SIUE study links homicides to heat

By Kelsey Landis, The Telegraph, Alton, Ill.

Hotter temperatures and an increase in homicides may be connected, according to the results of a study by two Southern Illinois University Edwardsville researchers.

The researchers found there is a 6 percent average increase in homicides for every degree Celsius increase in average temperatures. The year-and-a-half-long study was conducted by Dennis Mares, a professor in sociology and criminal justice, and Ken Moffett, a political science professor.

“A lot of people think, if it’s one or two degrees warmer, why should I care? But there are unintended consequences,” like an increase in homicides, said Mares. “Society is very interconnected. We think of ourselves as being independent of nature, but we’re not. It’s important to point that out.”


Mares and Moffett collected and compared data on average temperatures, homicide numbers and economic data in 57 countries between 1995 and 2012. The article was recently published in Climatic Change, an international journal.

Two theories can potentially explain the connection between homicides and hotter temperatures, Mares said. One theory says that as the weather becomes more pleasant, more people will be out and about, and statistically speaking, a homicide is more likely.

The second theory says that more people become aggravated and aggressive in hot weather. Whatever the cause, the connection between the weather and homicide is stronger in certain areas of the world. The study looked at western and non-western countries with a focus on urban areas, which the researchers found are more likely to see an increase in homicides as a result of increased temperatures.

This may be because urban areas are already affected by higher levels of homicide. Urban areas where more people are on the street — like where gangs operate, Mares said — are more likely to be influenced by the weather simply because people are outside more often. The connection was weaker in more affluent areas.

“The region hit hardest are the poorest regions on the planet. The effects are by no means equally distributed,” Moffett said.

Such variation indicates that climate change may acutely increase violence in areas that already are affected by higher levels of homicides and other social dislocations. The most acute effects also happen to occur in the poorest areas of the world, such as some African and South American nations.

In those places, for each degree Celsius increase, there was a nearly 18 percent increase in the homicide rate.


There are two “major implications” of the study for society, Mares said.

“One, it’s worse than we thought before,” Mares said of earlier studies with a more limited scope. “Two, climate change does a whole bunch of things outside of just rising temperatures.”

Mares and Moffett’s area of study is relatively new, Mares said, because it combines research in climate, crime and socio-economics. The researcher added he believes there is room for growth.

There is not much debate in the scientific community about whether or not climate change is real, Mares said, and the amount of studies show the evidence is overwhelming.

“In our study, too, we see the exact same pattern the climatologists have found,” Mares said.

Moffett said he and his colleague ran externally verified data models upwards of 1,000 times. .

“You can drop out countries, years, variables, even change modeling technique. No matter how you slice and dice the data, you get the same result,” Moffett added.

Data on homicides came from the United Nations, which breaks information down by year and country. The UN collects its information on homicides from law enforcement agencies and the World Health Organization. The researcher’s climate data came from the Global Historical Climatology Network, and economic data came from the World Bank.

Kelsey Landis can be reached at 618-208-6460, Ext. 1396 or on Twitter @kelseylandis.

(c)2015 The Telegraph (Alton, Ill.) Visit The Telegraph (Alton, Ill.) at

Distributed by Tribune Content Agency, LLC.