Harris M. Cooper

Professor and Director of Education
Duke University

Thanks for the opportunity to cogitate on the issue of educational research and its impacts. It is an extremely important topic but thinking about it also hurts my head. 

Much of the pain arises from our own standards of evidence. If we were to apply these to demonstrating the impact of educational research we would, in the first instance, have to devise an experiment in which, for example, randomly selected schools or communities were exposed to research outcomes while others were kept in the dark. Then, we would measure the impacts on policies, practices, and student educational attainment. I know of no study with these characteristics. 

We are left, then, with a search for ex post facto case studies, and we know these are generally unconvincing. Most notably, the time lag between when research appears and when it might have an impact on policies, practices, and attainments can be short or it can take years. This is because the existence of a new research finding is rarely sufficient to instigate change. Rather, research becomes relevant to an issue in policy or practice instigated by other social forces. It can be part of a cluster of causes that move policy and practice. 

Also, educational research can become part of a decision-making process without being a causal factor. For example, when a Dover, PA , school board member (recently retired) cites research that supports intelligent design and the board subsequently votes to put a warning label on the teaching of evolution, is this a case of research having an impact on education policy? Here, I think research is used but the causal direction is reversed from that which you seek; a policy position caused the use of research (with supporting outcomes) rather than research causing a policy position. Which was the case in the judge’s decision to rule that the board policy was unconstitutional? 

A final problem involves the generally decentralized nature of education policy making and adoption of schooling practices. Often, change in education is a ground up exercise, occurring in classrooms, schools, and school districts, rather than at the state or national level. When most people seek evidence about the impact of research they want to see the big effects, not the small, hard-to-document ones that happen in individual classrooms, schools and communities. 

Now that I’ve anesthetized myself, I will make three nominations. 

1. Class size

I suspect you will get several nominations involving the Tennessee STAR study. Others will be more knowledgeable about this than me, but I suspect there are pretty direct links between the results of this study and education reform initiatives in Tennessee and California. 

2. Cooperative learning

This is more in the bottom-up type of impact I mentioned above. A large number of teachers and schools use formal and informal student learning groups. I believe if you spoke with David Johnson at the University of Minnesota he would tell you that he started with a basic social psychological theory of group behavior, conducted laboratory and field studies testing models, developed research-based pedagogical strategies, and then packaged these for teachers. 

3. Homework

Finally, I will nominate one line of my own research. My first meta-analysis of homework was completed in 1989. Subsequently, I translated this into a brief monograph for policy makers and practitioners (titled The Battle over Homework: Common Ground for Administrators, Teachers, and Parents). The results of the meta-analysis related to how much time students at different grade levels should spend each night on homework. They supported a set of guidelines jointly recommended by the NEA and the National PTA. Publicity surrounding my findings led several school districts to revise their homework policies. This is also an example of a bottom-up impact, so there are still numerous districts that “ignore” the research. 

What I learned by developing my response to your request was that the first step needed in assessing research impacts is to carefully define and catalog the many different ways research might influence policy, practice, and educational attainment. This would be a fun exercise. Then, the compelling examples -- the documented case studies -- would come easier. 

I hope this helps.