This research focused on predicting short-term variations in police service calls in Minneapolis, Minnesota, based on weather changes and temporal variations.
Data consisted of all nonduplicate police service calls received by the Minneapolis Police Department during 1985, 1987, and 1988. Final data files contained 2,920 records for each year. Weather variables included ambient temperature, wind speed, relative humidity, presence or absence of precipitation, and presence or absence of fog. Temporal variables included time of day, day of week, month, presence of light, major and minor holidays, local festivals, and public high school closings. The first day of the month was also examined, since it has been suggested that certain crimes, particularly assault, robbery, and domestic abuse, increase on the day of welfare checks. Results suggested that police service calls were remarkably predictable over short periods of time using weather and temporal variables. Variables most effective in predicting police service calls appeared to be temporal variables, especially time of day, day of week, month, and presence or absence of light. Police service calls increased during afternoon and evening hours, on weekends, during spring and summer months, and during periods of darkness. The first day of the month and temperature were significant predictors of police service calls, while major and minor holidays, local festivals, and public school closings were not significant predictors. The authors believe findings are compatible with basic premises of routine activity theory. 51 references, 8 notes, and 3 tables