TERENCE SMITH: One of the fundamental tools of election analysis was conspicuously absent from media coverage Tuesday night-- the exit poll.
CLAIRE SHIPMAN, ABC News: Remember, we don't have any exit polling.
TERENCE SMITH: Exit polls serve not only to help project races before all the returns are in, they provide other data that is dissected by print and broadcast outlets in an effort to gauge voter attitudes.
ANTHONY MASON, CBS News; We are learning how people voted, but we've lost some of the detail: The who voted and why that help us discern patterns in tight races.
TERENCE SMITH: Late Tuesday afternoon, Voter News Service, a consortium of the major networks and the Associated Press that runs an extensive exit polling operation, told its members and subscribers that it could not give them reliable exit poll information.
SPOKESMAN: Remember how you and I were going to be talking about exit polling tonight? Well, not so fast. It's more like we're using mid-1960s technology tonight. And so, to sum up, if you liked Johnson-Goldwater, you'll love this evening.
TERENCE SMITH: Faulty computer software prevented VNS from processing data they collected in 44 states. Some organizations, like CNN, fell back on their own independent precinct surveys; others used phone banks to survey voters; most waited on raw vote tallies, but this information had none of the potential scope and depth of the VNS exit polls.
TERENCE SMITH: Fresh in the minds of the media were the spectacular failures of election night 2000 when VNS supplied data that produced two erroneous calls in the presidential race.
SPOKESMAN: Mr. Gore has won in Florida.
SPOKESMAN: Florida pulled back into the undecided column. Computer and data problem.
SPOKESMAN: George Bush is the president-elect of the United States. He has won the state of Florida, according to our projections.
TERENCE SMITH: Later that night, the race was again declared too close to call.
VNS has spent the two years since that debacle retooling as part of a multimillion dollar effort geared towards the 2004 presidential cycle.
Tuesday night was supposed to be a trial run.
TERENCE SMITH: Joining me now are Mark Halperin, political director for ABC News; Warren Mitofsky, an election analyst for CNN and CBS News; and Maralee Schwartz, national political editor for The Washington Post. Welcome to you all.
Mark Halperin, explain what exit polls are and what value they have in an election like this one.
MARK HALPERIN: They have basically two purposes: First of all, they serve to allow us to understand who voted on Election Day, how they voted, and why. It's an incredible experience for America every two years, going to the polls; we're able to tell who the electorate was made up of. It's a wonderful way to learn about America and to look at America on Election Day.
The second purpose they serve is to allow us to project races, who the winners are in individual races, and that is something that the country likes and it's something the media likes, but it's not a gimmick; it's an important function for the press, a free press, to be able to tell people as quickly as possible who has won national elections. And the exit polls, by surveying voters as they leave the polls around the country, allow us to do that.
TERENCE SMITH: Maralee Schwartz, for a newspaper like The Washington Post that puts so much effort into the coverage of politics, these are important?
MARALEE SCHWARTZ: Oh, yes. Yes. I feel that echoing what Mark said, we were deprived... our readers were deprived of a portrait of America at this moment of time. Election night, we rely much less on exit polls for projections. We don't ever call races on projections.
But at 3:00 that morning, or 3:30 when we had finished our editions for Wednesday, I suddenly realized looking ahead to Thursday's paper, how do we explain this? Who voted? Why did they vote?
And while we'll do a lot of reporting and did do a lot of reporting to help explain that, you're missing a level of demographic information about the American public and that really I think cheats everybody. It's a huge disappointment, and particularly in such an interesting and important election, and as you look ahead to 2004, we're at a bit of a loss.
TERENCE SMITH: Warren Mitofsky, you're steeped in the science of how this sort of thing is done. What went wrong?
WARREN MITOFSKY: Well, VNS was unable to produce the election polls because they couldn't get all of the data through their computer. They had selected the samples, they made all the models, but the computers didn't work fast enough to handle everything that was coming in to them on election night.
TERENCE SMITH: Simply a matter of volume and system overload?
WARREN MITOFSKY: Well, it's a matter of design as much as anything else. You know, a computer's designed to take a certain amount of information, and there's no way to properly test it until you get to election night and you see just what the volume is.
TERENCE SMITH: Mark Halperin, what was... when you learned, I guess sometime Tuesday afternoon, that you finally would in fact be getting this exit poll information, what did you do about it?
MARK HALPERIN: Well, we planned for the possibility that we wouldn't have it, and that we wouldn't be able to see everything we wanted to see.
We assumed we might not see any of what we wanted to see, so we planned in terms of both our news coverage and in terms of election night programming for the possibility that none of the data would work, and we had a system in place that would allow us to do our function of performing a watchdog function over the election returns, as well as being able to tell people as quickly as possible who had won the key races.
TERENCE SMITH: But did it get... Mark Halperin, to the question of why people voted the way they did and their motivations?
MARK HALPERIN: No. Maralee Schwartz is exactly right. We're hoping the data eventually comes out, but to not have it on election night, to not be able to tell people that night and the next day in the newspapers why people voted when the country's paying maximum attention, we weren't able to do that. And as Maralee said, in such an important and interesting election, that loss is particularly painful and not a good thing for the country.
MARALEE SCHWARTZ: And I think the other thing, following up on that is conventional wisdom develops immediately after an election, "oh, it was the presence of the president in all these late races," or "voter turnout that are Democrats were demoralized," and that very well may be true, but our only way of presenting that to the readers or the viewers now in anecdotally.
TERENCE SMITH: Right. And so, not only would you... go ahead, Warren, I'm sorry.
WARREN MITOFSKY: That's exactly why exit polls came into being in the first place. It was to keep politicians from spinning the vote of the people to favor a particular campaign or partisan interest and it was to keep reporters from making up reasons. The exit poll was the voice of the people.
TERENCE SMITH: So, it had a double benefit. For example, Maralee Schwartz, when you try to assess the impact of President Bush's campaigning, do you really know if it made a difference in a given state unless you can talk to those voters?
MARALEE SCHWARTZ: Well, again, your sources of information, as Mr. Mitofsky said, are not totally without bias.
The campaigns, Republican campaigns and the Democratic campaigns, have an advantage to tell you, "oh, my gosh, he made a huge difference." Our voter turnout came up, our own internal polling showed it was better, but... and I'm not saying this may not be the truth. It may be the truth. Look at Saxby Chambliss. Bush was there, Talent in Missouri, maybe Thune in South Dakota, all did exceedingly well and closed gaps and rose ahead after the President was there.
Maybe that's why they did. We just don't know, and we didn't have any voters say, "I went to the polls specifically because George Bush urged me to."
TERENCE SMITH: Mark Halperin, do you expect to get this information? You suggested you were still hoping for it. And if so, at this stage, what would you do with it?
MARK HALPERIN: Well, VNS is still hopeful that they'll be able to aggregate the data and put it out, and we hope they are able to because it is so important.
As Maralee said, in some ways it's too late because this conventional wisdom has set in. My guess is that the exit poll will in fact show what the conventional wisdom is, that President Bush did have a big role in determining the outcome of the election.
But when we get it, it will be used for what it's always been used for and the networks, the Associated Press, the VNS members deserve a measure of credit for having done this over the years, an incredibly complicated and expensive undertaking. It distracts a lot of our resources for material that doesn't necessarily show up on ABC News or even on the Associated Press as much as it's used by historians and newspapers, other people in the academy, and by campaigns themselves to learn about who voted and why they voted.
So I'm hoping when we get it, it can be used for a lot of those same purposes because it is important to understand why what happened, and not just the result.
MARALEE SCHWARTZ: And of course if in fact the conventional wisdom that the president's stops made a big difference is proved... is true, what we don't know and won't know without exit polls, is why it did. Is it because the American people in specific states are supporting him on the war, or that they trust him to make the economy better?
Those are the issues we don't know, and that affects legislation, it affects candidacies down the road, presidential, senatorial. The agenda that will be put forth over the next two years really does often come from the American people, believe it or not, and this is how we find out about it.
TERENCE SMITH: Warren Mitofsky, you were consulting for CNN and CBS. What did CNN do to try to make up the gap?
WARREN MITOFSKY: Well, what CNN did, or at least what they had me do, was select samples of precincts in states with very tight races and gather real vote returns.
The purpose of that effort was to make projections that would be reliable and have a second source of information so that they didn't get led into the trap that they found themselves in in the 2000 election with the mistake in projections in Florida.
TERENCE SMITH: You know, given the problems of two years ago, which you and I discussed on this show, and now the problems of Tuesday night, are we asking too much, or is the world of journalism asking too much of computers and their capacity at this stage to really give a portrait of the American voter on the night of the day that that voter votes?
WARREN MITOFSKY: Oh, I don't think so at all.
Computers are entirely capable of it. There were successful exit polls on Tuesday night. They weren't all done by the networks, but The Los Angeles Times had a successful poll, there was a successful exit poll in Wisconsin.
Computers are more than capable of handling this. I think it's a problem of getting them set up properly, programmed properly.
TERENCE SMITH: So it sounds like the people, not the technology?
WARREN MITOFSKY: Oh, it's the people. It's not the technology.
TERENCE SMITH: Mark Halperin, what... if you look at the record of VNS's essentially 0-2 in the last two elections, what do you do about 2004? How do you solve this problem in the future and protect your own reporting?
MARK HALPERIN: Well, we're evaluating what went wrong this time, we're evaluating what went right there this time, trying to figure out the best way to do the basic functions that VNS does, both vote counting and the national exit poll for the purposes of understanding the electorate and for the purposes of telling people as quickly as possible who won the major elections.
We don't have the luxury of an election to practice on, so the goal is to as quickly as possible evaluate what we saw right and what we saw wrong this time and get ready for the next election.
TERENCE SMITH: Would the networks, Mark Halperin, go back to the business of doing it themselves? I know it's a very costly business.
MARK HALPERIN: Well, I think it's important to do this the right way, and one good thing is we generally agree about what the goal is: The trick is to find the right way to do it. I agree with Warren. There are computers that can handle this problem.
The question is: Can we build a system that will work well?
And again, it's going to take a little bit of time, although we don't have the luxury of waiting month after month, but it's going to take us some time now with the election just passed to study what happened and try to figure out the best way to go forward and do this important function, important not just for ABC News and CNN and the other members, but important for the country.
TERENCE SMITH: Maralee Schwartz, what about for newspapers? I mean, you make even greater in-depth use of in material.
MARALEE SCHWARTZ: Oh, absolutely. And we marry it with much of our own reporting, our own talking to voters, and our own polling that we do with ABC. And all... I don't... we're at the beginning of this reevaluation process, but the seniors editors of The Washington Post and our own polling director have said we will not enter 2004 elections without some system to determine voter attitudes.
TERENCE SMITH: So if not this system, then some other system?
MARALEE SCHWARTZ: Yes.
TERENCE SMITH: Warren Mitofsky, what would you recommend as the improvement or the solution next time around? And what would you recommend to news organizations that want to back it up in some way?
WARREN MITOFSKY: Well, the first thing I'd recommend is that they don't take too much time making up their mind about what it is they're going to do for 2004, because while they can throw money at the problem, they can't create time. And the first primary in 2004 is in the middle of January. That's not very much time to build a new computer system or fix the one they have. So I hope whatever decision they make, they make it quickly. That's the only way they're going to solve the problem in time for the 2004 election.
TERENCE SMITH: Mark Halperin, final word from you very briefly. What would you do differently?
MARK HALPERIN: I think we just need to, as Warren suggested, we need to move quickly and we need to be smart and we need to preserve the functionality that this has provided the country over the years. Two bad elections in a row means you've got to really focus, really bear down and make smart decisions. And I think, again, you can learn from what went wrong and the things that also went right.
TERENCE SMITH: Okay, thank you all three, very much.