By contrast, only 23 percent of white respondents said they had a favorable view of Dixon. Sixty-nine 0percent said they had an unfavorable view.

Opinion regarding a Dixon run for mayor in 2015 also broke along racial lines.

A plurality of black voters — 45 percent to 39 percent — said they would consider supporting her.

But whites oppose the idea overwhelmingly. Just 17 percent said they would consider backing her, against 79 percent who said they would not.

Among all poll respondents, 34 percent said they would be open to a Dixon comeback. But her "very, very high" negative ratings among all groups pose a clear hurdle to whatever political ambitions she might harbor, said pollster Steve Raabe, president of OpinionWorks in Annapolis, the firm that conducted the survey for The Baltimore Sun.

The major candidates in the Democratic primary next month all have positive favorable-to-unfavorable ratios. Rawlings-Blake, a prospective Dixon rival, stands out from the pack with a better than 3-to-1 favorable ratio.

OpinionWorks, which is based in Annapolis, conducted the telephone survey of 742 likely Democratic primary voters from Aug. 22 through 24. It has a 3.6 percent margin of error.

After largely vanishing from the public eye immediately after her resignation in February 2010, Dixon, 57, has re-emerged as an influential figure in Democratic circles.

During the primary campaign, she has been sharing advice and the experience she gained in more than 20 years as a City Council member, council president and mayor with several of the challengers who are looking to deny Rawlings-Blake a four-year term — even as she warns her protegees that she might take them on herself when the ban on her seeking office is lifted next year.

Dixon was found guilty in late 2009 of stealing about $500 in gift cards given to her by a developer who said they were intended for the needy. She agreed in January 2010 to resign, donate $45,000 to charity and perform 500 hours of community service to settle theft and perjury charges.

Under the plea deal, her record is to be wiped clean at the end of a two-year probation. She was allowed to keep her $83,000 annual pension.

Dixon, who has been working for the Maryland Minority Contractors Association, told The Sun last month that she enjoys life outside politics. She said she has spent more time with her children, a son in high school and a daughter who recently graduated from college.

She has been volunteering with a mentoring group and a nonprofit organization that works with the homeless to fulfill the community service requirement of her probation.

Yet she has also found time to imagine getting another chance as mayor to focus on priorities she had at City Hall, including repaving roads and "cleaning and greening" Baltimore.

"I definitely love public life, being a public servant," she said. "Even today, people come to me for help. …

"There's a job I want to finish that I started."

Susan Walther thinks it's too late for that. The 64-year-old Charles Village woman said America loves political second acts, but Dixon's failure to come clean when she had the chance has squandered her chance at a do-over.

"She didn't end it the way that would have allowed that door to be opened," Walther said. "She waited until she was effectively removed."

"It was just a blot on the city," she said.

scott.calvert@baltsun.com

The Sun Poll

The Baltimore Sun commissioned a telephone survey of 742 likely Baltimore Democratic primary voters from Aug. 22 to 24. The Sun's pollster, OpinionWorks of Annapolis, used the Baltimore City Board of Elections database to identify registered voters with a history of voting in municipal primary elections, and gathered survey results from those who ranked their likelihood of voting in the upcoming September primary "50-50" or higher. The Sun's sample was designed to approximate the racial, gender, geographic, and age breakdown of the City's Democratic primary voting population, based on turnout patterns averaged over the last three primary elections. The margin of error for questions that reflect the entire sample is 3.6 percentage points, which means that in 95 times out of 100, the answer obtained by surveying every Baltimore City Democratic primary voter would be within 3.6 percentage points of the answer obtained by using the sample.