistical bounding on sampling error can happily say that there is a 95 percent chance that kind of the true vote is within 50 of the point estimate produced by the true model; that there's a 99.9 percent chance...

UNIDENTIFIED MALE: Let's stop at that one. So, you're saying there's a 95% chance that if your point estimate is somewhere in this range...

SAVIDGE: We are watching testimony inside of the courtroom of Judge Nikki Clark. This is a case from Seminole County that is basically going over the validity of potentially thousands of absentee ballots here, and it quite literally could make all the difference in the presidential election, depending on how the judge may rule. We'll take a break and come back to testimony in just a bit.



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