The questionnaires or course evaluation forms (CEF) are routinely administered to university students to obtain information about the quality of courses and instructor effectiveness. There is an extensive variety of formats but many use a combination of Likert-type items and open-ended questions. Percentages, but more often means and standard deviations, are used to report student response to CEFs. A review of the literature reveals that CEFs may not elicit accurate information from students concerning course quality because there are a number of extraneous variables that influence student responses. There is some question about the accuracy of information obtained by treating nominal scale data with statistical techniques designed for interval or ratio scale data. Some schools of social work are administering CEFs of unknown reliability and validity. A procedure is presented for developing a functional CEF, and suggestions are offered for determining whether differences in student ratings of course quality are significant. It is concluded that defensible and usable data will result only from CEFs that possess demonstrable reliability and validity, and the CEFs of value must yield data that faculty members will understand, accept, and apply to course and curriculum development. (SW)