Association for Public Policy Analysis & Management

Policymakers have called for school-based professional learning communities and organizational structures that promote regular opportunities for teachers to collaborate with colleagues. These initiatives assume collaboration enables teachers to strengthen their instruction, thus improving learning outcomes for students. However, researchers are only beginning to understand how teacher collaboration affects student achievement. There is substantial evidence that schools characterized by higher levels of collaboration also have higher levels of student achievement. However, we do not know whether this relationship is causal, nor do we understand the mechanisms by which instructional collaboration impacts students’ learning (e.g., changes in school culture or teaching practice). Furthermore, we do not know whether certain kinds of collaboration (e.g., about assessment or students) may be more beneficial than other kinds.

This study draws upon survey and administrative data on over 9,000 teachers in 336 Miami-Dade County public schools over two years to investigate the kinds of collaborations that exist across the district and whether these collaborations predict students’ achievement. Using survey items about the extensiveness and helpfulness of collaboration about different instructional domains, we constructed a domain-general collaboration quality measure and three domain-specific collaboration quality measures (collaboration about instructional strategies, students, and assessments). After reporting on descriptive analyses of the nature of collaboration, we then use multilevel regression modeling approaches to estimate a teacher’s valued added to student achievement (in reading or math) as a function of these collaboration measures and various teacher and school characteristics. Finally, we use a teacher fixed-effects approach to test whether teachers’ returns to experience are greater when working in schools with better collaboration quality.

This paper contributes to the literature on instructional collaboration in at least four ways. First, we provide a descriptive account of the kinds and quality of collaboration that exist across a large, urban district, and we show how collaborations vary for different kinds of teachers and schools. This district-level view—which, to our knowledge, is the first of its kind—reveals which instructional domains (e.g., assessment, teaching strategies) receive more and less attention as well as which kinds of collaboration teachers perceive to be more and less helpful. Second, we test whether certain kinds of collaboration are more strongly related to student achievement than others. We find that high quality collaboration that addresses a variety of instructional domains, rather than domain-specific, is most often predictive of value-added across math and reading. Third, we investigate possible mechanisms by which collaboration influences student achievement. We make progress in differentiating the effects of working in a more collaborative school from being a more collaborative individual, finding evidence that both contribute to achievement gains. Finally, we provide some of the strongest evidence to date that collaboration improves student achievement by controlling for extensive school and teacher characteristics and by testing various threats to a causal interpretation of estimates. Of particular note, we find that a teacher improves at a faster rate when working in schools with better collaboration quality as compared to that same teacher in a school with worse collaboration quality.