In "The Wizard of Oz," the Wiz explains to Scarecrow that all he needs to prove he has a brain is a piece of paper -- a diploma.
Melvin A. Steinberg, candidate for governor, has a need, too. He needs to prove he's the front-runner in the Democratic field.
In recent weeks, the lieutenant governor has had a piece of paper that he says attests to his political clout. It's called a poll, and he's using it to impress political money men and other prospective supporters.
"It helps with regard to the inside people," Mr. Steinberg said. "People who contribute to political campaigns like to contribute to a winner."
The lieutenant governor said that he, unlike the man he views as his major rival, Prince George's County Executive Parris N. Glendening, has no jobs to pass out or contracts to award to induce support.
"All I have is my experience, my vision and that I'm a likely winner," he said. "This substantiates that aspect."
The question is, is the poll reliable? Or is it, as Mr. Glendening maintains, "not worth the paper it's printed on."
The statewide survey of 814 likely voters, conducted Dec. 13-19 by Mr. Steinberg's pollster, shows the lieutenant governor with a healthy lead over his three Democratic rivals.
According to a Jan. 5 "advisory memorandum" by the polling firm, Cooper & Secrest Associates of Alexandria, Va., Mr. Steinberg was the first choice of 26 percent of the respondents.
Mr. Glendening was the choice of 16 percent. Two relatively recent entrants in the race, state Sen. Mary H. Boergers of Montgomery County and state Sen. American Joe Miedusiewski of Baltimore, drew 7 percent and 6 percent respectively.
Forty-five percent said they were undecided. The margin of error was put at plus or minus 4 percent.
The memorandum containing some of the poll results, though not the poll itself, was leaked to The Sun a few weeks back by a source close to the Steinberg campaign.
When Mr. Glendening was asked about the poll, he denounced it, saying it contained numerous questions that so poisoned the well that the results were grossly distorted and essentially meaningless.
It should have been subtitled "Ten Reasons to Hate Parris Glendening," he fumed.
He's right, at least about the questions, though not necessarily about the results.
The Sun tracked down two people surveyed for the poll, one in Mr. Glendening's camp and the other in Ms. Boergers'.
Both recalled several questions that portrayed Mr. Glendening in an unfavorable light and a couple tending to denigrate Ms. Boergers.
But the two respondents said the attack questions -- routinely asked by pollsters to determine an opponent's vulnerabilities -- were posed after they were queried on the candidate they supported.
Alan Secrest, the pollster, stands behind the survey, saying the negative questions did follow the candidate preference query.
At times, he said, respondents are asked a second time, after the attack questions, about the candidate they prefer, but those responses were not reflected in the poll results cited in the memorandum.
"What you have in that memorandum is solely first blush numbers," said Mr. Secrest, a nationally known pollster.
Emily Smith, Mr. Glendening's campaign manager, said, "I think the numbers they got were after they tested negatives, not prior to."
The problem is that Cooper & Secrest, like most pollsters, does not make available its questions or working data. So candidates cheer their findings or trash them, as it suits their purposes, while waiting for the poll results that really count, the ones they tally on election day.
John T. Willis, Mr. Glendening's chief of staff and political guru, will soon be leaving his $107,000-a-year county job to join the Glendening campaign as a consultant.
Mr. Willis, a Harvard-educated lawyer who might know Maryland political history as well as anyone in the state, has come under criticism in some quarters for activities that seem more political than governmental.