Crimes of opportunity or crimes of emotion? Testing two explanations of seasonal change in crime

While past research has suggested possible seasonal trends in crime rates, this study employs a novel methodology that directly models these changes and predicts them with explanatory variables. Using a nonlinear latent curve model, seasonal fluctuations in crime rates are modeled for a large number of communities in the U.S. over a three-year period with a focus on testing the theoretical predictions of two key explanations for seasonal changes in crime rates: the temperature/aggression and routine activities theories. Using data from 8,460 police units in the U.S. over the 1990 to 1992 period, we found that property crime rates are primarily driven by pleasant weather, consistent with the routine activities theory. Violent crime exhibited evidence in support of both theories. Sociologists have long had an interest in how seasonal climatic changes may interact with social structures to influence the behavior patterns of individuals. Early work in this area includes Durkheim’s classic studies of seasonal differences in suicide rates (Durkheim 1952:107-18), a topic that has seen renewed attention in recent decades (Bollen 1983; Warren 1983). Seasonality in birth and death rates has also been investigated (Land & Cantor 1983), as has the linkage between seasonal changes in testosterone production and sexual activity, with mixed success (Smolensky et al. 1981; Udry & Morris 1967). This * We thank the members of the Carolina Structural Equation Modeling (CSEM) group for their comments and suggestions on earlier drafts of this article. The research on which this article is base