Students are asked on a regular basis to evaluate the lectures and seminars that they are attending. Until recently this has been done by paper and pencil surveys in a classroom setting. In recent years some systems became available to approach student using web surveys. Since this methodology has various advantages (in terms of cost and time, data protection), some methodological concerns have been raised:
(1) Over-coverage and under-coverage: In order to be included into the study, all students have to be registered with a valid e-mail address. Since almost every student has an e-mail address nowadays this no longer seems to be an issue. However, there is a small population who abandons the Internet or are not willing to provide their e-mail address. These populations cannot be reached using web surveys. By contrast, using web survey methodology we can address also students who abandoned the class or course in the course of the semester. As a result, the two survey modes address overlapping, however, slightly different populations which has consequences for the results.
(2) Non-response: Experiences from several waves of online student feedback web surveys suggests that nonresponse is of great concern. Compared to a class room setting where social pressure leads to almost 100 percent cooperation, students feel less obliged to participate in web surveys. Several strategies in reducing non-response for the web survey condition were tested.
(3) Measurement: Completing a paper form in class in the presence of the teacher implies a completely different social setting compared to responding to a web survey at home, in the dorm or at some other place. Thus, concerns were raised, that measurement bias related to the mode of administration should be of great importance. However, experimental research (split half design) suggests, that only minor differences of the data occur. The paper summarizes this line of methodological research for student feedback and course evaluation web surveys based on a series of four studies conducted over the past several years.