A pool of items used for student ratings of college instructors was studied by means of two factor analyses. Major objectives were to confirm the existence of the major dimensions of student ratings as identified in previous research literature, and to determine whether categories or levels of item specificity could be identified. As to the first intent, the dimensions found in the factor analysis of the Form A questionnaire (Course Management, Communication Effectiveness, Outcomes, Difficulty/Workload, Warmth, Encouragement/Stimulation and Assignments) coincided quite well with previously identified dimensions. These results were substantially confirmed with the analysis of Form B using a Likert scale. The most significant unintended outcome from the study was the results of the hierarchical factor analysis. It was a fairly widely held belief that student ratings are subject to one general factor, but the discovery of two general factors from the hierarchical analysis led us to question this assumption. Regarding the attempt to identify levels of specificity, results of the hierarchical analysis permitted empirical classification of items into at least five categories. Logical analysis of these classifications indicated that they are related to item specificity. (Author/CTM)