Elementary and secondary school teachers were trained in the use of two observation instruments for teacher evaluation--the Classroom Procedures Assessment Instrument and the Interpersonal Skills Assessment Instrument. Training was conducted in workshops which used a series of ten videocassette sequences showing teachers working with students. Accompanying printed materials presented questions and answers which explained the meaning of specific indicators on the two instruments; these questions helped the trainee rate the videocassette teachers' behavior. Groups viewed the videocassette and then participated in either: (1) oral discussion with feedback from an instructor; (2) directed discussion with an instructor; (3) self-tutorial using a manual, without discussion; or (4) no treatment. The raters were also categorized according to the grade level they taught (K-5 or 6-12). Their attitude toward the training they received and their performance in rating teacher behavior in a 30-45 minute videocassette were studied. Analysis of variance indicated that type of training did affect performance in rating; use of written responses with either oral discussion or printed feedback was more successful than oral discussion only. Practice and rating in one's own grade level diminished these effects. No significant attitudinal differences were found. (GDC)