Student opinion surveys are important but widely misunderstood tools for evaluating teaching effectiveness. In this brief review, an analogy is drawn between the use and interpretation of observational data for public health and biomedical research and the use of student opinion data in evaluating teach ing effectiveness. Sources of systematic error in the form of selection bias, information bias, and confounding are defined and illustrated. Original data concerning intermittent "quid pro quo" confounding (i.e., the effect of expected grades on student evaluations of teaching) are presented. Finally, the principle of faute de mieux ("lack of anything better") and the interpretation of less-than-pristine data are considered.
B. Burt Gerstman. "Student evaluations of teaching effectiveness: the interpretation of observational data and the principle of faute de mieux" Excellence in College Teaching (1995): 115-124.