The main purpose of this study was to determine the predictive value of course and instructor characteristics which have not been linked to student evaluations yet or expected to show a different pattern in Turkish higher education context for students' perception of instructors' effectiveness. These included class size, academic rank, instructors' workload and total number of students, gender and students' disciplines. According to the regression analysis, class size, instructors' total number of students, work load, year of experience and students' disciplines were predictive of students' perceptions of instructors' effectiveness. The higher workload, experience, class size or total number of students the instructors have, the lower ratings they had. However, these effects were not consistent across disciplines and instructors' gender. That is, the better understanding of the predictive value of course and instructor characteristics for student evaluations requires more detailed studies wherein teaching context is taken into consideration.