Teacher Evaluations

Only available on StudyMode
  • Download(s) : 106
  • Published : December 4, 2007
Open Document
Text Preview
Literature Review

It is well known that students' rating of instructors is perhaps the most widely used method of assessing instructor effectiveness (Centra, 1993, 2003). Many issues concerning validity of student evaluations have been examined including bias (Marsh, 1984), which was found to be minimally present based on the grade the student expected to receive (Centra, 1993, 2003), and student characteristics (e.g., grade point average, academic ability, gender, age); which some studies found did not affect student evaluations (Centra & Creech, 1976; McKeachie, 1979; Centra, 1993, 2003), and some studies found did affect student evaluations (Basow & Howe, 1987). Since students are the recipients and raters of instruction, they can provide an important, unique, and necessary perspective on judging teacher effectiveness. Student evaluations are commonly used to make decisions regarding faculty salary, awards, and promotion and tenure; therefore, it is imperative to consider all the major variables that can influence student evaluations. Although there has been extensive research on student ratings of instruction, the research is only now focusing on those who partake in the evaluation process: faculty, students, and administrators (Chen & Hoshower, 2003). Specifically, when attempting to assess teaching quality through student evaluations, it is important to be aware of the specific behaviors exhibited by students as they evaluate faculty. One area of student behavior that has been examined is rating errors, which are inadequacies of one sort or another in performance appraisals and the most common rating errors are halo effect, leniency or severity error (Centra, 1993), and central tendency; which have been examined extensively in the past in instructional evaluations. Halo effect occurs when a rater has the tendency to think of a person as being generally good or inferior, which would yield relatively high inter-correlations between different dimensions...
tracking img