Teachers and Student Achievement in the Chicago Public High Schools

Using unique administrative data on Chicago public high school students and their teachers, we are able to estimate the importance of teachers on student mathematical achievement. We find that teachers are educationally and statistically important. To be sure, sampling variation and other measurement issues can strongly influence estimates of teacher effects, and, in some cases, account for much of the dispersion in teacher quality. Even after correcting for these problems, we find that one semester with a teacher rated two standard deviations higher in quality could add 0.3 to 0.5 grade equivalents, or 25 to 45 percent of an average school year, to a student's math score performance. Additionally, our teacher quality ratings remain relatively stable for an individual instructor over time, are reasonably impervious to controlling for non-math teachers, and do not appear to be driven by classroom sorting or selective reporting of test scores. After relating our measured teacher effects to the standard observable characteristics of the instructor, we find that traditional human capital and demographic measures, including those used for compensation purposes, explain little of the total variation in teacher quality.