Complex research on weather factors and crime across long periods of time or in numerous locales requires the handling of extremely large data sets, and this has been possible only with the development of sophisticated data-gathering meteorological instruments; the advent of the computer; and the development of analytical techniques to handle large, complex data sets. As a consequence, and despite significant early studies conducted with the limited data and analytical techniques available, the study of detailed weather patterns and resultant crime changes is only a little over three decades old. In that short history there are some contradictory results in the research, often based on results obtained only for very limited geographic areas or periods of time. There are also a number of questions that have not yet even been addressed. Despite those problems, however, there do appear to be some basic conclusions we have reached that can be taken as a starting point for future research:
- Of all of the weather variables measured, only higher temperatures, often augmented by higher humidity, show a consistent and robust relationship to crime.
- In the United States there are pronounced seasonal patterns for rape, assault, burglary, and larceny, with all of these crimes increasing during the summer months. Other nations also display seasonal patterns for specific crimes, but they do not always have the same summer peak seen in the United States.
- Both murder and robbery show regular monthly patterns, but not along seasonal lines. From this, it is important to understand that the impact of weather, whether seasonally or day by day, on any population is mediated by the culture, the social and economic status, and the demographic structure of that population.
- The most effective model explaining the observed relationship between weather or season and crime is routine activities theory. Activities common to hotter weather or hotter seasons tend to directly or indirectly influence the probability of a convergence of motivated offenders, suitable targets, and reduced guardianship. Interactionist models that consider changes in stress produced by differing behavior patterns occurring in different weather situations or seasons also show promise in explaining changes in personal crime.
- Climate as an explanatory variable for crime differences or the alleged criminality of any population has been largely discounted.
It is inherent to the nature of science that conclusions that are accepted at one point in time will change as more research is conducted to test those conclusions. Criminology scholars need to continue studying the relationship of weather and homicide in order to sophisticate theory, to test previous results with more accurate data and better analytical techniques, and to produce policy recommendations that can help reduce crime.
There is a need to improve criminological theories that address the relationship of weather and crime. Criminology needs better integration of the existing theoretical models, and there is a need to continue research that specifically tests theories of the weather–crime relationship. Some studies have done this, but future research needs to develop very specific testable propositions that would enable us to integrate (or to distinguish between) two or more theories and then perform the research necessary to test those propositions.
We now have ever more complex and detailed data. Where early research often had only the number of crimes reported in some area over some period of time and, at best, daily weather data for those areas and times, modern technology can now secure hour-by-hour weather and crime data for any number of places over long periods of time. However, only improvements in theory can lead to more carefully selected data and more precisely targeted analysis. The research conducted from the 1970s through the 1990s began to explore the possibilities. Now, in the 21st century, criminology has enough consistent findings, much more sophisticated analytical techniques, and equipment capable of applying these techniques to massive data sets.
With all of this, further research is needed across broader geographic areas and over longer periods of time. We are becoming aware that weather “works” in interplay with temporal data, for example. We know that weather conditions can change the impact of time of day or day of the week, and we suspect this applies to major holidays as well. If some assumptions of interactional stress are correct, we need to begin to examine these interactions in more detail, including examining whether these weather and time interactions are different in hotter versus cooler climatic areas.
It is also significant for the future study of weather effects on crime that we are controlling our weather environment far more than we used to. We air-condition our homes, our cars, our businesses, and our places of entertainment. The data may be hard to obtain, but criminologists need to begin to consider what impact that has, particularly as it spreads (or fails to spread) to subgroups in our society, specifically, the poor. If these changes allow most people to mitigate much of the impact of increased heat and concomitant stress, but are not available to the poor, what impact will that have on crime rates in those neighborhoods in which crime is already a significant problem?
In this regard, criminologists need to begin to consider ways to apply the knowledge obtained. We cannot change the weather—at least, not yet. However, if we understand the impact of weather conditions on different areas, different times, and different populations, then we should become better able to prevent, or at least reduce, increases in crime resulting from this impact. It is possible that some of the most basic understandings of how temperature affects people’s routine activities or increases their levels of social stress, with resulting increases in crime, might enable us to act to head off some of those increases. We cannot yet put climate-controlled weather domes over our major cities, but with well-developed theory leading us to examine detailed data, we might be able to find some ways to address the weather–crime interaction with the technology we now have.