Comparisons were made of qualitative and quantitative information collected in nine second grade reading classes in a study that examined the respective strengths and weaknesses of using such information to evaluate teacher behavior. Data that were collected included trained ethnographers' narrative observations of teacher behavior (the qualitative information) and the codings of four "frames" of classroom teacher-student interaction per hour of observation (the quantitative information). When comparisons were based on summary profiles of the two types of observations, there was overall agreement between the two data sources, especially with regard to the teachers' tendency to deal with different units of students (individuals, small groups, and whole classrooms) and how much observed time was spent on academic versus behavioral matters. Slight elevations in such affective behaviors as criticism or punishment were found to affect the narrative data much more heavily than the quantitative data. In general, when the teacher's behavior in the classroom was consistent, the qualitative and quantitative profiles showed good agreement. When the pattern of teacher behavior was less clear, the narrative data provided invaluable information regarding the presence of unusual classroom activities or the absence of expected ones. (RL)