Quality assurance and evaluation research, like other fields of social research and its application, are confronted with a series of problems. In the present paper, I want first to give a list of such problems, although necessarily incomplete. It is then claimed that while there is no "perfect" solution to these problems, critical multiplism may be a set of approaches which might attenuate the problems or at least make them more visible so that one can deal with them. Critical multiplism is a general methodological framework which cannot be fully accounted for here; rather, I want to present some major issues and on their base make some suggestions about how to conduct social research. (Contains 3 footnotes, 1 table, and 3 figures.)