Improving Teaching Through Collaboration

Our guest author today is Matthew Ronfeldt, Assistant Professor at the University of Michigan School of Education. Ronfeldt seeks to understand how to improve teaching quality, particularly in schools and districts that serve historically marginalized student populations. His research sits at the intersection of educational practice and policy and focuses on teacher preparation, teacher retention, teacher induction, and the assessment of teachers and preparation programs.

Learning to teach is an ongoing process. To be successful, then, schools must promote not only student learning but also teacher learning across their careers.* Embracing this notion, policymakers have called for the creation of school-based professional learning communities, including organizational structures that promote regular opportunities for teachers to collaborate with teams of colleagues** – also here and here. As the use of instructional teams becomes increasingly common, it is important to examine whether and how collaboration actually improves teaching and learning. The growing evidence, summarized below, suggests that it does. 

For many decades, educational scholars have conducted qualitative case studies documenting the nature of collaboration among particular groups of teachers working together in departmental teams, reading groups, and other types of instructional teams. This body of work has demonstrated that the kinds and content of collaboration vary substantially across contexts, has shed light on the norms and structures that promote more promising collaboration, and has set the stage for today’s policy focus on “professional learning communities.” However, these studies rarely connected collaboration to teachers’ classroom performance. Thus, they provided little information on whether teachers actually got better at teaching as a result of their participation in collaboration.

More recently, a number of large-scale studies have looked across many schools to investigate whether teacher collaboration specifically improves teaching and learning – see here for a review. Goddard et al. (2007) found that elementary schools in which teachers reported more extensive collaboration on surveys also had better student achievement, even after controlling for a set of student and school characteristics. In a follow-up study, Goddard and colleagues (2010) similarly found a direct relationship between collaboration and achievement and an indirect relationship, mediated by teacher collaboration, between principal leadership and achievement.

Though these correlational studies provide initial, suggestive evidence that teacher collaboration causes student achievement to improve, other explanations are also possible. First, unobserved factors could explain observed relationships; for example, schools that are better at retaining teachers may be likely to have both better achievement and better collaboration. Second, more collaborative teachers might non-randomly sort into higher achieving schools. Finally, it is possible that stronger achievement causes teachers to collaborate rather than the other way around. To rule out these alternative explanations, experimental and quasi-experimental research is needed.

A pair of recent, quasi-experimental studies provide credibly causal evidence that supporting instructional teams to engage in inquiry around student data increases student achievement – see Gallimore et al (2009) and Saunders, Goldenberg, and Gallimore (2009). The researchers designed a school-level intervention that trained instructional leaders to promote frequent teacher collaboration based upon an inquiry-focused protocol. Treatment schools showed substantially greater achievement gains than control schools. These studies begin to build the case that collaboration causes instructional effectiveness to improve, but it is difficult to ascertain whether collaboration specifically, other aspects of the intensive intervention (e.g., trained instructional leaders, structured protocols), or both caused the observed improvement. Even if collaboration were responsible, finding such carefully orchestrated collaboration to spur improvement does not necessarily mean that more typical forms of collaboration are equally beneficial.

My colleagues and I investigated the various, naturally occurring forms of collaboration that exist among teachers in instructional teams across Miami-Dade County Public Schools, one of the largest, urban districts in the U.S. – see Ronfeldt et al (2015). Our goal was to better understand the landscape of the more typical forms of collaboration that exists across a district and its relationship to student achievement and teachers’ effectiveness at raising achievement, as signaled by value-added measures. We collected over 9,000 surveys across two years, including 18 questions about the extensiveness and helpfulness of teachers’ collaboration in different instructional domains (e.g., discussing specific student needs, state test results). Using these items, we constructed a general measure for the quality of collaboration across all instructional domains, as well as a set of domain-specific measures for collaboration – about students, instruction and assessment – to investigate whether collaboration about specific instructional topics might be especially beneficial. We then linked collaboration measures to value-added measures of school and teacher effectiveness at raising math and reading achievement.

Consistent with prior research, we found that schools in which teachers reported better quality collaboration – regardless of the content of the collaboration – had better average achievement gains in math and reading. Assuming that collaboration quality is actually causing schools to be more effective at raising achievement, how did it do so? The most likely explanation is through improving the quality of instruction among the teachers participating in this collaboration. Were this the case, then we would expect: (1) teachers who report engaging in better quality collaboration to be more effective than peers who report engaging in worse quality collaborations, and (2) teachers to improve at faster rates when working in schools with stronger collaboration. We found evidence on both fronts. Students of teachers who reported experiencing better quality collaboration across a wide range of instructional topics (domain-general), as well as specifically about assessment, had better achievement gains in math than students of teachers who reported engaging in worse quality collaboration in these areas; students of teachers who reported better quality collaboration about instruction experienced achievement gains in reading. Additionally, teachers’ effectiveness at raising math achievement increased significantly more each year when employed in schools with better quality collaboration about multiple instructional domains (domain-general) and about assessment than when employed in schools with worse collaboration in these areas.***

These findings suggest that collaborative schools can function as organizations for teacher learning. If so, we would expect the same kinds of schools to make ideal settings for training prospective teachers during initial preparation, when they are just beginning to learn to teach. In a separate study, I tested whether teachers who learned to teach in field placement schools with better collaboration quality were more effective after graduating and becoming a full-time teacher of record – see Ronfeldt (2015). The results suggested that they were. Teachers who completed their clinical training in schools with one standard deviation better collaboration quality had 9 percent of a standard deviation better math achievement gains. In other words, compared to their newly hired peers who had completed their pre-service preparation in less collaborative settings, on the first day of class these teachers performed as though they already had about half a year more of full-time teaching experience under their belts. I found no significant differences, though, in terms of reading achievement gains.

If the evidence summarized above isn’t enough to sway skeptics about the benefits of teacher collaboration, then a new randomized control trial in Tennessee should. The intervention consisted of pairing teachers who scored highly on particular dimensions of the state’s observational evaluation rubric with colleagues from the same school who scored lower on those same dimensions. After these teacher pairs were encouraged to work together on instruction for a year, the authors found that the schools assigned to the treatment had meaningfully greater achievement gains than other schools. On average, teachers who participated in the intervention had better performance. These findings offer some of the strongest causal evidence to date that encouraging teachers to work together on instruction directly improves the teaching quality of all involved.

It has become increasingly common for practitioners and policymakers to build opportunities for teachers to collaborate in instructional teams, including the creation of professional learning communities, as a way to promote instructional improvement, combat isolation, and increase retention among teachers. Existing research evidence, which I summarized here, overwhelmingly supports these efforts. Collaboration appears to be good for teachers as well as for the students that they teach.


* See Feiman-Nemser, S. (1983). Learning to teach. In L. S. Shulman & G. Sykes (Eds.), Handbook of Teaching and Policy (pp. 150-170). New York: Longman.

** Carroll, T. G. (2007). Teaching for the future. In B. Wehling & C. Schneider (Eds.), Building a 21st century U.S. education system, (pp. 46-58). Washington, D.C.: National Commission on Teaching and America’s Future.

*** We used a teacher fixed effects approach to compare the rate of improvement of a given teacher working in a school with better collaboration to the rate of improvement of the same teacher in a school with worse collaboration.