By considering such traditional features as validity and reliability, as well as utility, and political and ethical considerations, this paper attempts to establish that student ratings are not credible as sources of information about teacher effectiveness. Several of the most common problems in the related literature are outlined and include how ratings are related to grades awarded; the relationship between course content, or major areas, and student ratings; agreement between student raters and other raters; the relationship of student learning and teacher rating; the halo effect on teacher rating; the influence of course or class level and class size on instructor ratings; reliability of teacher rating; and the politics of evaluation. If sense is to be made of student rating instruments, the author suggests a threefold approach. First, he proposes a moratorium on student ratings as evaluative measures. Second, existing instruments need to be refined until they satisfy minimal criteria of objectivity, reliability, sensitivity, validity, and utility. Finally, he proposes that student ratings be studied in depth. Their use, for the time being, must be experimental and not evaluative. Suggested research directions are proposed. (RC)