• April 2018
  • Vol. 19, No. 3

Printer-Friendly version of article

Risk Terrain Modeling Predicts Child Maltreatment

Effective primary prevention methods are imperative to averting the lifelong consequences of child maltreatment, which can include shorter life expectancy, chronic disease, obesity, alcohol and drug use, intimate partner and sexual violence, depression and anxiety, and more. A recent article in Child Abuse & Neglect discusses the first study to apply risk terrain modeling (RTM) to child maltreatment primary prevention efforts.

RTM is a relatively new statistical and geospatial analysis technique that was originally developed by the Rutgers University Center on Public Security for traditional criminal justice applications. RTM can analyze the relative influence of a variety of environmental factors, such as poverty or substance use, on a dependent variable (e.g., child maltreatment), which helps develop more valid predictions and identify risk clusters, which are locations where risk factors for maltreatment accumulate. It has been successful in accurately predicting shootings, robberies, and other crimes and offers a significant advantage over other types of predictive analysis for child maltreatment, such as traditional hotspot maps, that rely on retrospective analysis of past occurrences.

The study team reviewed address-level data for 10 risk factors: commission of aggravated assaults, robberies, murders, domestic violence, narcotics crimes, the presence of gangs and prostitution, runaways, poverty, and the presence of bars and nightclubs with a license to serve alcohol past midnight. Then they used specialized software to examine the geographical distributions of all risk factors in order to determine where the risk of maltreatment is most likely to occur. To test the model's accuracy, the RTM model based on 2013 data was then overlaid with the actual locations of substantiated child maltreatment during 2014 to reveal the following results:

  • With the RTM model, 52 percent of all substantiated instances during 2014 occurred in the one-tenth of the city's area that was shown to have the highest risk, which was better than the 43 percent for the hotspot prediction model.
  • Nearly all (98 percent) substantiated future cases of maltreatment occurred in areas that were flagged as having an elevated risk by RTM.
  • Only 2 percent of incidences of maltreatment occurred in areas that were not identified as having an elevated risk.

RTM offers researchers and prevention providers a predictive tool that can more accurately predict where child maltreatment may occur than other models, which will allow for more targeted interventions, improved allocation of prevention resources and services to the locations with the highest need, and additional information about the risk factors that contribute to the elevated risk.

"Risk Terrain Modeling Predicts Child Maltreatment," by Dyann Daleya, Michael Bachmann, Brittany A. Bachmann, Christian Pedigo, Minh-Thuy Bui, and Jamye Coffman (Child Abuse & Neglect, 62), is available at https://www.sciencedirect.com/science/article/pii/S0145213416301922.

 

<  Previous Article   Next Article  >