Huffpost Impact

Featuring fresh takes and real-time analysis from HuffPost's signature lineup of contributors

Robert E. Slavin Headshot

Tailoring Evidence-Based Reform to Different Problems

Posted: Updated:


Not long ago, I gave a speech at the American Psychological Association's convention in Honolulu (all right, fighting for evidence-based reform does have its pleasures). Readers of this column will not be surprised at anything I said, but I got one question that provoked some thought. My questioner wanted to know why I kept referring to such easy-to-define-and-measure problems as ensuring that children can read or understand algebra, rather than much more complex problems such as how to lead schools.

I didn't say it at the time, but I think this is both a silly and a profound question. The silly part is its implication that if evidence can't solve all problems, it is of little value. In fact, is there anyone out there who thinks that it is not important to identify effective and replicable approaches to teaching reading, algebra, and all the other relatively easy-to-define, easy-to-measure problems of education?

Yet solving these does still leave some very important but less-well-defined problems that may take different approaches. These approaches should still be informed by evidence, but perhaps different types of evidence from the design-build-evaluate-disseminate model that usually leads to proven and replicable approaches to reading or algebra, if anything does.

Take leadership, which was my questioner's example. I don't think anyone will ever develop and evaluate a "principal protocol" for all schools, but it's easy to imagine many innovations that could help principals be more effective, if they turned out to work in well-designed evaluations. If you break the principal's role into its components, this is easy to see. For example, principals play a key role in teacher evaluation, and if anyone designs a teacher evaluation strategy that improves teacher performance overall, this should improve students' performance, and that is easy to measure. Principals can play a key role in such schoolwide issues as attendance and behavior problems, and solutions for these exist and can readily be implemented and measured.

Principals play a leading role in managing resources, and the impact of each resource can have its own evidence base. Given a certain level of discretionary funding, should principals hire classroom aides, reduce class size, adopt particular programs, implement after-school programs, or purchase playground equipment? There is already evidence on most of these; the principal should be aware of this evidence and take it into account, and more evidence of this kind and better dissemination would be helpful.

Principals should be able to collaborate with staff to set goals, and then motivate and enable the staff to achieve the goals and monitor progress toward doing so. This is less cut-and-dried, but professional development for goal-setting and continuous progress toward targets is available, and the use of specific professional development models can be evaluated and replicated.

The kinds of leadership skills I suspect my questioner was thinking about are perhaps harder to measure and harder to influence. For example, positive relationships with staff, students, parents, and community leaders. Or the ability to make good decisions under pressure. Or the ability to communicate enthusiasm and high expectations for learning, or to serve as a positive role model and moral exemplar. Principals are probably more likely to learn these skills from observing their own principals and through other general life experience, but I would never rule out the possibility that professional development and coaching could build these skills as well. Research might also focus on identifying and selecting extraordinary leaders and keeping them on the job and growing in wisdom and capability over time.

Of course, if principals and school staffs chose and effectively implemented proven classroom programs, proven attendance and behavior programs, proven programs for struggling students, proven parent involvement programs, and so on, then the job of being principal would be a lot easier, and a lot more fulfilling. So rather than worrying about areas in which develop-evaluate-disseminate models don't directly apply, it might be a good idea to use what we already know how to do, expand the range of proven approaches, and establish incentives and supports to create a school culture that respects and seeks proven solutions.

This may not solve every problem faced by every principal, but it would be a heck of a start, and then we could use different research and development methods to solve the remaining problems.

NOTE: You can obtain my APA address at I'm sorry you couldn't be in Hawaii to hear it!