When Democratic Gov. Martin O'Malley began airing advertisements describing his main opponent as a lobbyist for Big Oil, former Republican Gov. Robert L. Ehrlich Jr. focused on the bright side.
"The first two negative O'Malley ads got us even in the polls," Ehrlich told WBAL-AM. "We think this might put us up by four or five" points.
And, sure enough, within days a little-known group called Magellan Strategies published a survey showing that Ehrlich had pulled ahead of O'Malley, though numbers for the two were still within the margin of error.
Republicans say Ehrlich simply had an accurate read on the political landscape. Democrats say the coincidence is too striking.
"We're going to have to get used to the fact that there's going to be a new Republican robo-poll every week," O'Malley said. "And none of them are going to be very favorable to us."
The candidates' comments reflect the growing role of partisan pollsters in political campaigning. While financially struggling news organizations are less inclined to lay out money for the detailed independent surveys that once helped to set the campaign narrative, operations such as the conservative Magellan Strategies and its liberal counterpart, Public Policy Polling, are stepping in to fill the void.
The Maryland governor's race has seen about twice as many telephone polls this year than at the same point four years ago. The great majority have been "robo-polls" — automated, push-button surveys that some say are less accurate than the far pricier live interviews — conducted by partisan outfits as a form of advertising.
"It's buyer beware with any poll," said Larry J. Sabato, director of the Center for Politics at the University of Virginia. "They can be designed to tilt the system in one fashion or another because it's just so easy to structure a poll so that it leans Democrat or Republican."
Both the Ehrlich and O'Malley campaigns are using the partisan pollsters to rally their bases and raise money. But the dearth of independent surveys means unaffiliated analysts also are now relying more heavily on them as well.
The nonpartisan website Real Clear Politics last week used the results of three automated surveys — two by the right-of-center Rasmussen Reports, one by the left-leaning Public Policy Polling — to rate the Maryland governor's race a toss-up.
Ehrlich called that rating good news.
"Every outside poll that comes in we are ahead or tied translates into dollars," he said. "Some people who have been on the sidelines get in the game when they see real momentum."
The goal of an authentic horserace poll is to find out which candidate would win if the election were held on the day of the survey. Demographics and likely voter turnout are among the many factors a pollster considers.
But there are several ways to skew the results: Asking leading questions, focusing on friendly segments of the population, ordering the questions to elicit favorable responses.
The Magellan survey of 752 likely voters showed 46 percent supporting Ehrlich and 43 percent supporting O'Malley. Released on July 2, it appeared in several Washington-area news outlets during the Independence Day weekend, and made its way to Baltimore television and radio by Monday.
Annapolis-based Democratic Party activist Steve Lebowitz was among those who called the poll a sham. In a post on the liberal website the Daily Kos, Lebowitz accused Magellan of "supporting a campaign theme that Team Ehrlich has been pushing for a few weeks — overstating Ehrlich's numbers to bolster fundraising, scare the opposition and distract attention from Mr. Ehrlich's running mate."
Mary Kane's ties to her husband's company, which is being sued federally, had made headlines the weekend the Magellan poll came out.
On its website, the four-year-old Magellan Data and Mapping Strategies boasts that its founders "have managed some of the most challenging data and technology projects for the Republican Party and conservative movement in the past 18 years."
Magellan CEO David Flaherty said his firm has no clients in Maryland and that the poll was "not paid by anybody. It's a marketing strategy. It's a way to generate business."
Lebowitz and others accused Magellan of underestimating the number of likely black voters as a way to push Ehrlich ahead of O'Malley. Magellan's sample was 20 percent African-American. In 2008, with then- Sen. Barack Obama on the presidential ballot, black voters made up a record 25 percent of the electorate. In 2006, the last gubernatorial election, African-Americans accounted for 23 percent of the vote.
Flaherty called Lebowitz's criticism "fair" and said he would likely use 23 percent in future Maryland polls. And there will be more, he said.
"This is why we release everything," Flaherty said. "We're not trying to play games."
The survey that prompted O'Malley's "Republican robo-polls" comment was conducted by Rasmussen Reports, which has released results for four automated Maryland governor's race polls.
The latest, released last week, showed Ehrlich ahead of O'Malley 47 percent to 46 percent — a statistical tie within the 4.5 percentage-point margin of error. Earlier surveys — all have been of 500 likely voters — had shown the Republican gaining on the Democrat.
Scott Rasmussen, a New Jersey-based pollster who has been in the business since the mid-1990s, said O'Malley's criticism was to be expected.
"Whenever a poll comes out that a candidate doesn't like, whether a Republican or Democrat, they want to dismiss it," he said.
Rasmussen, the first pollster to find that Republican state Sen. Scott Brown could win the Massachusetts seat held by Democratic Sen. Edward M. Kennedy, said his results speak for themselves. He said he plans to poll in Maryland at least once a month, more frequently as the November election draws near.
Mark Blumenthal, director of pollster.com — which was just acquired by the liberal website Huffington Post — said Rasmussen is one of the more mysterious polling operations.
"They don't tell us party balance or how many people they are screening out," Blumenthal said.
Rasmussen said his is a media company that sells subscriptions and advertisements. Subscribers get access to more information about the survey and who exactly was polled. No one pays for his polls, he said.
Rasmussen said he always asks the horserace question first, before any issues questions that could affect responses about individual candidates.
A day after the latest Rasmussen Reports survey, Public Policy Polling released a survey showing another statistical tie between O'Malley and Ehrlich. But conservatives say that firm is at least as partisan as the right-leaning pollsters.
"PPP polls tend to have Democrats making up a larger share of the sample than one might expect in a midterm with strong headwinds against the incumbent party," Jim Geraghty wrote last week on the website of the conservative National Review.
North Carolina-based PPP, founded in 2001 by businessman Dean Debnam, plays an active role in that state's politics, taking on only Democrats as private polling clients. Director Tom Jensen said the company has no Maryland clients and conducted the governor's race poll here as a way to advertise.
Jensen, who said PPP had also predicted the Brown win in Massachusetts early on, called criticism that the numbers favor Democrats "a convenient card to play when you don't like the results."
Julie.bykowicz@baltsun.com
http://twitter.com/bykowicz
An earlier version of this story misidentified the publication for which Jim Geraghty wrote about Public Policy Polling. It is the National Review. The story also stated that Public Policy Polling backs Democratic candidates. The firm says that while its founder backs Democratic candidates; the firm itself does not. The Sun regrets the errors.
Recent polls
•Public Policy Polling, released July 13, O'Malley 45 percent, Ehrlich 42 percent (automated survey of 569, Democrat-funded firm)
•Rasmussen Reports, released July 12, Ehrlich 47 percent, O'Malley 46 percent (automated survey of 500, firm funded by subscriptions)
•Magellan Strategies, released July 2, Ehrlich 46 percent, O'Malley 43 percent (automated survey of 752, Republican-funded firm)
•Rasmussen Reports, released June 10, O'Malley 45 percent, Ehrlich 45 percent (automated survey of 500)
• Washington Post, released May 10, O'Malley 49 percent, Ehrlich 41 percent (live interviews of 1,030, conducted by Abt SRBI Inc., a firm with mostly media clients)