The purpose of this research report was to improve the utilization of classroom observation systems in instruction and research by identifying (a) new and unique dimensions of teaching, and (b) the commonalities of several existing observation systems. The design of this study called for the coding of classroom behaviors by individuals highly trained in the use of the nine observation systems selected for this study. A sample of 50 taped lessons demonstrating classroom performance was obtained, and each coder was instructed to code each lesson. The materials were then returned and analyzed. Factor analysis was used to find the appropriate number of independent dimensions necessary to adequately describe the phenomenon of teacher behavior. The number of variables input into the first factor analysis was summarized in terms of standardized scores for each factor generated. These standardized scores became the input for a second factor analysis. The results of this study indicate an overlap of dimensions among the category systems and raise questions about the continual unsystematic development of observation schemes. (RC)