Fifty-eight undergraduates took part in an experiment evaluating a full-spectrum light closely approximating natural sunlight and a cool-white fluorescent light commonly used in institutional settings. Four classes of a single teaching assistant were used and each class was compared against itself under different lighting conditions. Data were gathered during each seminar (with videotapes) and after each seminar (with questionnaires). The results show that there was no differential effect in self-reported attitudes or videotaped behavior attributable to the respective lighting conditions. At the end of the experiment students were asked to judge the respective lights on 20 dimensions and express their overall preference for one light or the other. No significant overall preference was shown; however, there was a clear difference in the qualitative characteristics of the lights when direct observations and appraisals were made. (Author/MLF)