Why the numbers lie


THE POLLS have weighed in, and fully four months before the general election they've already knighted a Republican as governor and re-elected a Democrat to the U.S. Senate.

But before that, the surveys of public opinion, as they're loftily called, chose Parris Glendening as the Democratic nominee for governor and Re. Helen Delich Bentley as his iron-maiden GOP opponent.

And, of course, the polls renominated Democrat Paul Sarbanes to the Senate seat he's held since 1976 and his Republican opponent in the November match-up is -- Ta ta! -- Ruthann Aron of Montgomery County. But Mr. Sarbanes wins re-election again.

Polls are fun. They're terrific reading. They reinforce what people already believe. They comfort the afflicted and afflict the comfortable. At best, they're snapshots of a frozen moment in time.

At worst, they can be dead wrong. And the revealed wisdom is that any two reputable polls taken at the same time will show essentially the same results.

In modern campaigns, polls are used to define a candidate's message, shape media strategy, direct a candidate's scheduling, prop up soft spots and amplify strong points, even determine the way a candidate dresses and speaks.

But as important as polls are in contemporary campaigning, there are certain elements in campaigns that polls are unable to determine.

Polls are great at picking front runners, identifying issues and isolating candidates' assets as well as vulnerabilities, favorables and unfavorables, they're called.

But polls can't measure the depth and breadth of organization strength.

So who's to say that while the professorial Mr. Glendening is the statistical pack leader among Democrats the backslapping lieutenant governor, Melvin "Mickey" Steinberg, hasn't assembled enough pockets of organized support across Maryland to overcome his new fourth-place berth in the polls?

Recall that in 1987, Mayor Kurt Schmoke was leading Clarence "Du" Burns by 35 points in the trial heats and wound up winning the election by only a hairbreadth 5,000 votes.

And before that, in 1978, Mr. Lost Ball In Tall Grass, Harry R. Hughes, was in fourth place in the polls until a last-minute surge blew away the pack, including the numerical leader, Blair Lee 3rd, and delivered the Democratic nomination as well as the governorship to Mr. Hughes.

In 1990, Gov. William Donald Schaefer was many decimal points ahead in the numbers game. But on Election Day, he ended up losing 14 counties and pulled in 60 percent of the vote.

Nor are polls weather forecasters. Even the most resilient voters are more likely to turn out on a festive sunny-side-up day than in a thunderclap downpour. So despite the best intentions of voters and polls, the weather often produces unforeseen consequences.

In 1978, Mr. Lee. then the acting governor, was the statistical leader in the polls going into Election Day. He had the broad support of the political organization as well as the black community and most of its leaders.

But at 5 a.m. on Election Day, the clouds boiled over and uncorked their wrath. Drenching rains drilled down until after the polls closed at 8 p.m. Black voters, who historically vote late in the day and into the evening, never materialized.

The rest, as they say, is history.

Polls are relentlessly present-tense and unable to predict the future. They are not works in progress. Polls can't for example, forecast the reaction of voters to heavy negative advertising.

And while polls can measure basic core support, they are unable to predict with any degree of accuracy which way the undecided vote will break.

To this date, there has been no published evidence that the polls are measuring the intensity of the regional rivalry between the Baltimore area and the Maryland suburbs around Washington.

Washington-area suburbanites believe they're entitled to their first governor since Oden Bowie in 1867. Baltimore voters just haven't caught on yet.

Nor has there been any indication of whether voters are more or less inclined to support a candidate who's been endorsed by Mr. Schmoke, as has Mr. Glendening.

Another problem with polls is that they become self-fulfilling prophecies that assume a life of their own. In a sense, they tell voters how to vote because everybody likes a winner.

After every benchmark poll, news stories predictably begin: "Parris Glendening, the Democratic front runner . . ." or "Lt. Gov. Melvin "Mickey" Steinberg, the early Democratic front runner who has plummeted to fourth place . . ."

So the next time you see poll results, wink and take them for what they are -- a passing profile that has the shelf-life of the day or two in which they were conducted.

Frank A. DeFilippo writes on Maryland politics from Owings Mills.

Copyright © 2019, The Baltimore Sun, a Baltimore Sun Media Group publication | Place an Ad