Understanding the Risk Factors of Spousal Violence Victimization Using Machine Learning and Network Approaches

Abstract

Abstract Purpose Identifying pertinent risk factors is an essential first step for early detection and upstream prevention of spousal violence. However, limited research has examined the risk factors of spousal violence in the Asian context. This study aimed to understand the spousal violence issue in Singapore by (1) identifying the pertinent risk factors that could predict the likelihood of applying for a Personal Protection Order (PPO) - an order restraining a respondent from committing family violence against a person, and (2) understanding the relationship between various risk factors and the likelihood of PPO application. Method Linked administrative data of ever-married Singapore residents born in 1980 and 1985 (N = 51,853) were analyzed, using machine learning and network approaches. Results Results indicated that the pertinent risk factors associated with PPO application included lower educational attainment, staying in a public rental flat, early marriage and parenthood, childhood maltreatment, prior history of being respondent to PPO, offending behaviors, and mental illness. Conclusions Findings could aid in identifying individuals and families at-risk and informing upstream efforts to combat spousal violence issues. First responders, such as police or social workers, could utilize the relevant risk factor as a guide in cases of suspected family violence to identify at-risk individuals and families in a timely manner and minimize adverse effects.