Typologies of domestic and family violence perpetrators: A data-driven study

Summary

Background

Domestic and family violence (DFV) perpetrators are often treated as a single group, but they differ widely in their backgrounds, risk factors and contact with government services. Prior research has identified different types of perpetrators, but most empirical studies have relied on limited data sources, such as police or court records. Other studies have developed theoretical or clinically-driven typologies based on small samples.

This study uses linked administrative data from multiple NSW Government systems, including criminal justice, child protection, Out‑of‑Home Care, and housing and homelessness services. The dataset includes records for all individuals with at least one DFV offence finalised in a NSW criminal court between 2008 and 2023, and contains over 270 variables.

We used a two‑step approach to summarise patterns of offending, victimisation and service contact, and then group individuals with similar profiles. Groups are identified separately for males and females. The main analysis focuses on 71,896 individuals born after 1987, for whom childhood system contact is fully observed. The resulting typology is characterised and assessed for robustness.

Key findings

Eight distinct groups of DFV perpetrators were identified (four male and four female). They differ in socioeconomic characteristics, age at first DFV court appearance, Aboriginal representation, and patterns of contact with criminal justice, child protection and Out‑of‑Home Care, housing and homelessness systems, as well as DFV‑related outcomes.

Clear differences emerge between males and females. Among males, groups are mainly distinguished by lifetime criminal justice involvement and recency of DFV‑related offending. Among females, differences are more strongly linked to lifetime criminal justice involvement, prior DFV and violent victimisation, age at first DFV court appearance and housing or homelessness service engagement.

Groups with elevated lifetime and recent DFV offending tend to have higher DFV reoffending outcomes, while revictimisation outcomes are highest among female groups with extensive victimisation histories and housing engagement. Sensitivity analyses show the typology is reasonably robust across cohorts and time periods.

Conclusion

A data‑driven typology of DFV perpetrators can support more targeted responses. Differences between groups, particularly between male and female perpetrators, highlight the importance of not only supervision and compliance responses for higher‑offending groups, but also integrated service responses for groups with complex victimisation and housing engagement.