| Schools face a great challenge in recruiting and retaining quality teachers, given the documented importance of, variability in, and difficulty observing and predicting teacher quality. One option schools have is to identify what more effective teachers do and use that information to train less effective teachers to get better. Unfortunately, there is little empirical support for much traditional teacher training, as measured by gains in student test scores. Models of collaborative, team-based learning---such as Professional Learning Communities and Japanese lesson study---have been widely touted, and there is some evidence that they may be effective in certain contexts. Economic theory suggests this could be because of peer monitoring, peer pressure, specialization, knowledge-sharing, or market failure in pre-service training, particularly if learning to teach is primarily experiential. However, not all collaboration is good due to concerns about free-riding and substituting for more productive individual activity, so unbridled enthusiasm for collaborative professional development may need to be tempered.;This dissertation examines the effectiveness of a specific form of teacher collaboration in the form of inquiry teams, groups of teachers and administrators jointly engaged in action research projects with the aim of uncovering innovative instructional strategies and sharing effective approaches. It takes advantage of the phase-in of teams, eventually to all teachers in a large, urban school district in the northeastern United States from 2007--2010 to estimate the results of three natural experiments using difference-in-differences and instrumental variables approaches. The effects of teamwork on teacher value-added, teacher retention, and student test scores are small and sensitive to year, specification, and outcome, although results are mostly positive and occasionally statistically significant, suggesting that overall effects are potentially positive but modest at best. Further examination of heterogeneity and four qualitative case studies of teams suggest that small average effects mask considerable differences in team processes, and that under certain conditions, inquiry team work may be far more effective. A cost analysis reveals that, although it is costly to do inquiry work well, given the low-intensity of average treatment and the large number of students affected, the benefits of inquiry work could exceed the costs if the policy were more targeted. Overall, the policy recommendation is to temper unqualified enthusiasm about teacher collaboration, as without appropriate structures and supports it has little measurable effect on the outcomes examined here. As a policy lever, a universal mandate to participate on collaborative inquiry teams is unlikely to be effective or pass a cost-benefit test. Nonetheless, smaller-scale, higher intensity forms of collaboration that allow for more active leadership support and participation may be more promising, and more cost-effective than alternative forms of professional development, particularly for some sub-groups of teachers such as those in their first year of teaching. |
