Two experimental studies of attribution for a hypothetical accident involving a male pedestrian being knocked down by a car while crossing a street used the same 2 (subject's sex) x 2 (sober vs. drunk stimulus person) x 2 (proper vs. improper behavior of stimulus person) x 2 (mild vs. severe outcome) factorial design, the dependent measure being fault attributed to the victim in the first experiment and perceived likelihood of the accident in the second. The subjects were undergraduate students of the University of Bombay. Greater attributed fault and higher perceived likelihood were obtained in the improper rather than proper behavior condition. Greater fault was attributed to the drunk victim as compared to the sober victim only by female subjects and higher likelihood of accident was perceived for the drunk victim as compared to the sober victim only in the improper behavior condition. The results suggest the greater relative importance of an immediate cause (behavioral impropriety) vis-a-vis a prior cause (insobriety).