The Effects of Biasing Variables on Students' Ratings of Courses [Zum Einfluss von Biasvariablen auf die Bewertung universitaerer Lehre durch Studierende]

Author(s)

Christiane Spiel, Martin Gössler

Abstract

The present study deals with a question that is often controversely discussed: Do student ratings reflect a confound of (biasing) background variables, such as characteristics of teachers or students, with teaching performance? In the study we investigated whether the students' ratings of a course can be predicted by biasing variables. In addition, we investigated whether there are differences in the amount of variance explained depending on discipline and the type of course (lecture or seminar) and reason for participation (required courses versus elective). In the Summer semesters of 1994 and 1995, data were collected from 756 students from 66 courses at the University of Vienna. Courses were selected based on a stratified research design. Results suggested that the amount of variance that can be explained by biasing variables varying across data. In particular, the amount of variability explained for elective seminars is very small. In agreement with previous research results indicated that only students' interest is substantially correlated with their ratings. However, to decide whether ratings are biased by interest an additional data collection (before course participation) is needed. In any case, students' interest should be statistically eliminated when evaluating courses.

Organisation(s)

Department of Social and Cultural Anthropology

Journal

Zeitschrift für Pädagogische Psychologie

Volume

14

Pages

38-47

No. of pages

10

ISSN

1010-0652

Publication date

2000

Peer reviewed

Yes

Austrian Fields of Science 2012

5010 Psychology

Portal url

https://ucrisportal.univie.ac.at/en/publications/575e51bc-e09e-479d-9d90-a7a4017d6ac6