Archive: Mark Mellman

Veteran CBS News White House Correspondent Bill Plante is the host of CBSNews.com's Smoke-Filled-Room. Each week, Bill invites a top political expert into the Smoke-Filled-Room to answer your questions. This week's guest is Democratic pollster and strategist Mark Mellman.

Plante: Our first viewer, Rev. F. Kevin Murphy writes, “With poll numbers jumping all over the place and the race so very tight in so many states - who is your best bet for a win, and why?”

Mellman: This is a very very close race. What happens in these next 18 or 19 days is going to determine the outcome which is unusual in most presidential contests. In recent years it’s been pretty well decided by now. But this race is very very close. I think some of the polls that have been jumping around are somewhat misleading. The Gallup/CNN/USA Today poll for example, jumps quite a bit with their likely voters, jumps much less with their poll of registered voters, which tells you that a lot of the jumping around has to do with the way they determine who’s likely to vote, and who’s not likely to vote. At the end of the day we won’t really know who’s turned out to vote until Election Day.

Plante: ”Are there methodological reasons for the volatility of poll results in the Presidential race, or are people frequently changing their minds?” asks John Baldridge.

Mellman: It’s an excellent question. There are several methodological reasons, but none of the methodological reasons obviate the possibility that people could be changing their minds. But there are not as many people changing their minds as some of these polls would have us believe.

There are a couple of methodological problems. First, a lot of these polls force people to tell more than they know. People haven’t decided – lots of people haven’t decided who they’re going to vote for. Indeed the exit polls, the post-election polls, will tell us probably that somewhere between 15 and 25 percent of people made up their mind in the last two weeks. So when you now have polls showing there’s 5 percent undecided, something’s clearly wrong. What’s wrong is that people are being forced by pollsters to come to a decision on the phone. It’s not a decision they’ve made in their own minds. Because it’s not a real decision, it could change tomorrow. It’s sort of a random response to the interviewer demanding an answer.

The second kind of methodological problem I alluded to before is these “likely voter” models. Not to become overly arcane and statistical, the reality is, in the CNN/Gallup/USA Today polls, done by people who I respect a great deal, that they have chosen some questions that help them determine who’s likely to vote. There’s more volatility, more change up and down in hose questions than there is in change in who people say they’re going to vote for. A lot of the volatility you’re seeing in the CNN poll is not produced by people changing their minds about who they’re going to vote for for President; rather it’s people changing their answers to questions about whether they’re likely to vote or not. The reality of course is, that the people likely to vote isn’t changing radically from day to day, and that suggests that the questions they are using may be not be the best ways to determine who’s likely to vote and who’s not likely to vote.

The final methodological point I’d like to make has to do with margin of error. This both overused and under understood concept. Often times we talk about the margin of error in polls and we say a particular poll has a margin of error of, say, four points. Well, that’s true, but that margin of error exists on each number – that is – on the Gore number and the Bush number. So the margin of error on the margin is actually eight points. The margin of error for comparing two polls is about one and a half times the margin of error of each. So when you’re comparing the margins, sometimes you have a margin of error of 12 points. The truth is, people are trying sometimes to make too fine a statistical point that the data will support. If you look at the overwhelming majority of polls today they agree on one thing, which is: this is a very very close Presidential race. With one or two exceptions the polls are saying it’s a one or two point difference between Gore and Bush at this stage of the game.

Plante: I've listened to the DNC claim that Internet voting seems to be slanted highly in favor of the Republicans. Most of the national polls have somewhere between 500-2000 respondents, and the "snap polls" after the debates had about 500 respondents. The Internet voting I've participated in shows massive differences in numbers with as many as 60,000 on some of the media’s homepages. The majority of these polls show a drastic difference in what the national polls are telling us. A person can conclude one of two things: (1)- The Internet Polls show a biased Republican voter group as claimed by the Democrats or (2)- the country really is more politically conservative than a recent poll showed. Any opinion on this? S. W. Adams is curious.

Mellman: What gives a poll of 500 or 1000 people statistical reliability, with a certain margin of error, is the fact that every person or voter has an equal probability of being contacted. That random sample is what really makes a difference. What you get in Internet voting is not only excluding half the population that does not have Internet access - and those people tend to be more down-scale, they tend to be less well-educated, they tend, in fact, to be more Democratic. That prevents them from being able to participate a all, but even among the people who have Internet access – the 50 or 60 thousand that choose to go to those sites and participate in voting there, are likely to be different from the population as a whole.

The fact that you have Internet access in the first place, and that you’re motivated enough to go to an Internet site – it doesn’t matter whether there’s 60 thousand or 6 million, or even a thousand people doing that. The reality is that it’s only a particular slice of the electorate. A sample of 500, that’s properly drawn, has much more validity than a sample of 60 thousand that all comes from a certain segment of the electorate.

Plante: Ken asks, “Do the voters still care about campaign finance reform? Neither Bush nor Gore talks about it very much. Why aren’t Nader and Buchannan getting more of the "reform" voters who went for H. Ross Perot?”

Mellman: Bush really can’t talk about campaign finance reform because he’s against it and because he’s raised more special interest money than any Presidential candidate in human history. He’s decided to ignore the reform rules that were put into place to limit special interest money into Presidential campaigns. It’s obvious George Bush is not in a position to talk about this issue.

Gore has talked about campaign finance reform quite a bit. He’s said at the debates, and on the trail, that the first thing he does as President will be to submit the McCain/Feingold campaign finance reform bill to Congress.

There is a very profound difference on this issue. Most voters don’t say it’s the most important thing to them; they say that some substantive issue like education or health care or social security is more important. But there is a broad recognition that it’s very difficult to make progress on things like social security, Medicare, HMO reform unless we limit the power of these special interests. There’s very strong support for limiting the powers of these special interests, even though people say the substantive problems tend to rank higher on their agendas.

Plante: ” I 've been reading lots of neck-and-neck polls and stories, but I have yet to hear anybody project where the Nader voters are (or are not) going to go,” notes Mark Barrett. “With Nader holding 3-4 percent in many polls, could his voters cost Gore the election?”

Mellman: Yes, that is certainly possible. Ralph Nader, at this point, has no chance of winning this Presidential election. I don’t think there’s a serious analyst anywhere that would say Ralph Nader has any possibility at all of being the next President of the United States. Some of the people who are voting for Nader, if he wasn’t on the ballot, may well would vote for Bush. But I think somewhat more of them – some probably wouldn’t vote at all &150; but I think more would end up voting for Gore than for Bush. There may be some close states where that is decisive. At the end of the day I don’t think it will be. I don’t think we’ll have a different outcome if Nader had not been on the ballot.

Iit is possible that the people who have Ralph Nader as their first choice and George Bush as their third choice are going to face a difficult decision as they get closer to election day because they are going to have to decide whether to vote for their first choice and perhaps end up getting their third choice, or whether to go with their second choice to avoid the third choice. It’s a relatively small number and it’s hard to know where they’re going to go, and Nader is going to keep some of them – there’s no doubt. Most of history tells us these third party candidates tend to lose support, rather than gain support as Election Day approaches.

Plante: Have any polls been taken to see if people are voting for a candidate simply because they’re against the opponent and not because they believe in the candidate they’re voting for? Tony Pucci is curious.

Mellman: That is asked quite a bit. It’s asked most often in exit polls after the election. I think we’ll get a very good read on it if you watch CBS on election night or on their web site. At this point I have not seen a recent poll question that goes to that point.

Plante: Long before the polls opened back in 1980, Pat Caddell told Jimmy Carter he was going to lose. Dennis Womack asks, “How far ahead nowadays will the polling community know the outcome of a race with a reasonable certainty?”

Mellman: In these elections where people have won by ten and twenty points it’s pretty easy to tell someone the morning of the election who’s going to win and who’s going to lose. But right now this is a very close election – one or two points. If it stays this way we may not know until election night who wins this election. No pollster may be accurately able to predict. People can make some guesses, people can make some judgments, but the truth is if it stays a one or two point race either way, that’s well within the margin of error of any poll and well within the uncertainty of who’s actually going to turn out and vote. If the election is this close it’s going to be guesswork. When it’s a ten and twenty point margin in the polls, it’s pretty easy to tell your client whether they’re going to win or lose.

Plante: Our final questioner Ruth Ann Grueling writes, “I have been reading that a significant majority of voters who support Clinton's policies, but dislike him personally, are planning to vote for Bush. Why is Gore not able to capitalize on the Clinton administration’s success? With this in mind do you beliee that President Clinton should play a stronger role in the final days of the campaign?”

Mellman: The truth is most people who like Bill Clinton are voting for Al Gore. There are some people who say they like Bill Clinton’s policies, but don’t like him personally, who are voting for Bush. But those who are most ambivalent about a Democratic president, for whatever reasons, are less likely to vote for the Democratic candidate to succeed him, than people who are ambivalent. On the other hand, there’s no question that the Gore campaign has made a very strong effort to have Al Gore be seen as his own man. He doesn’t want to run as an extension of, or third term of Bill Clinton. He wants to be there as his own man, an independent voice, someone with his own vision, his own values, his own point of view. At the same time, he deserves a lot of credit for the successes of the Clinton administration and ought to be getting that credit from the public. It’s a shame that he isn’t.

Plante: Do you think President Clinton should be out there more?

Mellman: I think Clinton will be out there campaigning more. I don’t think you’re going to see them together, because I think it is important for Al Gore to be seen as the leader he is, and as a leader in his own right. I think the President will be out there, helping to make the case and helping to rally the faithful.