Both scholars and professional standards identify teachers' knowledge of students as important to promoting effective instruction and student learning. Prior research investigates two such types of knowledge: knowledge of student thinking and teacher accuracy in predicting student performance on cognitive assessments. However, the field presents weak evidence regarding whether these constructs are amenable to accurate measurement and whether such knowledge relates to instruction and student outcomes. Without this evidence, it is difficult to assess the relevance of this form of teacher knowledge. In this article, evidence from 284 teachers suggests that accuracy can be adequately measured and relates to instruction and student outcomes. Knowledge of student misconceptions proved more difficult to measure, yet still predicted student outcomes in one model.