Louise Dixon, Annabelle Wride
9 min
Intimate partner aggression (IPA) is a complex, heterogeneous phenomenon. While various classification systems (typologies) have been developed to categorize perpetrators, these systems have largely failed to improve treatment outcomes. The authors argue that this failure stems from a mismatch between the purpose of these classifications and the units used to create them. Most existing typologies, such as the seminal Holtzworth-Munroe and Stuart (1994) model, rely on descriptive risk factors (e.g., severity of violence, psychopathology, or criminal history) rather than the causal mechanisms that drive the behavior.
The authors critique the Holtzworth-Munroe and Stuart (1994) typology, which divides perpetrators into 'Family Only,' 'Generally Violent/Antisocial,' and 'Dysphoric/Borderline' categories. They find this system lacking for three main reasons: it lacks clear definitions for its dimensions (e.g., what constitutes 'moderate' severity), it conflates statistical correlates with causal constructs, and it fails to provide discrete categories. Because these categories are based on risk markers rather than the 'why' behind the aggression, they offer little guidance for clinicians on how to design or target specific interventions.
To move beyond the limitations of current models, the authors propose adopting the Functional Offending Behavior Classification Framework (FOBCF). This framework shifts the focus from static risk factors to the motivational systems that organize human behavior. By identifying the underlying goals (e.g., status, security, or pair bonding) that an individual is attempting to achieve through aggression, clinicians can develop more nuanced, individualized treatment plans. This approach treats aggression as a goal-directed behavior, allowing for a more holistic understanding of the offender's needs and providing a clearer path for intervention that addresses the root causes rather than just the symptoms.
By shifting toward a functional classification, researchers and practitioners can move away from 'one-size-fits-all' treatment models. This approach recognizes that the same aggressive outcome can stem from different motivations, and that individuals may use different strategies to meet their goals over time. This shift is essential for developing more effective, inclusive, and targeted interventions that can address the broad spectrum of harm associated with IPA, ultimately improving outcomes for families and reducing the intergenerational cycle of violence.
Alex: Which existing classification do they use to examine that problem? I’d want to know whether they picked an approach that actually aimed to inform treatment.
Sam: They examine Holtzworth-Munroe and Stuart’s typology from 1994. It explicitly aimed to identify underlying processes in men’s marital violence and potentially improve treatment matching. Its developers reviewed 15 existing typologies. They organized differences around violence severity, violence beyond the family, and psychological or personality problems.
Alex: Those dimensions sound descriptive. What kinds of people were they supposed to distinguish?
Sam: One group was called Family Only: men whose violence was limited to family members, with fewer other identified problems. Another was Generally Violent/Antisocial, with violence inside and outside the family. The Dysphoric/Borderline group was described through distress and personality difficulties, with violence primarily within the family.
Alex: Were those groups actually found, or did they remain a proposed way of organizing the literature?
Sam: A later validation study used a community sample of 102 physically violent men and their wives. It also included a non-violent control group. The researchers broadly recovered the proposed groups, but also identified a fourth, called Low-Level Antisocial.
Alex: An extra group could just refine the classification. What evidence makes the authors question whether the categories distinguish different underlying processes?
Sam: The generally violent and distressed groups did not differ as expected on antisociality. Many proposed differences in associated risk markers were not statistically significant. Follow-up research retained many group differences, but some overlap persisted. Other studies also struggled to distinguish those two groups.
Alex: So the critique is not that every distinction disappears. It’s that the boundaries don’t consistently match the explanatory story.
Sam: There’s also a conceptual problem: severity is typically an outcome of physical violence, not an explanation for it. And “low,” “moderate,” and “high” were defined relative to the groups, without clear standalone thresholds. That makes the classification harder to apply to an individual without a comparison cohort.
Alex: What happens to the Family Only group? If they have fewer recognized problems, do we learn less about why they aggress?
Sam: That is a central concern. They were originally expected to represent about half of the men. Yet the group is largely characterized by problems they lack compared with other groups. Absence of those risk markers does not explain their aggression, or establish that their behavior cannot cause serious harm.
Alex: Let’s turn to the alternative. What would a classification based on goals ask that these existing labels do not?
Sam: Dixon and Wride consider Ward and Carter’s Functional Offending Behavior Classification Framework. It uses motivational systems: psychological mechanisms that organize goals and responses to pursue them. Instead of starting with severity or risk markers, it asks what the person’s behavior is trying to achieve.
Alex: Can you give an example where that changes the explanation? Otherwise “goals” could become another broad label.
Sam: The framework includes a fear-and-security system, concerned with detecting and avoiding threats. Problems in that system might appear as anxiety or excessive dependency. Harassment or stalking might also serve a goal of reducing fear through proximity to another person. Those are possible functions to investigate, not motives established by the act alone.
Alex: That last distinction seems essential. Observing stalking does not automatically tell you why that particular person is doing it.
Sam: The approach uses functional analysis: assessment of behavior’s goals and the context shaping it. A practitioner and client explore developmental and learning history, behavioral patterns, and circumstances. They examine both offending and non-offending problem behaviors. Then they ask whether those behaviors achieve the goals, and what resources or constraints shape the person’s options.
Alex: How would that assessment translate into treatment rather than just produce a richer account?
Sam: The proposed target is the difficulty meeting the underlying goal. Treatment would develop alternative strategies for meeting it without problematic behavior. Several problems might reflect attempts to meet the same goal, rather than separate risk factors. The authors suggest this could address difficulties beyond partner aggression itself.
Alex: But a plausible mechanism is not evidence of better outcomes. How much confidence should listeners place in this alternative?
Sam: Treat it as a research proposal, not a validated replacement. This review neither constructs a partner-aggression classification nor tests treatment matching with it. Which motivational systems matter most remains open. So does its applicability across demographic groups, including gender identities.
Alex: And the authors aren’t saying to abandon risk assessment. They’re separating the prediction task from the explanation task.
Sam: They retain a role for risk assessment and management. But they argue that treatment outcomes should extend beyond officially recorded reoffending. Psychological aggression and other family problems can cause distress even when someone has low estimated risk of criminal recidivism.
Alex: Who should read the full document, and where would you send them first?
Sam: Researchers developing partner-aggression theories or classifications, and clinicians designing treatment, should read it. Start with the critique of Holtzworth-Munroe and Stuart, then “How can classification of IPA best inform treatment?” Read the alternative as a framework needing tests, not a finished system. For everyone else: a category that predicts harm does not necessarily tell you how to prevent it.
Alex: Thanks for listening.