Student Self-Assessment in Higher Education: A Meta-Analysis

Abstract

Quantitative self-assessment studies that compared self- and teacher marks were subjected to a meta-analysis. Predictions stemming from the results of an earlier critical review of the literature (Boud & Falchikov, 1989) were tested, and salient variables were identified. Factors that seem to be important with regard to the closeness of correspondence between self- and teacher marks were found to include the following: the quality of design of the study (with better designed studies having closer correspondence between student and teacher than poorly designed ones); the level of the course of which the assessment was a part (with students in advanced courses appearing to be more accurate assessors than those in introductory courses); and the broad area of study (with studies within the area of science appearing to produce more accurate self-assessment generally than did those from other areas of study). Results of the analysis are discussed and differences signaled by the results of the three common metrics examined. The distinction between relative and absolute judgment of performance is drawn. It is recommended that researchers give attention to both good design and to adequate reporting of self-assessment studies.