The purpose of this study was to identify factors that contribute to the differences in teachers' and principals' perceptions of the principal's leader behavior. Data were collected from a systematically selected sample of principals in the 19,046 public high schools in the database of a commercial school mailing list company. Survey packets were sent to the principals. One hundred sixteen survey packets were returned with 106 being usable.

Principals completed Part I of a questionnaire designed to measure the principal's perceptions of his or her own leader behaviors. Each principal was asked to purposely select six teachers"one in English, math, science, social studies, vocational education, and special education"and have them complete a questionnaire to measure the teachers' perceptions of the principal's leader behavior. The criterion variables were determined by subtracting the mean teachers' response from the principal's response on each item for each school. The mean teachers' responses on Part II of the questionnaire served as measures of the predictor variables.

Principal components analyses were conducted to reduce the data and create meaningful scales. The data were then statistically treated three different ways: (1) by identifying the criterion variables using the difference scores, (2) by identifying the criterion variables using the difference scores when principals' responses only from the questionnaires were used, and (3) by identifying the criterion variables using the difference scores when teachers' responses only from the questionnaires were used. Predictor variables were the principal's modeling of ideal behavior, the principal's skill in teacher evaluation, the teachers' overall awareness of the school, the teachers' perceptions of discipline procedures, and the demographic categories of principal's and teachers' experience, principal's gender, school size, and school type.

Multiple regression analyses were used to determine relationships between the criterion and predictor variables. Principals' modeling of ideal behavior was a significant predictor of every criterion variable in every model. Principals' skill in teacher evaluation was a significant predictor of the principal's quality of communication in every model. School awareness, discipline procedures, and demographic variables were not significant predictors of the criterion variables.