SPEA study shows links between land use and violent crime rates
Land use matters when it comes to predicting violent crime rates, according to results of a study by two professors in the IU School of Public and Environmental Affairs at Indiana University-Purdue University Indianapolis.
Thomas Stucky and John Ottensmann show that rates of murder, rape, robbery and aggravated assault are generally higher in areas with high-density residential developments and commercial property, and generally lower in areas characterized by industry, parks and schools.
Illustration by Ned Shaw
But the correlations aren't always straightforward. They are influenced, to varying degrees, by socioeconomic factors such as poverty rates and neighborhood residential stability.
"We found you couldn't look at either the socioeconomic characteristics of a neighborhood or the land-use configurations separately," Stucky said. "It's critical that you understand both in order to understand the crime patterns."
The study, titled "Land Use and Violent Crime," was published in the November 2009 issue of the journal Criminology. It employs geo-coded Uniform Crime Report data for the city of Indianapolis, along with information on 30 categories of land use and demographic information from the 2000 U.S. Census, to map relationships between land use and crime. The research was sponsored by the Indiana University Public Policy Institute (https://policyinstitute.iu.edu/index.aspx).
While other studies have examined crime rates by geographical units such as street blocks or Census tracts, Stucky and Ottensmann took an innovative approach. They used data for 1,000-by-1,000-foot grid squares, providing objective and precise plotting of land-use types and crime locations. Also, previous studies of crime and land use tended to focus on specific uses, such as proximity to taverns or schools; and they often looked at land use independently of socioeconomic factors.
Some of the results are, on the surface, not unexpected -- for example, that there are more robberies in commercial areas. But putting both land-use categories and socioeconomic factors in the mix led to complex and sometimes surprising findings. For example, in "disadvantaged" areas with no commercial land use, rates were higher than average for homicides but lower than average for other violent crimes. At the same time, in better-off areas with commercial land use, rates are higher than average for robbery but low for other violent crimes.
"People might expect the rates for homicides and robberies to both be higher in disadvantaged areas, but we didn't find that," Stucky said. "This allows you to think in more nuanced ways about where you would expect to see different crime configurations."
The study found higher rates of all types of violent crime in areas of high-density residential land use, even after controlling for overall population. The correlation was more pronounced in disadvantaged areas but held true in other areas as well.
"There seems to be something about (high-density residential) units that is associated with all types of serious violent crime, even controlling for the other factors in the model," the authors write. "Apparently, high-density housing units promote serious violent crime."
Generally speaking, the study found higher rates of robbery, aggravated assault and rape in commercial areas, and higher rates of all violent crimes in areas traversed by major streets. It found generally lower violent crime rates in areas with parks, cemeteries and schools.
Stucky is a criminologist and former law enforcement officer, while Ottensmann is an expert in urban land use, especially the development of land-use models. Their collaboration took root several years ago when Stucky attended a presentation by Ottensmann on LUCI, the Land Use in Central Indiana model, which facilitates urban planning by showing the relationship between policy choices and development.
They realized that, with massive data sets available on both land-use patterns and crime, it made sense to combine the topics -- and their research specialties -- and look for relationships.
"It's a perfect example of the kind of collaboration that comes out of serendipity, being in the right place at the right time and open to new opportunities," Stucky said.
