Holding teachers accountable: An old-fashioned, dry, and boring perspective

Abstract

Few would disagree with the desirability to hold teachers accountable, but student evaluations of teaching and department head evaluations of teaching fail to do the job validly. Although this may be due, in part, to difficulties conceptualizing teaching effectiveness and student learning, it also is due to insufficient attention to measurement reliability. Measurement reliability sets an upper bound on measurement validity, thereby guaranteeing that unreliable measures of teaching effectiveness are invalid too. In turn, for measures of teaching effectiveness to be reliable, the items in the measure must correlate well with each other, there must be many items, or both. Unfortunately, at most universities, those who are tasked with teaching assessment do not understand the basics of psychometrics, thereby rendering their assessments of teachers invalid. To ameliorate unsatisfactory assessment procedures, the present article addresses the relationship between reliability and validity, some requirements of reliable and valid measures, and the psychometric implications for current teaching assessment practices.