Teacher-generated and student-generated variability in teacher effectiveness ratings

The effects of teacher-generated and student-generated variability on student ratings of instructors were assessed using multiple regression procedures. Student variables included interest in the subject, reason for course enrollment, and educational level. Teacher variables were generated through dummy coding. Five different samples were studied: 24 classes taught by different instructors were divided into three samples of eight classes each, and 10 classes taught by the same five instructors were divided into two samples. Teacher-generated variability explained more of the variance in student ratings than did student-generated variability.