The following article originally appeared in The Harvard Education Letter (volume 15, number 6). Copyright 1999 President and Fellows of Harvard College. All rights reserved.

It requires hard work to foster and keep good schools in poor communities, and that work has never been so important. With the trend toward resegregation and with the virtual abolishment of affirmative action, Title I remains one of the few means to narrow the achievement gap between affluent and disadvantaged children.

Title I has been the largest federal program for impoverished public schools for more than 30 years. This legislation is expiring and must be renewed by Congress, which seems determined to turn the dollars over to the states with no real protections against their misuse. Politicians who say they care deeply about the fate of our children should look carefully at the evidence on Title I and vote on the basis of proven research, rather than on political expedience. If Title I is the ark that carries vulnerable children from dysfunctional and overwhelmed schools to hopeful futures, it has to be rebuilt with the greatest care.

Unfortunately, little serious research has been done on why Title I programs have not produced larger gains. In an effort to inform the current debate, The Civil Rights Project at Harvard initiated a research project on Title I's impact and potential. The project, which involved researchers from Harvard, RAND, Johns Hopkins, Teachers College, and other institutions, critically examines some of the basic assumptions underlying the education reform efforts of the last two decades. The research also contributes solid evidence about paths to educational gains and underscores the civil rights implications in this legislation that affects millions of black and Latino students.

This new research suggests that Title I dollars should be directed toward ensuring greater accountability in the program. That means funding programs that have solid evidence of impact on learning, such as Success For All and Reading Recovery; enforcing accountability provisions for schoolwide use of funds; and terminating funding for existing policies where no benefits can be documented in independent evaluations.

According to a recent report by the American Institutes of Research, only a few of the 17 eligible research-based programs listed in the Comprehensive School Reform Demonstration Program show strong evidence of benefits. Lawmakers should also consider ensuring that research and experimentation on additional approaches are independently evaluated with random assignment experiments when possible.

To be successful, the new Title I must permit sufficient autonomy for serious long-term implementation of school-wide reforms, rather than disrupting them with inconsistent policies and assessment practices. It should also be extended to higher grade levels and higher order skills, including a serious high school program like the Talent Development High School under development at the Johns Hopkins CRESPAR Center.

In addition, research indicates that funding should be directed toward improving the quality of teaching. This requires new policies and incentives to get good teachers and administrators to work in Title I schools and to stay there. It also means that we must encourage long-term retraining of teachers who will need to implement curricula supported by tests assessing the more complex skills involved in those approaches.

Other research-based recommendations include:

• lowering class size in the early grades;
• using school choice, magnet schools, and other techniques to permit students to transfer from high-poverty, low-achieving schools to more successful schools, as prescribed in the 1994 legislation;
• sustaining long-term commitments to maintaining strong curriculum materials, appropriate assessments, and serious teacher education programs;
• focusing assessments not only on raising average achievement in schools, but also on raising the achievement of each racial and ethnic group;
• conducting more accurate assessments of limited-English-proficient students, and building support for language-development programs into the core of Title I.

The Clinton administration's proposal for Title I emphasizes state accountability systems of uneven quality that are often not focused on program evaluation and might actually do additional harm to low-income children. Strong points of the president's plan include efforts to assure that low-income children have teachers trained in their subject, more teacher training on the job, lower class size, and forcing failing schools to change.

But the president is advocating excessive decentralization and the popular policy against "social promotion," which, according to a recent National Academy of Science (NAS) study, produces even higher dropout rates for low-income students without educational gains. The Education Department has softened this proposal with more emphasis on remediation, but the social promotion idea is spurring counterproductive policy in Congress.

The voucher movement and other attacks on public schools are clear warnings that continued failure of schools serving impoverished kids will not be tolerated by the public. There is now enough information to demonstrate credibly that Title I can actually bring about educational benefits by focusing on successful approaches and implementing them consistently over the long run. But it will not work if we simply graft onto it ideas and proposals that have popular ratings in the opinion polls.