Teachers’ Instructional Collaborations Across an Urban District

• In Event: Investigating Policies and Practices That Promote Teacher Quality

Abstract

It is commonly assumed that teacher collaboration benefits both teachers and students. Large-scale, quantitative studies usually examine these relationships at the school-level, providing some evidence that schools with more collaboration have better average achievement. However, schools with more collaboration tend to be better functioning in other ways that could explain these relationships. Little prior research has investigated direct links between collaboration and the performance of individual teachers. Among studies that have, collaboration is typically treated as a unitary and homogenous construct. We do not know whether different types of collaboration are more beneficial than others. Thus, this paper investigates the different kinds of instructional collaborations that teachers experience, how they vary within and across schools, and what impacts they have. It asks: Do collaboration frequency and type vary by teacher and school characteristics? Are teachers who report more frequent or different kinds of collaborations more effective at raising student achievement or more likely to remain in their schools?

This study draws upon survey and administrative data from a large, urban district. In 2011 and 2012, 8,245 teachers completed the survey, including multiple items about collaboration. Using these items, we constructed six factors to signal different features of instructional collaboration, including measures of frequency, extensiveness, helpfulness, and content (e.g., collaboration about students, instruction, and assessment). From administrative data we constructed value-added scores (VAM) for mathematics and English/language arts teachers.
Using linear regression, we estimated a teacher’s collaboration score as a function of teacher and school characteristics. We also modeled a teacher’s VAM as a function of her collaboration scores, controlling for teacher and school characteristics. Retention analyses used discrete-time hazard models with the same sets of predictors.

Teachers who are female, non-White (especially Black), and without advanced/ specialized degrees reported more frequent, extensive, and helpful collaborations across instructional categories (assessment, students, instruction). As compared to secondary teachers, elementary teachers indicated more frequent, extensive, and helpful instructional collaborations across measures, with one exception--they reported significantly less extensive/helpful collaboration about assessment.

In general, teachers who reported more frequent, extensive, and helpful collaborations were no more or less effective at raising students’ achievement, nor more or less likely to remain in the same school. A notable exception was that teachers who reported more extensive and helpful collaborations about assessment had significantly better math VAM. Additionally, we found differences by school performance level. More extensive and helpful collaborations predicted better VAM in higher-performing schools and better retention in lower-performing schools.

This study offers the only large-scale, teacher-level evidence, of which we are aware, for relationships between different types of instructional collaborations and teachers’ effectiveness. Though it is often assumed that collaboration helps teachers improve their performance, our preliminary results suggest few significant relationships. Only collaboration around assessment predicted math VAM, providing some empirical support for reform efforts that intend to improve teacher quality by promoting collaboration around assessment. Our findings also suggest that policymakers aiming to use collaboration to improve instruction or retention should consider how its effects may vary by school context.

Authors

• Matthew Ronfeldt, University of Michigan

• Susanna Farmer, University of Michigan

• Susanna Loeb, Stanford University