The object of this research was to establish a reliable test battery to determine probable future success for those persons who wish to be trained as expert classroom observers. The focus of this investigation was to predict coding performance. This report presents the findings of a replication of an earlier study that indicated that coder performance could be predicted from measures of training performance and coder attitudes toward classrooms. These findings were partially confirmed. Included in this report are results of tests conducted to assess accuracy in coding and reliability of observation. Also included are samples of a self-concept test, and two Likert-type questionnaires on attitudes about classroom behavior and beliefs and attitudes about teaching and the teaching process. (JD)