Four general problems (two substantive, two methodological) were addressed in a research project: (1) Does the nature of the pupil or the setting make a difference in the teaching style which is most effective? (2) Does the cognitive level of the learning objective make a difference? (3) How can relationships within the classroom be analyzed? and (4) How should composite measures of classroom behavior be formed? Chapter one of this final report presents a brief overview of the project, which tested relationships between teaching style and pupil gain in a first and a fifth grade sample of pupils who differed in sex, race, socioeconomic status, IQ, anxiety, motivation, cognitive style, dependency, locus of control, or pre-test standing. Chapter two is a review of literature on research studies utilizing different methodologies: process-product relationship; trait-treatment interaction; and analysis at different levels of aggregation. The third chapter reports on a reanalysis of data collected from fifth and first grade classrooms in a previous project during 1969-70. Chapter four presents and discusses problems studied in this project: (1) to develop a set of process measures to serve as alternatives to the factor scores which had previously been used; (2) to develop procedures for analysis within classroom relationships; (3) to analyze interactions between characteristics of pupils and the teaching style which is most effective; and (4) to seek differences in the classroom behaviors which are associated with most effective learning for outcomes which differ in cognitive level. The fifth chapter presents a summary of the project and implications of the findings. (JD)