Abstract
Analysis of responses from end-of-course evaluations of 24 graduate statistics classes taught by the author on a 26-item evaluation form yielded highly reliable evaluations for both classes (.93) and items (.96). Reliable differences were found for six consecutive stages of four classes each with strong positive trend and weak but significant cubic trend apparent. Items accounted for .423 of the variance with stage explaining .166 and interaction of stage and item .085. A stable hierarchy of item characteristics indicated consensus with mean rating correlating —.76 with standard deviation. Most favorable ratings were given on instructor-student relations, motivation-stimulation, reasonable work load and tests, and clearness of grading procedures, with least favorable ratings on subject organization and competence.
Get full access to this article
View all access options for this article.