Study Validates Brief Anger and Aggression Questionnaire Among Minnesota Inmates

A simple questionnaire developed nearly four decades ago may offer correctional systems a powerful new window into violent behavior behind bars and after release. In a study published in the American Journal of Criminal Justice, researchers Grant Duwe and Valerie A. Clark of the Minnesota Department of Corrections report that the Brief Anger and Aggression Questionnaire, or BAAQ, holds up remarkably well when administered to a large incarcerated population, and that its scores do far more than measure a passing mood. In a sample of 1,522 people imprisoned in Minnesota’s state prison system, the scale behaved as a statistically coherent measure of a single underlying construct, and higher scores were significantly associated with assessed recidivism risk, particularly the risk of committing a future violent offense.

The BAAQ was originally created in the late 1980s as a rapid clinical tool for identifying anger and aggression problems, conditions that psychologists have long recognized as potent drivers of interpersonal violence. Its brevity has always been its central selling point: where comprehensive hostility inventories can run to dozens of items and considerable administrator time, the BAAQ compresses the assessment of anger arousal, aggression, and related behavioral tendencies into a short self-report format that can be completed in minutes. Yet despite its long history and its clinical appeal, the authors note that rigorous validation work on correctional populations has been strikingly limited. Most of the existing evidence for the instrument was built on psychiatric outpatients, batterer intervention samples, and other clinical groups, leaving open the question of whether the scale functions appropriately among people whose anger and aggression histories are, in many cases, directly tied to their incarceration.

Duwe and Clark addressed that gap using data from Minnesota’s prison system, where the BAAQ had been incorporated into a broader needs and responsivity assessment framework. The validation exercise confronted three distinct psychometric questions. First, does the instrument actually measure one thing? Second, do its scores correlate, as theory predicts, with related assessments of risk? And third, does it predict anything that matters in the real world, such as misconduct inside prison or reoffending after release? The answers, on all three counts, were affirmative.

On the structural question, the researchers found that the BAAQ measures a single construct. In practical terms, this means that the individual items on the questionnaire hang together statistically, loading onto one latent dimension of anger and aggression rather than fragmenting into weakly related subcomponents. That unidimensional structure is a meaningful finding for practitioners. A scale that taps a single coherent construct can be scored and interpreted straightforwardly: higher scores indicate more anger and aggression, and there is no ambiguity about how subscale scores should be combined or weighted. For a tool intended for rapid administration at intake in busy correctional settings, that simplicity is not a cosmetic virtue but a functional necessity. The unidimensionality also aligns with how anger-focused interventions are designed, since treatment programs typically target anger and aggression as a unified behavioral problem rather than as several independent traits.

The concurrent validity findings were equally consequential. The BAAQ showed a significant, positive association with assessed recidivism risk, meaning that individuals who scored higher on the anger and aggression measure tended to be rated, through independent risk assessment instruments, as more likely to reoffend. But the association was not uniform across offense types. When the researchers disaggregated the findings, the relationship proved to be driven largely by the BAAQ’s stronger link with assessed violent recidivism risk specifically. This pattern is theoretically coherent. Anger is a well-documented proximal factor in violent crime; laboratory and field research alike have connected heightened anger arousal and aggressive response tendencies to assaultive behavior, while property and drug offenses are more often linked to economic need, opportunity, and associations with criminal peers. A measure of anger and aggression that predicts violent risk more strongly than general risk is, in a sense, doing exactly what it was designed to do.

Perhaps the most striking results concerned the BAAQ’s predictive power for prison misconduct. Duwe and Clark found that scores on the questionnaire significantly predicted disciplinary infractions during incarceration, and, mirroring the concurrent validity results, the effects were larger for violent infractions than for nonviolent ones. Prison misconduct is not a trivial outcome. Infractions ranging from verbal altercations to assaultive attacks on staff or other incarcerated people generate substantial costs for correctional agencies, compromise institutional safety, and disrupt programming and rehabilitation efforts. Moreover, a long tradition of research on behavioral continuity suggests that aggression inside prison is not disconnected from aggression outside it; individuals who were violent in the community tend to bring those behavioral patterns with them through the prison gate. A brief intake measure that flags who is most likely to generate violent incidents during confinement therefore has obvious operational value.

The study sits within a larger scientific and policy conversation about how correctional systems assess the people in their custody. Modern corrections in the United States is organized largely around the Risk-Need-Responsivity model, a framework developed by Canadian psychologists James Bonta and Donald Andrews, which holds that intervention intensity should match an individual’s assessed risk level and should target the criminogenic needs, dynamic factors such as criminal thinking, antisocial associates, and substance abuse, that drive offending. Within that architecture, responsivity factors like anger and anxiety do not typically predict reoffending directly, but they shape how well a person engages with and benefits from treatment. A person with severe anger dysregulation may struggle in a standard cognitive-behavioral program unless the anger itself is addressed first. Instruments like the BAAQ give correctional agencies a way to identify those responsivity barriers at intake and route individuals toward anger management or emotional regulation programming before other rehabilitative work begins.

The Minnesota context matters. The state’s Department of Corrections has been a national leader in developing and validating assessment instruments, including the Minnesota Screening Tool Assessing Recidivism Risk, known as MnSTARR, and its successor, as well as a comprehensive needs and responsivity assessment system whose validity Duwe and Clark evaluated in earlier work. In the current study, the BAAQ functioned as part of that assessment architecture, which gave the researchers the rare opportunity to compare self-reported anger against professionally scored risk assessments and against objective behavioral records of prison misconduct. Few validation studies of brief anger measures have had access to samples of this size, or to such rich administrative follow-up data.

The implications of the findings cut in several directions. For clinicians working inside prisons, the results suggest the BAAQ is a defensible, efficient screening choice: it measures what it claims to measure, it aligns with independent risk assessments, and it anticipates future behavior. For administrators, the study hints at resource savings, since a short instrument that accurately flags treatment needs can help agencies allocate anger management programming, mental health services, and enhanced supervision where they are most likely to reduce violence. And for researchers, the findings add a correctional data point to a psychometric literature that has long been dominated by civilian clinical samples. Previous work on anger assessment has catalogued a dizzying array of instruments, from the classic Buss-Durkee Hostility Inventory of the 1950s to newer aggression scales, but comparatively few of those tools have been stress-tested on the populations where anger and aggression carry the highest stakes.

The authors are careful to frame their findings in terms of potential rather than promise. A significant statistical association between anger scores and violent outcomes does not mean the questionnaire can be used mechanically to decide custody classifications or sentencing, and the researchers position the instrument as one input among many in a comprehensive assessment system. Nor does measuring anger explain it. Anger and aggression in incarcerated populations are shaped by histories of trauma, adverse childhood experiences, mental health problems, and the criminogenic environments many individuals return to upon release. But identifying a measurable, modifiable risk factor is a first step toward treating it, and a growing body of research supports the effectiveness of anger-focused and broader psychosocial interventions in reducing aggression among adults.

The timing of the study is notable. Anger self-management treatments, once considered a soft-edged supplement to correctional programming, have accumulated randomized controlled trial evidence in settings as varied as traumatic brain injury rehabilitation and domestic violence intervention. At the same time, a 2023 review commissioned by the National Institute of Justice of the Federal Bureau of Prisons’ needs assessment system highlighted how unevenly validated many correctional assessment practices remain across U.S. prison systems. Against that backdrop, the Minnesota study offers a concrete example of the diligence the field increasingly demands: an old, inexpensive tool, tested rigorously on a modern incarcerated sample, validated against both concurrent measures and real behavioral outcomes.

What emerges from the analysis is a modest but persistent signal: anger, measured in minutes at intake, echoes through years of imprisonment and into the calculus of reoffending risk. With nearly two million people incarcerated in the United States, even small improvements in identifying who is most likely to act violently, and in targeting them with interventions that actually work, could translate into safer prisons and safer communities. The BAAQ, a questionnaire old enough to have outlived most of the research technology of its era, may turn out to be one of the simplest and most cost-effective instruments in that effort.