Student questionnaire data concerning the evaluation of instruction were collected over a 3-day period. Multivariate analyses of variance were used to investigate the statistical analyses of evaluation data. It was found that: departments differed significantly on overall ratings of instruction; class size and level of instruction affected all overall ratings except those of instructor effectiveness; the effect of class size and level of instruction varied dramatically across departments; and ratings of teaching skills were different across departments, levels of instruction, and class sizes (within departments these relationships varied further). The interest/atmosphere factor was the best predictor of overall ratings of instruction. When this relationship was examined in different subgroupings, it was found that lower level students in general placed more value on organization and clarity. It is possible to pinpoint courses which are generally rated lower or higher than others, and this does not seem to be dependent on level: both the highest and the lowest ratings were most often received by level 4 courses. The generation of questions which could be answered from analyses of questionnaire data is also discussed. (Author/PN)