Scaling Issues in Faculty Evaluations

Abstract

148 students were asked to evaluate particular survey questions with scales that varied in number of scale points, type of labeling, and labeling of anchors only vs all scale points. No significant difference was found between variations on 4-and 5-point scales, so these data were collapsed. Analysis indicated that students generated scale-point exemplars significantly better for 4-point scales (96.7%) than for 5-point (87.8%) and 10-point (79.3%) scales. These results are discussed with respect to the administration of student evaluations and the care required in selecting a scale type and interpreting the subsequent outcomes.