Criminality, Victimization and Social Interactions (CIVICS)

A large social science literature tries to describe and understand the causes and consequences of crime, usually focusing on individuals’ criminal activity in isolation. The ambitious aim of this research project is to establish a broader perspective of crime that takes into account the social context in which it takes place. The findings will inform policymakers on how to better use funds both for crime prevention and the rehabilitation of incarcerated criminals.
Criminal activity is often a group phenomenon, yet little is known about how criminal networks form and what can be done to break them up or prevent them from forming in the first place. Overlooking victims of crime and their relationships to criminals has led to an incomplete and distorted view of crime and its individual and social costs. While a better understanding of these social interactions is crucial for designing more effective anti-crime policy, existing research in criminology, sociology and economics has struggled to identify causal effects due to data limitations and difficult statistical identification issues. This project will push the research frontier by combining register datasets that have never been merged before, and by using several state-of-the-art statistical methods to estimate causal effects related to criminal peer groups and their victims.

More specifically, we aim to do the following:

Part 1: Criminal networks. Using rich administrative data, we will document the existence of criminal networks in relation to several dimensions, including neighbourhoods, prisons, schools and racial and immigrant groups. We will use recent advances in network modelling to describe the structure and density of criminal networks, and to identify the key players within criminal networks. We will also focus on the jail-mates of incarcerated criminals to study whether or not they form new criminal networks or promote the merging of existing criminal networks. To better understand interactions among criminals, we will explore how the arrest, incarceration or death of a criminal from a network affects other members of the group.
Part 2: Victimization. Using register data, we will link all victims of charged crimes in Norway from 2004-2015 to their offenders. To our knowledge, this will be the first large panel dataset in any country with such offender-victim linkages. We will describe incidences of victimization along several dimensions, including demographic characteristics and crime type, while noting whether or not victims stay in contact with their offenders after a crime. We will then estimate victimization costs, including the actual measures of lost earnings, disruptions in family relationships, and physical and mental health problems. A key advantage of this method is that we will not need to rely on retrospective survey data that is likely biased, but can use actual outcomes, many of which have not been previously linked to victims. Finally, we will explore whether sending offenders to prison affects long-term health, family and labour market outcomes for victims and their families, an area for which no causal estimates currently exist.

Part 3: Prison rehabilitation programs. Our third, and riskier, part of the project aims to better understand which prison programs are most effective at rehabilitating criminals. We will be running randomized controlled trials for specific training and education programs offered within Norwegian prisons. Linking the data from these experiments to our register data, we will be able to estimate the spill-over effects from these experimental interventions on other members of criminal and victim networks.