Abstract
BCD, the borderline between comfort and discomfort, is a useful tool in studying subjective reactions to discomfort glare. This paper takes data obtained when inexperience on servers made BCD judgments of luminaire combinations in a simulated office and applies an analysis of variance to the data. The data in log form are reasonable distributed, particularly for the situations most closely resembling the lighting in an actual office. Based on the results of the analysis of variance, recommendations are made for improving the design of experiments in which inexperienced observers make BCD evaluations of discomfort glare.