It was hypothesized that base rate differences in the number of complaints made during daylight and nighttime hours were responsible for a previously reported, nonlinear relationship between temperature and domestic violence. This hypothesis was tested by subjecting calls for service in 1987 and 1988 in Minneapolis, to moderator-variable regression analyses with controls for time of day, day of the week, season, and their interactions as well as linear trend, major holidays, public school closings, the first day of the month, and other weather variables. Temporal variables explained 75% of the variance in calls for service. As hypothesized, the base rate artifact was responsible for an apparent downturn in violence at high temperatures: Fewer complaints were received during afternoon hours, because they happen to be the warmest time of the day. The results were interpreted in terms of routine activity theory.
