Intimate partner violence remains a public health problem of extraordinary scale. According to recent estimates from the World Health Organization, 25.8 percent of women aged 15 to 49 have experienced physical or sexual violence from a partner or ex-partner. Beyond the immediate physical and mental health consequences for victims, the societal costs ripple through families, courts, health systems, and communities. One of the principal tools for prevention is the intervention program mandated for convicted perpetrators, typically delivered as a condition of suspended sentences. Yet these programs vary widely in effectiveness, and researchers have increasingly argued that their success depends on tailoring treatment to the specific risks and needs of each participant, in line with the Risk–Needs–Responsivity model widely used in correctional rehabilitation.

The research team, led by Cristina Expósito-Álvarez of the University of Valencia alongside colleagues Rocío Pérez-Sabater, Manuel Roldán-Pardo, and Marisol Lila, examined a sample of 1,126 men convicted of intimate partner violence against a female partner under Spanish law. All participants were enrolled in the Contexto Program, a community-based, court-mandated intervention at the University of Valencia that spans roughly eleven months and combines cognitive-behavioral techniques, motivational interviewing, and an ecological framework grounded in both psychological and feminist perspectives. Groups of nine to twelve participants, convened between June 2006 and June 2024, received 35 weekly two-hour group sessions supplemented by five individual motivational sessions. During intake interviews, facilitators used the sixth item of the Spousal Assault Risk Assessment Guide, a widely validated clinical checklist, to determine whether each participant had been a direct victim of or a witness to family violence during childhood or adolescence. In total, 330 men were classified as having a trauma history and 796 as having none.

The researchers then compared the two groups across a sprawling social-ecological landscape of variables: sociodemographic characteristics, psychological adjustment, substance use, social-relational factors, attitudes toward women and violence, and violence-related measures. The assessment battery was rigorous and standardized. Anger was measured with the State-Trait Anger Expression Inventory; depressive symptoms with a seven-item short form of the Center for Epidemiologic Studies Depression Scale; impulsivity with the Plutchik Impulsivity Scale; and difficulties identifying and expressing feelings with the Toronto Alexithymia Scale. Substance use was screened with the Alcohol Use Disorders Identification Test and the Severity of Dependence Scale for cannabis and cocaine. Social support was gauged with the Multidimensional Scale of Perceived Social Support, and self-reported violence with the Revised Conflict Tactics Scales. Facilitators also rated each participant’s stage of change according to the Transtheoretical Model and completed the full Spousal Assault Risk Assessment Guide as a measure of recidivism risk. Statistical analyses employed Welch-corrected ANOVAs, chi-square tests with false discovery rate corrections, and hierarchical regression models designed to isolate which variables most strongly distinguished the trauma-exposed group.

The univariate comparisons painted a consistent and sobering picture. Men with a history of childhood or adolescent trauma exposure were younger, less educated, and more likely to be immigrants. They reported significantly higher state and trait anger, a higher overall anger expression index, greater impulsivity, more depressive symptoms, and greater difficulty identifying and describing their own emotions — the hallmark of alexithymia. Substance use was elevated across the board: more problematic alcohol consumption, greater cannabis dependence, and higher cocaine use. Socially, they experienced more stressful life events in the preceding six months and perceived less support from both family and friends, though support from significant others did not differ. Their attitudes were also more troubling, with higher levels of hostile sexism and victim-blaming beliefs about intimate partner violence. On the violence side of the ledger, they self-reported more physical and psychological violence against partners, were further along in their stage of change, and — most strikingly — were assessed by facilitators as being at substantially higher risk of reoffending, with an effect size for recidivism risk large enough that trauma history alone accounted for roughly 15 percent of the variance.

But raw differences between groups can be misleading, because many of these variables travel together. To disentangle them, the researchers built a hierarchical binary logistic regression model, entering variables in four sequential blocks: sociodemographics first, then individual psychological and substance-use factors, then social-relational and attitudinal variables, and finally violence-related characteristics. As each block was added, the picture sharpened. Sociodemographic characteristics alone explained 9 percent of the variance in trauma history, with younger age, lower education, and unemployment all significant. When individual variables entered, impulsivity leapt to the foreground with an odds ratio of 2.74, meaning that higher impulsivity more than doubled the odds that a participant belonged to the trauma-exposed group. By the final model, which explained 20 percent of the variance, four variables remained robustly and independently associated with a trauma history: younger age, higher impulsivity, lower perceived family social support, and unemployment.

The second major question — whether trauma history predicted recidivism risk beyond everything else — was addressed with a hierarchical linear regression using the total Spousal Assault Risk Assessment Guide score as the outcome. The results were unambiguous. After sociodemographic characteristics alone, adding trauma history increased the explained variance in recidivism risk from 11 percent to 22 percent, and trauma remained a significant positive predictor even after individual variables, cannabis use, and social-relational and attitudinal factors were accounted for. In the final model, which explained 31 percent of the variance, trauma history, cannabis use, and hostile sexism were each independently and positively associated with recidivism risk, while younger age, lower education, and unemployment were negatively associated with it — meaning these demographic markers of vulnerability predicted greater risk.

The theoretical weight of these findings is considerable. Several frameworks have long been offered to explain the link between early violence exposure and adult perpetration. The general aggression model suggests that early exposure to violence leads men to internalize aggressive scripts and to normalize violence as a conflict-resolution strategy, in line with the intergenerational transmission of violence hypothesis. Attachment theory points to how early abuse and neglect create maladaptive internal working models of relationships, producing disorganized, avoidant, or insecure attachment styles that shape later intimacy. Feminist perspectives add another layer, proposing that trauma may intersect with patriarchal norms to amplify a compensatory drive for control and dominance in intimate relationships. The new study’s data are consistent with all three: the trauma-exposed men showed the emotional dysregulation predicted by aggression models, the relational mistrust implied by attachment theory, and the hostile sexist and victim-blaming attitudes flagged by feminist analyses.

Mechanistically, the researchers point to the disruption of the emotion regulation system as a central pathway. Early trauma contributes to dysregulated nervous systems characterized by hyperarousal, heightened threat sensitivity, and impaired social information processing. These impairments make it harder to identify one’s own feelings — the alexithymia observed here — and harder to inhibit aggressive impulses during interpersonal conflict. Substance use, in this view, often functions as a maladaptive coping mechanism for the psychological distress of unresolved trauma, which in turn further elevates the risk of violent escalation. The finding that perceived family support, but not support from significant others, was independently lower among trauma-exposed men suggests that mistrust born within the family system may specifically cut men off from the relational resources that could otherwise buffer trauma symptoms.

The authors are candid about limitations. Trauma history was assessed through participants’ self-reports during initial motivational interviews — a moment when therapeutic alliance is still forming and when many perpetrators struggle to reflect on how childhood adversity shaped their present behavior, which may have led to undercounting. The design is cross-sectional, so the regressions identify associations rather than causal or mediational pathways, and recidivism risk was measured by facilitator ratings at intake rather than official records over follow-up. Future work, the team argues, should use longitudinal and mediational designs to test the mechanisms proposed by the theoretical frameworks, and should reassess trauma at multiple points during intervention.

Still, the practical implications are immediate. Trauma-informed approaches, the researchers emphasize, do not necessarily mean trauma-focused therapy delivered inside a batterer intervention program. Rather, they mean building intervention environments sensitive to participants’ histories: screening for impulsivity and substance use, strengthening social support networks, addressing socio-economic vulnerability, challenging normalized sexist beliefs, and above all avoiding retraumatization while maintaining accountability for violent behavior. Similar trauma-informed interventions have already shown promise among veteran populations, and meta-analyses suggest such approaches can reduce recidivism. As the study’s title asks — “Wounded boys, violent men?” — the answer emerging from this large, methodologically careful dataset is nuanced: early wounds do not determine violence, but they reliably mark a higher-risk profile, and programs that ignore them are likely to fail the very participants who need them most.