By Jim Schutze
By Rachel Watts
By Lauren Drewes Daniels
By Anna Merlan
By Lee Escobedo
By Eric Nicholson
"There's a delicious irony."
Off the mark
How could the numbers be so wrong?
Six days before Dallas voters cast their ballots for mayor, the lead story in the Dallas Morning News reported Ron Kirk only modestly ahead of his two major challengers, Domingo Garcia and Darrell Jordan. According to the News' survey of 600 likely voters, Kirk was headed for a certain runoff, with 30 percent of the vote, compared to 23 percent for Garcia, 19 percent for Jordan, and 22 percent undecided.
The actual results, of course, were dramatically different. Kirk racked up 62 percent of the vote--twice what the News poll indicated. Jordan took 23 percent, and Garcia just 13 percent.
Was there a sea change within the electorate during the campaign's final days--an unforeseeable shift in preferences after the poll was conducted, as a result of dramatic campaign events?
Or did the News simply screw up?
The answer, in large part, is the latter.
Yes, Domingo Garcia took a pounding in the campaign's final days--with editorial criticism and unflattering stories in the media. Yes, Kirk snatched the editorial endorsement of Dallas' Only Daily after the poll results were in--a coup his well-funded campaign trumpeted in TV commercials.
But it is also clear that the News made critical mistakes.
The first involves the sample.
According to the News' April 30 story on its poll, 87 percent of its 600 respondents were white; 11 percent were black; and only 2 percent were Hispanic. The poll, noting "that the city keeps no official records of voting activity by ethnicity," assumed that those numbers would reflect actual turnout.
In fact, according to a post-election analysis for the News by respected independent demographer Dan Weiser, usual turnout in a Dallas municipal election is 70 percent white and 30 percent minority (far different from the poll's assumptions). But this year's turnout was 51 percent Anglo and 49 percent minority.
And, given the presence of viable black and Hispanic candidates for mayor, Weiser told BeloWatch, everyone should have expected an unusually high minority turnout. "I was predicting ahead of time," says Weiser, "that you'd get one of the largest black turnouts you'd ever seen--and also a [big] Hispanic turnout."
Another oddity: the News poll showed Kirk receiving only 70 percent of the African-American vote. Even Anglo Democrats receive more than 90 percent of the black vote "consistently," according to Weiser. With a strong black candidate in the race, it was reasonable to expect that he would receive an even bigger margin. Indeed, according to Weiser's analysis for the News, Kirk won 97 percent of the black vote.
Adding to the inaccuracy of the News poll, says Weiser, was an unreasonably large number of those surveyed who failed to state a preference; 22 percent were termed "undecided" and 4 percent "declined to answer." Additional questions would have produced more stated preferences, Weiser says.
According to Weiser, the News' survey suffered from being an inexpensive, "plain-vanilla" poll--done by formula, instead of taking into account through careful analysis the factors that would likely affect this year's results. "They were light on past history. There's an awful lot of science, and pseudo-science, which is based on [the notion that] tomorrow's like today which is like yesterday," he told BeloWatch.
"But there's a lot of times when you can tell tomorrow's not going to be like today."
"It makes a difference how you ask questions. You have to do them in depth. They're not set up to do that."
News research director Barbara Wells, identified as one of two poll supervisors, did not return BeloWatch calls for comment.
In the days after its poll came out, the News' ill-fated poll began to define campaign reality. The paper's political stories regularly spoke of Jordan and Garcia battling it out for the second position in a likely runoff.
After the results were in, the News offered no explicit reference to or explanation of the inaccuracy of its poll results. In fact, the only mention of the paper's survey was a comment that "Mr. Kirk was the front-runner in a poll conducted by The Dallas Morning News shortly before the election."
Well, they did get that right.
Last week's BeloWatch incorrectly reported that News 'Viewpoints' columnist Phil Seib is a political consultant. Seib has not done political consulting work for more than a decade. The item also should have identified him as a professor on the SMU faculty.