Background: Criminal behavior is a significant public health threat, serving as both cause and consequence of violent actions. Violent assaults constitute over 1 million emergency room visits annually and represent the 8th leading cause of non-fatal injuries. However, the underlying environmental causes of violent behavior are still poorly understood. Our study investigates the role that daily temperature and temperature deviation plays on the risk of crime incidence. We further consider the ongoing debate as to whether temperature affects crime through a temperature-aggression relationship or by modifying routine activities.Methods: We used a hierarchical time series model to estimate the risk of criminal behavior associated with daily temperature and daily departure from normal temperatures in 436 U.S. counties. Departure from normal temperatures denote the deviation of daily temperature from the expected conditions. Daily crime incidence was acquired from the FBI’s National Incidence Based Reporting System and daily weather data from gridMET. We explored seasonality of crime risks and estimated linear and non-linear exposure-response relationships.Results: We found each 10°C increase in daily mean temperature increased the risk of violent crime by 11.92% (95% CI: 11.57, 12.27) and each 10°C increase in departure from normal temperatures increased violent crime risk by 10.37% (95% CI: 10.05, 10.69). The exposure response curve revealed increasing risk with absolute daily temperature until flattening off above 25°C. For departure from normal temperatures, we observed a strong linear relationship with crime, and significantly higher risk during fall and winter seasons. Similar overall effects, but with lower risk, were observed for non-violent property crime.Conclusions: Our results suggest that short-term changes in daily temperature may be associated with greater risk of criminal behavior. Based on the exposure-response curves and seasonal trends, our findings provide evidence of a routine activities pathway, rather than the temperature-aggression hypothesis.
Read full abstract