The purpose of this study was to identify effective and ineffective strategies for teaching Title I students. The data collection took the form of direct observations of teacher-student interactions and classroom activities in the classes of 79 Oklahoma City Public Schools Title I teachers. The data on achievement gains came from the California Achievement Test administered in May 1981 and May 1982. The results for reading indicated that, across all grade-levels, the teachers who allocated more time for academic activities and who interacted more with their students tended to have students with higher achievement gains. The teachers whose students had higher gains also tended to rely more on one-to-one interactions, used some challenge in the lesson content, used opinion questions and "higher order" questions, and used feedback strategies that sustained the interaction. Although some of the results for math agreed with the reading findings, there were also some notable differences. In general, the profile of the effective math teacher was more characterized by public interactions and by faster-paced activity than in reading. Small group activity was positively associated with gain, and, as in reading, sustaining feedback was a better strategy than redirecting a question to another child. (BW)