This article describes some of the conceptual and methodological issues that arise when researchers use teacher logs to measure classroom instruction. Data and examples come from the Study of Instructional Improvement, which used teacher logs to study patterns of literacy instruction in schools implementing three comprehensive school reforms. Over the course of this study, more than 75,000 logs were collected from nearly 2,000 teachers in Grades 1 through 5. This article discusses why teacher logs were chosen as the data collection strategy, various psychometric issues associated with their use, and some of the substantive findings that emerged as part of the study. (Contains 2 tables and 4 notes.)