Examining gender bias in the policing of intimate partner violence

Intimate Partner Violence (IPV) remains a pressing public health and criminal justice issue. Concerns have been raised about police misidentification of IPV victims as offenders - particularly women - due to factors such as self-defensive violence and non-conforming victim behaviour. Misidentification can have serious consequences, including criminalisation of victims and reduced access to support services.

This study investigates whether police are more likely to charge women with IPV offences by examining whether different evidentiary thresholds are applied to male and female persons of interest (POIs) in IPV assault cases. We use a robust instrumental variables approach to compare conviction rates among ‘marginal’ defendants—those who are on the cusp of being charged or not, depending on the attending officer’s individual tendency to proceed with legal action.

The analysis draws on 52,423 IPV assault incidents reported between July 2010 and February 2023, involving 38,413 male and 14,010 female POIs. By comparing conviction rates among marginal male and female POIs, we assess whether police require less evidence to charge women. If so, this would suggest gender bias in police charging decisions and support concerns about misidentification.

Key findings

Figure 1 presents our estimates of the impact of being proceeded against on the probability of conviction by gender for marginal POIs, after controlling for case characteristics, demographics and prior offending. 

Marginal female POIs had a conviction rate of 81.7%, while marginal male POIs had a rate of 67.2%. Although this 14.5 percentage point difference suggests stronger cases against women, it was not statistically significant. 

This indicates that police officers apply similar evidentiary thresholds when deciding to charge male and female POIs in IPV assault incidents. The findings suggest no gender bias in police charging decisions among experienced officers. However, the analysis does not account for potential bias in other parts of the justice system or among less experienced officers.