Statistical Issues in Criminology

Sunday, February 19, 2017: 10:00 AM-11:30 AM
Room 309 (Hynes Convention Center)
As a result of several factors, including a 2009 National Academies report, the role of statistics in forensic science and criminology has become a prominent issue. Many municipalities are exploring the use of predictive policing, in which statistical algorithms are used to forecast the locations and times when criminal activity is most likely to occur; however, the accuracy and value of such tools is still an open question. Similarly, methods from statistical epidemiology are being used to identify regions where drug abuse, rape, or murders tend to occur (and regions in which such crimes are successfully prosecuted). An epidemiological approach can identify common factors that may enable better causal models and more effective intervention. Additionally, several recent shootings by police, caught on camera, have led to the creation of datasets on police homicides. These data are being mined from a number of perspectives, both to improve law enforcement practices and to evaluate the role of race and other factors in police and public interactions. This session provides an overview of different kinds of statistical research and potential uses in addressing the pressing needs of modern criminology.
David Banks, Duke University
David Banks, Duke University
Kristian Lum, Human Rights Data Analysis Group
Societal Consequences of Biased Data in Predictive Policing