How Can Successful School Improvement Ideas Be Taken to Scale?

In this article in Educational Researcher, Catherine Lewis (Mills College School of Education) asks why a number of good ideas for improving student achievement are not “scaling up” – that is, having an impact beyond a small number of successful classrooms and schools. Other fields (including health care and automobile manufacturing) have brought about major gains in quality by applying “improvement science,” and so have Japanese educators using the Lesson Study model. So why not U.S. schools?

The reason improvement science hasn’t taken hold in K-12 education, says Lewis, is that we’ve been wedded to a completely different approach to spreading school improvement ideas – the experimental science paradigm. In this model, experts design a program, ask teachers to implement it with fidelity, carefully measure outcomes, and then try to get it working in other sites. The problem with the experimental science model is that there’s tremendous variation in classrooms and schools and few “pure” programs work as designed in all situations. But following this model, variations in implementation are problematic – the model depends on implementing an innovation exactly as it was designed. The result: narrow findings and programs that work only under very specific conditions.

By contrast, the improvement science model is bottom up and thrives on local variation. Instead of attempting to implement a model program, it starts with a learning goal – for example, the Common Core standard, getting students to make sense of math problems and persevere in solving them – and tries out different solutions following a PDSA cycle (plan, do, study, act) revolving around three basic questions:

• What are we trying to accomplish?
• How will we know if we have been successful?
• What change can we make that will result in improvement?

Instead of the knowledge being “in” the program, the knowledge is “in” the teachers trying out different ideas. Instead of striving for fidelity to the program, teachers are constantly making adjustments as they see what’s working and what’s not working with their students. In the experimental science model, variation is problematic. In the improvement science model, variation is a source of ideas to continuously improve instruction and arrive at programs that are more likely to be useful in a variety of school settings.

In Japan, says Lewis, 95 percent of schools follow the Lesson Study model, and the same trial-and-error improvement science process is used at the district, university, and national level, with all levels constantly sharing ideas – which explains the significantly higher student achievement of Japanese students. 

Can the improvement science model be used in U.S. schools? Some researchers are skeptical about importing ideas that have been successful in different cultures, but Lewis believes the basic idea, which has been successful in other enterprises in the U.S., can be used in K-12 schools. The key is shifting from the experimental science model and adopting the far more decentralized, bottom-up, adaptive, improvement science model of change – and then sharing successful practices across schools, districts, states, teacher-training programs, and researchers.

“What Is Improvement Science? Do We Need It in Education?” by Catherine Lewis in Educational Researcher, January/February 2015 (Vol. 44, #1, p. 54-61), available for purchase at http://edr.sagepub.com/content/44/1/54.abstract; Lewis can be reached at clewis@mills.edu.

From the Marshall Memo #577