Wracked by low achievement, abysmal dropout rates and plunging enrollment, the Detroit Public Schools have spent years struggling to launch and sustain a coherent program of school improvement.

As would-be reformers seek strategies that can finally turn the situation around, Mayor Kwame M. Kilpatrick is seeking more control over the school system. While mayoral control has become increasingly popular in big city school systems, with proponents pointing to the promising experience of Boston and Chicago, Kilpatrick’s efforts are more controversial in light of Detroit’s disappointing recent six-year experience with mayoral control.

In 1999, the Michigan legislature empowered Detroit Mayor Dennis Archer to take over the elected school board’s responsibilities. The reasons cited included years of failed school reform efforts, a graduation rate estimated to be around 30 percent, abysmal academic performance, declining enrollment and troubling financial and accounting practices — sound familiar?

The takeover did not deliver the hoped-for results. Six years of mayoral control ended with a return to an elected school board in 2006, capping a period of heated political conflict, stagnant performance and little substantive change. As University of Michigan professor Jeffrey Mirel has noted, racial tensions and political partisanship overwhelmed the mayor’s efforts.

It would be a mistake, however, to think the Detroit experience suggests that an elected school board is the best local bet for effective reform governance. In a district facing Detroit’s challenges, the ability of mayoral control to provide effective political leadership, an extended time horizon, meaningful accountability and a potential counterweight to the teachers union are vital. The experience under Archer makes clear, though, that these possibilities are just that — and come with no guarantees.

Given the challenges facing Detroit’s schools, sensibly designed and sensibly executed mayoral control is more likely to help than to hurt. However, that emphasizes the need to act with eyes wide open.

Contrary to the strong claims of some proponents, there is no rigorous evidence demonstrating the impact of mayoral control. Most studies find no systematic evidence that mayoral control leads to improved governance or achievement. Just one study to date has examined multiple districts and found achievement gains associated with mayoral control — and those researchers suggested caution in interpreting their findings.

As the previous conflict in Detroit illustrated, mayoral control can also raise legitimate concerns. Mayors may politicize school systems in self-serving ways. Marginal neighborhoods can have more difficulty being heard. In the case of New York City, critics of mayoral control have argued that transparency has suffered.

The reality, however, is that the Detroit Public Schools are so hidebound and district leadership so tangled that mayoral leadership may be imperative if the district is to be righted. That said, how mayoral control is approached next time will matter more than merely whether it is. It’s worth trying again, but it needs to be done right.

One issue, of course, is leadership style. Mayor Archer ran into predictable hostility from the powerful Detroit Federation of Teachers and was unable to marshal civic, parental and business support equal to the challenge. Moreover, allowing the takeover to be defined as a legislative incursion meant that the mayor was at odds with his core constituency. If Kilpatrick is serious, he should take a page from Mayor Adrian Fenty in Washington, DC, who has used aggressive salesmanship and a months-long process of public hearings to win key allies and build broad community support.

Beyond that, any proposal for mayoral control should ensure transparency by instituting disclosure requirements prior to any governance change. In the past decade, in the aftermath of malfeasance at firms like Tyco and WorldCom, corporate America saw that overly cozy boards dropped the ball when it came to providing essential scrutiny. Any plan should entail regular, public hearings and reports.

Those reports should feature agreed-upon metrics that need to be established before any change in control. Measures of performance should include more than test scores and graduation rates, they should also be established for areas like staffing, transportation, student safety, construction and finances.

Fourth, transparency is toothless without oversight. Mayoral accountability only works when local public officials, civic leaders and reporters are prepared to ask hard questions and insist on verifiable measures of performance.

Finally, mayoral control only works where, as in Boston or Chicago, mayors put their reputations on the line and their political clout to work. The mayor needs to be a willing, energetic partner if governance reform is to work.

Frederick M. Hess is director of education policy studies at the American Enterprise Institute and author of "Looking for Leadership: Assessing the Case for Mayoral Control of Urban School Systems." His e-mail is rhess@aei.org.