There are three primary measures of teaching performance: student test-based measures (i.e., value added), classroom observations, and student surveys. Although all three types of measures could be biased by unmeasured traits of the students in teachers’ classrooms, prior research has largely focused on the validity of value-added measures. We conduct an experiment involving 66 mathematics teachers in four school districts and test the validity of all three types of measures. Specifically, we test whether a teacher’s performance on each measure under naturally occurring (i.e., non-experimental) settings predicts performance following random assignment of that teacher to a class of students. Combining our results with those from two previous experiments, we provide further evidence that value-added measures are unbiased predictors of teacher performance. In addition, we provide the first evidence that classroom observation scores are unbiased predictors of teacher performance on a rubric measuring the quality of mathematics instruction. Unfortunately, we lack the statistical power to reach any similar conclusions regarding the predictive validity of a teacher’s student survey responses.