Told you David Duke would lose and lose...


November 20, 1991|By THEO LIPPMAN JR.

I TOLD YOU SO. I told you David Duke would lose and lose big. Well, maybe I didn't tell you, but I told several people.

For instance, on Nov. 14, two days before the vote, The Sun's national editor, Ed Goodpaster, asked me what I expected in Louisiana. I said I assumed Edwin Edwards would win with ease. "That's what West says, too," he said.

West is Paul West, The Sun's political correspondent, who was reporting from Louisiana on the race. Of course, it was from reading West's reporting and that of other newspaper types on the scene that I came to my conclusion.

I also told Brad Coker. He's president of Mason-Dixon Opinion Research Inc. of Columbia. He called Nov. 15 to complain about an editorial that morning that began, "If the polls are correct, David Duke will lose." The editorial then cited three polls, including a new Mason-Dixon poll that showed Edwards ahead 49-42 percent.

Coker said his poll did not mean Duke would lose. He said, good naturedly but seriously, "I don't want somebody from Dundalk calling me up after the election saying, because of that editorial, my poll was wrong." He said that since Edwards only got 49 percent of the vote in his numbers, Mason-Dixon wanted it made clear that its professional judgment was that he could lose. Coker said the 9 percent of the vote that he did not give to either candidate could split any number of ways, including all of it going to Duke.

"So Duke could win 51-49," he said. I said, "but I hear Edwards is going to win big." He said, "I know Paul West thinks that, but. . ."

So what's the point? The point is not to pick on Coker. Mason-Dixon is highly respected. Coker's analysis of his own poll data showing Howard County Executive Elizabeth Bobo ahead of Charles Ecker 46-43 percent in 1990 led Coker to predict Ecker would win by 300 votes, and he won by 450. The point is, polls can be very, very wrong.

Pollsters usually won't admit this, for obvious reasons. I imagine when somebody from Dundalk calls up Coker and says his poll was wrong because it showed Edwards could lose, and in fact he won with just over 61 percent of the vote, Coker might say something like this:

"Hey, the poll was right on the money! The margin of error is plus or minus 3.5 percentage points. Edwards' 49 percent, plus 3.5 equals 52.5, and when you add all the 9 percent undecided to Edwards, he gets 61.5 percent, which is only 0.38 percent off of his actual 61.12!"

Which is true, technically. Which is why pollsters can always say they're never wrong. But in fact, they all took a bath in Louisiana. They all showed the race close, and they all said Duke always did better than he polled.

Oh, back to "what's the point?" There are two. For journalists, it is to trust each other more than you trust pollsters. And for voters, it is don't stay away from the voting booth when the polls tell you your candidate is a loser or a cinch.

Baltimore Sun Articles
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.