The Center for Analysis of Longitudinal Data in Education Research (CALDER) is a project that conducts research on a variety of questions, including factors that affect teachers’ effectiveness. CALDER is a collaboaration among researchers at the Urban Institute’s Education Policy Center, Duke University, Stanford University, the University of Florida, the University of Missouri, the University of Texas at Dallas, and the University of Washington. Funded by the US Department of Education’s Institute for Education Sciences
CALDER Research Question 3: Teacher Effectiveness
What factors account for teacher effectiveness in terms of student achievement outcomes? To what extent, and in what ways, does the importance of these factors vary for the performance of different types of students?
For three of our states (North Carolina, Texas, and Florida) and for New York City we have data that allow us to investigate the effect of a wide range of teacher characteristics on student performance. The particular factors vary somewhat from dataset to dataset, but we typically have information on teacher education level, degree, undergraduate institution, and teacher test performance in addition to demographic characteristics.
In some datasets we additionally have teacher coursework in college, the characteristics of the teacher training program they completed, their hometown, and the amount of professional development they have received on the job.
Because we are able to link teachers with students in some of our databases, we are able to identify characteristics and experiences of teachers that make a difference for student achievement. Perhaps most importantly, we are able to distinguish effects for students with different backgrounds, e.g., limited English, poverty, immigrant status, age level, achievement level, etc.
It’ll be interesting to see how this study progresses. I wonder whether CALDER will identify and employ a way to control for how teachers teach. Without controlling for this factor, I doubt that many of the other factors will affect enough of the variance to account for much in the way of outcomes; that is, I expect the effect sizes (the β weights?) will be small.