The Risky Business of Assessing Intimate Partner Violence Risk: Research to Practice

Introduction to the Celia Project

This presentation introduces the CELIA IPV Project, Common Language for Intimate Partner Violence Risk Appraisal—An Evidence-Based Policing Approach. The CELIA IPV Project is a collaboration involving the Ontario Provincial Police, Edmonton Police Service, Saint John Police Force, and researchers from the Waypoint Centre for Mental Health Care, MacEwan University, the University of New Brunswick, and the University of Toronto. The project’s goal was to establish formal partnerships between researchers specializing in intimate partner violence (IPV), risk assessments, and policing services in Canada to address existing research gaps in evidence-based policing for IPV. Researchers coded a randomized sample of domestic violence incidents reported to the police in 2018, resulting in charges (N = 771). In this presentation, you will hear about the origins of CELIA, a sample description, and emerging differences between men and women perpetrators of IPV.

A Common Language for Communicating Risk in Cases of Intimate Partner Violence

Effective criminal justice interventions accord more time and resources to high risk cases than to low risk cases. Risk assessments can only improve practice when interested parties share a common understanding of risk level labels. This is not always the case. The same label (e.g., "high risk", “low risk”) can have different meanings across evaluators, settings, professional groups, and risk tools. To address this problem, a collaboration between Public Safety Canada and the Justice Center of the US Council of State Governments advanced principles for creating a common language of standardized risk levels. They proposed a 5-Level system for general correctional populations (Hanson, Bourgon et al., 2017), which has subsequently been adapted for sexual recidivism risk (Hanson, Babchishin et al., 2017) and the risk for violence in general (van Dooren et al., 2024). This presentation discusses the foundations of standardized risk levels for IPV, informed by recent research on the Ontario Domestic Assault Risk Assessment (ODARA). Although the ODARA research provides a valuable testing ground for standardized IPV risk levels, standardized risk levels should be independent of any particular risk tool, and be revised as new evidence becomes available.

Ideal vs Reality: Police Use of the ODARA in the Field for IPV Risk Appraisal and Mitigation

The Ontario Domestic Assault Risk Assessment (ODARA) was introduced in New Brunswick to aid police with appraising risk of intimate partner violence (IPV) and formulate effective risk management strategies. However, the extent to which police officers utilize the ODARA to inform these decisions is unclear. We examined 186 police IPV case files from 2016-2018 including police-scored ODARAs, and police actions. Cases were followed for recidivism (M = 6.34 years; 29.3% IPV recidivism). 87.7% of files had an officer-completed ODARA and most completed it the same day as the incident, reflecting timely completion. However, multivariate analysis of covariance revealed that the ODARA did not have a significant impact on the risk management strategies used. Furthermore, logistic regressions revealed that specific types of risk mitigation strategies did not predict recidivism. This research identifies potential weaknesses in how officers are trained on the ODARA as a decision-aide for informing police response for IPV risk mitigation.

Violence in Non-Cohabiting Relationships: How to Score the ODARA

 The Ontario Domestic Assault Risk Assessment (ODARA) identifies the risk that an individual who assaulted their intimate partner will assault a partner again in the future. It was originally developed for men who assaulted their female “domestic” (marital or cohabiting) partners. Some recent studies have included couples who were not cohabiting at the time of the index assault or who had never cohabited, and the ODARA can be used in these cases. However, three ODARA items concern pre-index domestic and nondomestic partners, and the definition of “domestic partner” in these items has not been changed. This talk will describe a study that asked, What happens if we simplify the ODARA scoring by treating marital, cohabiting, and dating violence the same throughout the ODARA? We tested this modified scoring system in 566 cases of men who were charged following an assault on their female intimate partner, including a subgroup of 100 noncohabiting partners. Modified ODARA total scores were slightly lower than original scores. Both scoring methods were approximately equally valid for assessing risk of a new charge for a violent offence against an intimate partner, in the whole group and the cohabiting and noncohabiting subgroups.

Assessing Bias: The Ontario Domestic Assault Risk Assessment Tool and Indigenous Offenders

 The Ontario Domestic Assault Risk Assessment (ODARA) was developed to assess the risk of intimate partner violence (IPV) recidivism. However, little is known about its applicability to Indigenous populations. This study examined whether the ODARA is equally effective in predicting IPV recidivism among Indigenous and non-Indigenous offenders and explored whether criminogenic needs accurately predict reoffending for both groups. It was hypothesized that the ODARA will demonstrate greater predictive accuracy for IPV recidivism for non-Indigenous than Indigenous perpetrators, and the prevalence of criminogenic needs will differ between groups except for procriminal attitudes and poor use of leisure time. IPV cases (N = 360) that were reported to the Edmonton Police Service in 2018 were analyzed. Data has been collected and coded, and results of statistical hypothesis testing will be presented. The implications for using the ODARA with Indigenous individuals will be discussed.

Co-occurring Child Abuse and Risk for Future Child Abuse in Cases of Intimate Partner Violence

 Research demonstrates a significant co-occurrence of intimate partner violence (IPV) and child abuse. Despite this, there is little research examining co-occurring or future child abuse in IPV cases. Police attending IPV calls routinely examine the risk for recidivism, but existing IPV risk measures do not address co-occurring or future risk of child abuse. In this presentation, we will discuss a study that investigated the identification of co-occurring child abuse and IPV using the Salisbury risk factors for child abuse and the ODARA risk assessment tool for IPV recidivism. Additionally, we explore the potential for using IPV risk assessment to evaluate the risk of future abuse of children and discuss potential challenges to the research and applied practice for using these measures. Structured risk assessment tools may offer some practical promise for improving the identification of co-occurring as well as future child abuse and officers’ recognition and response to child abuse in IPV situations.

Assessing Coercive Control in Police Reports of IPV

 Coercive control is a pattern of behavior that often co-occurs with physical intimate partner violence (IPV), and in Canada, coercive control will soon become a criminalized behaviour. While prior research suggests that police can identify acts of coercive control, there remains considerable disagreement on how to define and measure it. This lack of consensus presents challenges for both researchers assessing its role in physical IPV risk and for criminal justice professionals tasked with addressing the offence. In this presentation, we will discuss three studies sought to establish a reliable third-party measure of coercive control for use in police reports involving physical IPV. Our findings indicate that coercively controlling behaviours can be identified and documented with moderate to good interrater reliability, particularly when using structured checklists with item-specific descriptors. Results also show that operationalizing coercive control through explicit behavioural examples improves coding reliability. These studies have important practical implications for ensuring consistency and reliability in the identification of coercive control, thereby improving police investigations and evidentiary burden of proof.

Seeing the Signs: What Police Reports Reveal About Coercive Control and Risk

Police officers often encounter domestic disputes, and not all dangerous situations involve visible injuries. This presentation discusses how certain patterns of behavior, such as jealousy, psychological abuse, stalking, and controlling behavior, can form a bigger picture of what’s called coercive control. These behaviors and the attitudes behind them (like denying abuse or justifying violence) are not always physically violent but can be just as dangerous over time. The key message is the implication of coercive control on the risk of recidivism and the level of physical injury. That means what officers document during a domestic call, even if there are no visible injuries or physical assault, could help prevent serious harm. Paying attention to controlling behaviors and attitudes might help officers assess risk better and take steps to protect victims before violence escalates.

Threats of Self-Harm in IPV and its Relation to Mental Health and Coercive Control

Threats of self-harm may be seen as coercive when made in the context of perpetrating intimate partner violence (IPV) but may also indicate a mental health concern. We compared 113 men who threatened self-harm with 453 who did not, in cases of men with a police report of IPV. Men who threatened self-harm were younger than men who did not, and they were more likely to be unemployed, be in unstable relationships, and have a previous IPV offence. They were also more likely to commit specific acts of violence against their partner and had higher risk of reoffending. The odds of threatening to self-harm increased with every additional reported coercive and controlling behaviour, but there was no relation between self-harm and mental health service referrals, suggesting that threats of self-harm are more closely related to coercive control in this context.

In this final session, we reviewed four categories (i.e., research needs, knowledge translation gaps, knowledge mobilization challenges, and ways to overcome knowledge mobilization challenges). View a summary of the round table discussions here