Frank Keegan: What do you think about that poll?

Opinion,Frank Keegan

Look out people! It?s poll season. Get ready for a flurry of them.

If we could figure out how to tax these things (and stop wanton government spending) the deficit would disappear overnight.

Since we can?t tax them, we might as well enjoy them.

The Baltimore Examiner and Baltimore Research teamed up to entertain as well as inform during what could be a dreary election season.

Starting today, every week between now and the election you will get a straightforward, simple "casual" poll result on a major Maryland race. You also will get answers to a fun question or responses on a major issue.

We should take these ? and all polls ? for what they are worth, then make up our own minds.

Polling in the 21st century is a slippery beast. As our Information Age fractures and splits into ever finer, more rapidly changing segments, getting hold of what the general public thinks at any given moment is like trying to grab a cloud.

Traditional, old-fashioned 20th century metrics fall apart. A kind of social and political Heisenberg Uncertainty Principle inexorably distorts results. By the time a large enough sample is selected, questioned, sorted and analyzed, reality has moved on. It might have been accurate when done. But now, who knows?

Plot the universe of all polls along their line of standard deviation, and I bet the aggregate accuracy would closely resemble luck.

The new millennium requires new methods.

Baltimore Research CEO Jay White thought so 10 years ago when he bought the three-decade-old company. He and partner Ted Watson were clients when they worked for the National Fruit Product Co.

They reinvented Baltimore Research and grew the business 10-fold. They conceived a wide array of unorthodox ? and adapted traditional ? methodologies to get a fast, tightly focused snapshot of public opinion.

Clients include NBC, CBS, General Motors, Coke, Johnson & Johnson, and even lawyers preparing for trial.

For example, conventional wisdom says a precisely selected sample of 500 or more people is required for an accurate survey. White says he can do it faster with fewer if need be, depending on the products, issues or questions.

So, how does Baltimore Research do in the tumultuous political arena?

You decide. The firm did the polling for Dennis Rasmussen, candidate for the Democratic U.S. Senate nomination. Gerry Patnode, a Baltimore Research consultant, ran his campaign.

One of their polls showed Rasmussen at 11 percent when no other had him above 5. He got less than 2 percent of the vote.

The Baltimore Examiner asked White for a casual poll with two questions tagged onto other polling Baltimore Research is doing every day for other purposes.

According to White, this means "the data ... have limited reliability and are in keeping with the ?pulse of the consumer?" we here at The Examiner asked for.

"This is not a statewide survey," White said. "The sample is selected logistically for our purposes for other surveys. We believed the point was to see just how close a ?casual poll? came to the predictions of more accepted ? and expensive ? methods, and finally how casual and scientific polls compare to the actual outcome of the election."

For this poll they questioned 303 people about their choice of candidate in the upcoming gubernatorial election. Of those responding, 30.4 percent said they planned to vote for Republican Robert Ehrlich; 37.3 percent said they planned to vote for Democrat Martin O?Malley; 2.3 said "Other"; 22.8 percent said they were undecided; and 7.2 percent said they did not planto vote. Quibble about accuracy all night long, but there is no arguing about speed. Results were on the table with a turnaround time of less than 12 hours.

One interesting comparison is how Maryland voters registered as of October 2002 and ?04.

Democrats were 56 and 55.3 percent respectively. Republicans: 30 and 29.2 percent. Unaffiliated 13.5 and 15.1.

No matter how close or off the political reality mark this poll may have been on one day nine days after the primary, the message to Gov. Robert Ehrlich and Mayor Martin O?Malley is clear: Lots of voters are undecided.

Undecided voters pay attention to what politicians say and do. Politicians pay attention to undecided voters.

So let the polls begin. But in the end, as always, the only one that really counts is on Election Day.

Frank Keegan is editor of The Baltimore Examiner. He can be reached at
View article comments Leave a comment