The study arrives at a moment of acute public anxiety in South Africa. Violent crimes such as cash-in-transit heists and mass shootings have increased substantially in recent years, dominating headlines and fueling debate about the effectiveness of policing and punishment. Yet the research draws attention to a quieter, less visible layer of criminality. Petty offences such as theft and burglary receive far less media coverage, but they remain pervasive and account for the majority of criminal cases in the country. Together, minor and major offences contribute to elevated public safety concerns and deepening doubts about whether the criminal justice system, as currently configured, is capable of delivering durable reductions in crime.

At the heart of the study is a concern that criminologists have long raised but that is rarely quantified at a national scale: imprisonment itself may, under certain conditions, intensify criminal behavior rather than suppress it. In South Africa, this concern takes on particular urgency because prisons are chronically overcrowded, and minor offenders are routinely housed alongside hardened criminals. The study argues that this mixing of populations creates environments that substantially elevate the risk of recidivism, meaning that a person convicted of a relatively minor offence may emerge from custody more deeply embedded in criminal networks and behaviors than before.

To test these ideas rigorously, Nyabadza built a mathematical model describing the population dynamics of individuals as they move among states defined by criminal behavior and rehabilitation status. Drawing on South African crime and prison data spanning 1995 to 2024, the model tracks pathways related to susceptibility, petty crime, serious crime, custodial sentencing, non-custodial sentencing, rehabilitation, and relapse. In this framework, a custodial sentence requires an offender to serve time locked in a prison or secure institution, while a non-custodial sentence allows the offender to remain in the community under specific rules and restrictions. By simulating how people flow between these states over time, the model can estimate how different sentencing policies alter the overall trajectory of criminal activity in the population.

The results are unambiguous in their direction. Indiscriminate custodial sentencing of petty offenders, the model shows, can amplify criminal activity across the system. By contrast, the risk that a petty offender escalates to serious crime is substantially lower when the person remains under community supervision. In other words, the pathway from a shoplifting or burglary conviction to involvement in violent or organized crime appears to run, at least in part, through the prison gate itself. The finding quantifies what has often been described anecdotally: that prisons in overcrowded conditions can function as schools for crime rather than institutions of reform.

According to the study author, the policy implications are clear. Evidence-based approaches that prioritize non-custodial rehabilitation, relapse prevention, and prison reform are described as essential for curbing recidivism and achieving durable crime reduction. Rather than reserving community sentences only for the least serious cases as a matter of leniency, the research suggests that keeping petty offenders embedded in their communities, with structured supervision and rehabilitation support, may be the more effective public safety strategy. This reframing positions sentencing reform not as a soft option but as a calculated intervention with measurable benefits for overall crime levels.

The study also illustrates how policy decisions leave detectable traces in national data. In a comment accompanying the research, Nyabadza noted that what struck him most was how clearly policy leaves its fingerprints in the data. He pointed to the sharp prison-population spike around 1999 to 2002, which coincided with the introduction of mandatory minimum sentencing, as an example of how legislative shocks can reveal insights that a smooth mathematical model alone cannot capture. That episode, in which a change in sentencing law produced a visible surge in incarceration, provides a natural experiment embedded in the historical record, allowing the model to capture the downstream consequences of abrupt policy shifts.

Like all modeling studies, this one carries important limitations that shape how its findings should be interpreted. The author notes that the results might not generalize to other countries, since the model is calibrated to South African crime and prison data and reflects the specific institutional conditions of that system. Pre-trial detention data are not included in the analysis, which could lead to an underestimation of criminal career duration, since time spent awaiting trial is invisible to the model. Additionally, the model allocates all released petty offenders to a high-risk class, a conservative assumption that may overestimate the escalation of criminal activity after release. These caveats mean the estimates should be read as indicative of direction and mechanism rather than as precise predictions for any individual case.

Further limitations concern what the model leaves out entirely. The methods do not capture the explicit socioeconomic drivers of crime, such as unemployment, inequality, or educational access, which are widely recognized as powerful influences on criminal behavior in South Africa and elsewhere. Nor does the study include a cost-effectiveness analysis of interventions, so while it can indicate that community-based approaches may reduce escalation, it cannot yet say how the financial costs of expanding non-custodial rehabilitation would compare with the costs of incarceration. These gaps define a clear agenda for future work, combining criminological modeling with economic evaluation and socioeconomic data to produce a fuller picture of the policy landscape.

Even with these constraints, the study offers a quantitative framework for one of the most consequential debates in South African criminal justice: what to do with the thousands of people convicted each year of petty offences. By translating the dynamics of sentencing, rehabilitation, and relapse into a formal mathematical structure, the research makes it possible to compare policy scenarios on a common footing and to see how choices made in courtrooms today ripple through prison populations and crime statistics for years afterward. The findings highlight the potential benefits of prison reforms that integrate deterrence with effective community-based rehabilitation, suggesting that the path to lower violent crime may begin with how the state responds to its least serious offenders. The full article is freely available in PLOS One for readers who wish to examine the model and its results in detail.

Subject of Research: Mathematical modeling of the criminogenic effects of incarcerating petty offenders in South Africa

Article Title: Incarceration may escalate criminal behavior in South Africa

Article References: Incarceration may escalate criminal behavior in South Africa. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: South Africa, incarceration, recidivism, petty crime, non-custodial sentencing, prison overcrowding, mathematical modeling, PLOS One, criminal justice reform, rehabilitation, mandatory minimum sentencing, crime policy

The design of the study is deliberately straightforward. Half of the participants, all of whom will have alcohol use disorder, opioid use disorder, or both, will be randomly assigned to receive once-weekly semaglutide injections for 52 weeks. Crucially, the medication will be offered as an additional support layered on top of existing care, not as a replacement for traditional addiction treatment. Participants will remain eligible for all other behavioral health supports throughout the study period. Recruitment will focus on individuals who are clinically appropriate for GLP-1 medications and who face an elevated risk of criminal-justice involvement, a targeting strategy that reflects the county’s interest in directing new interventions toward those with the greatest needs and its broader commitment to reintegration and community safety.

The scientific rationale rests on a growing but still preliminary body of evidence. Phase 2 trials conducted internationally have indicated that GLP-1 receptor agonists can decrease heavy drinking among people with alcohol use disorder and reduce opioid craving among people with opioid use disorder. Prior research published in The Lancet and JAMA Psychiatry, along with earlier work on the related drug liraglutide, has suggested that these medications, originally developed to mimic gut hormones that regulate appetite and blood sugar, may also modulate reward circuitry in the brain. Researchers involved in the trial describe the effect in practical terms: rather than feeling dominated by cravings, whether for sugar, alcohol, or drugs, many patients report that GLP-1 medications help quiet that internal noise, restoring a sense of personal agency that addiction erodes.

What sets this trial apart from most addiction medication studies is the breadth of its outcome measures. Where typical trials primarily track substance use, this study will exploit Allegheny County’s nationally recognized linked administrative data infrastructure to examine whether treatment improves real-world outcomes across entire systems. Researchers will follow participants’ employment, housing stability, health care utilization and costs, mental health, self-control and impulsivity, weight, and involvement with the criminal justice system, including rearrest, reconviction, and reincarceration. The team will also calculate total fiscal and social costs, allowing the study to answer not only a clinical question but a policy one: whether any clinical benefit is large enough to justify sustained public investment.

The partnership behind the trial grew out of an unlikely collaboration between a research organization and a county government. Alex Jutca, Director of the Allegheny County Department of Human Services, described reaching out to CASPR executive director and co-founder Nicholas Reville two years ago after reading his addiction medicine blog. The two began talking regularly about GLP-1 medications, addiction, recidivism, and the challenges the county was facing, and the trial concept emerged from those conversations. The collaboration combines clinical treatment expertise, the county’s integrated data systems and social service partnerships, economic analysis, and addiction-policy expertise within a single randomized study, a structure its architects say is extremely unusual for county government.

Clinical leadership will come from UPMC Western Psychiatric Hospital’s Addiction Medicine Services, which will serve as the lead clinical site alongside other addiction treatment partners in Allegheny County. Antoine Douaihy, MD, Professor of Psychiatry and Medicine at the University of Pittsburgh School of Medicine and Senior Academic Director of Addiction Medicine Services at Western Psychiatric Hospital, said the trial examines the effectiveness and safety of GLP-1 receptor agonists and their impact on outcomes including employment, housing, health, and recidivism, noting that drug use disorders are major drivers of criminal justice involvement and intersect with systemic inequities and disparities. He expressed enthusiasm that the trial holds promise to change the landscape of addiction treatment and expand the recovery toolbox.

The research team itself is multidisciplinary by design, bringing together economists, addiction psychiatry specialists, and policy researchers from MIT, the University of Pittsburgh, the University of California San Diego, and UPMC Western Psychiatric Hospital. Named investigators include Diana Martínez Heredia of the Allegheny County Department of Human Services, Frank Schilbach and Dhruv Gaur of MIT, J. Travis Donahoe of the University of Pittsburgh, Gordon Dahl of the University of California San Diego, and Antoine Douaihy of UPMC. This composition allows the trial to be studied from clinical, economic, and behavioral angles simultaneously, an approach intended to produce evidence that can guide decisions well beyond Pittsburgh.

Indeed, scalability is a central theme of the trial’s design. Reville emphasized that because the intervention is a medication rather than a complex service program, a positive result in Pittsburgh should translate almost everywhere, making the trial easy for other cities to adopt. Jutca added that if results are positive, the county will consider the approach on an ongoing basis for individuals with substance use disorders and will work with other cities and counties across the country to disseminate findings, consistent with the broader goals of the Allegheny Recovery Partnership, which aims to build a coordinated system of care that keeps people engaged in treatment over time. Both leaders stressed that new medications become most valuable when integrated into systems that consistently support recovery.

The trial also builds on CASPR’s earlier pilot work, including a program with OpenDoors in Rhode Island that provides semaglutide to women in a post-incarceration sober housing program. Reville reported remarkable outcomes from that pilot, with women describing the disappearance of drug dreams and reporting that they felt better than they had ever felt, allowing them to rebuild careers and family relationships once cravings quieted. That program has drawn national media attention, including coverage in The New York Times and The Washington Post. Reville argued that GLP-1 medications may prove especially impactful because many patients are already familiar with the drug class and eager to take it, a level of patient enthusiasm he suggested could make a sustainable population-level impact possible and mark the most significant advancement in addiction medicine in decades.

Organizers are also attentive to the ethical dimensions of the work. Jutca said the team is working closely with community leaders and partners to ensure that trial participants are treated with respect throughout the process, noting that while addiction is widespread, its costs are unequally distributed. The county wants to demonstrate that widening inclusion in medical research is both important and possible while upholding high clinical and safety standards. CASPR, which funds and publishes research with the VA, created the first-ever pilot programs giving GLP-1 receptor agonists to patients receiving addiction treatment, and advocates for policies enabling faster medication development, is funded by individuals and foundations and accepts no direct or indirect funding from pharmaceutical companies. If the trial succeeds, its architects believe the result will be a major advance: a way to give people a better chance to recover, stay healthy, find employment, maintain stable housing, and avoid returning to the justice system.

Subject of Research: A randomized clinical trial testing whether semaglutide reduces alcohol and opioid use and justice involvement among adults with substance use disorders.

Article Title: CASPR and Allegheny County announce major new randomized clinical trial: GLP-1RA recovery and stability trial

Article References: CASPR and Allegheny County announce major new randomized clinical trial: GLP-1RA recovery and stability trial. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: semaglutide, GLP-1 receptor agonists, alcohol use disorder, opioid use disorder, randomized controlled trial, addiction medicine, recidivism, Allegheny County, CASPR, Arnold Ventures, criminal justice involvement, recovery

The review, led by Marisol Lila, Enrique Gracia, Cristina Expósito-Álvarez and Rocío Pérez-Sabater of the University of Valencia, draws together systematic reviews, meta-analyses and methodologically rigorous primary studies, with particular emphasis on findings from samples of known male perpetrators rather than community surveys alone. The authors argue that understanding individual-level psychological processes is essential for improving prevention and intervention, even though violence ultimately arises from interacting influences across multiple social and ecological levels. Their synthesis organizes the evidence into two broad domains: developmental and clinical vulnerabilities on one hand, and cognitive-affective drivers on the other.

In the first domain, the reviewers examine childhood trauma exposure, personality disorders, insecure attachment and substance use. Meta-analyses consistently show that men who experienced maltreatment in childhood, whether abuse, neglect or witnessing violence between their own parents, face elevated odds of becoming violent toward their partners in adulthood. The association is reliable but not deterministic; most survivors of childhood adversity never perpetrate violence. Studies of probation samples and batterer intervention programs further indicate that cumulative trauma predicts offending severity, and that post-traumatic stress symptoms among perpetrators are linked to treatment engagement difficulties and higher recidivism. Recent work also highlights the role of masculine discrepancy stress, the distress some men feel when they perceive themselves as falling short of masculine ideals, in linking childhood trauma to later violence through emotion dysregulation.

Personality pathology emerges as another robust thread. A meta-analysis in Clinical Psychology Review found that personality disorders, particularly borderline and antisocial configurations, predict intimate partner violence perpetration. Related research on narcissism and psychopathy shows that these traits not only raise the risk of violence but also forecast worse treatment outcomes, including dropout and repeated reassault among program participants. The classic typology of male batterers proposed by Holtzworth-Munroe and Stuart in the 1990s, which distinguished subtypes differing in severity and generality of violence, continues to find empirical support, with subtypes differing in treatment completion and recidivism. Attachment insecurity adds a further layer: meta-analytic work shows that both anxious and avoidant attachment dimensions are associated with perpetration, often operating through psychological distress, substance use and affect dysregulation.

Substance use occupies a prominent place in the review. Meta-analyses confirm that alcohol problems are among the most consistent correlates of partner violence, and event-based diary studies show that physical aggression is far more likely to occur on days when men drink. The reviewers describe mechanisms such as alcohol myopia, in which intoxication narrows attention to immediate provocations and impairs inhibitory control, while noting that antisocial personality traits and trait jealousy can moderate the alcohol-violence link. Illicit drug use, including marijuana and opioids, shows similar associations in samples of men arrested for domestic violence. Importantly, substance use also predicts treatment attrition, meaning that offenders with alcohol and drug problems are less likely to complete the very programs designed to change their behavior, although those who do complete can show meaningful gains.

The second broad domain concerns cognitive-affective drivers: cognitive dysfunctions, anger and hostility, empathy deficits, and sexist or otherwise violence-supportive attitudes. Neuropsychological research summarized in the review points to subtle but reliable deficits in executive functioning among perpetrators, including impulsivity, poor attention switching and impaired working memory, which together undermine the capacity to inhibit aggressive responses during conflict. Meta-analytic evidence on anger and hostility confirms that these emotional states are among the strongest psychological correlates of perpetration, and clinical studies show that anger disturbances predict long-term criminal recidivism. The reviewers emphasize shared mechanisms that cut across domains: emotion dysregulation, hostile attribution of intent to a partner’s behavior, and impaired inhibitory control appear repeatedly as pathways through which childhood trauma, personality pathology, attachment insecurity and intoxication are translated into violence.

Empathy and related social-cognitive capacities also feature prominently. Studies of court-mandated offenders reveal deficits in emotional decoding, perspective taking and empathic concern, and some work documents atypical facial responses to others’ suffering, with low sadness and even elevated happiness displayed toward victims’ distress. Alexithymia, the difficulty identifying and describing one’s own emotions, is associated with impulsivity and cognitive and empathic dysfunction in perpetrators and offers new targets for intervention. Neuroimaging studies add biological texture, linking coercive partner behavior to lower brain volume in emotion-regulation regions and altered resting-state connectivity in social mentalizing networks, although the authors caution that such findings describe correlates rather than proven causes.

Attitudes complete the cognitive picture. The review examines ambivalent sexism, distinguishing hostile sexism, which protects men’s power, from benevolent sexism, which guards traditional gender roles, and cites meta-analytic evidence linking both to violence-against-women attitudes and behaviors. Perpetrators commonly deploy denial, minimization, justification and victim-blaming, and validated scales now measure the perceived severity, acceptability and responsibility attribution surrounding intimate partner violence against women among both the general population and offenders. Implicit attitudes toward violence have even been shown to predict treatment outcomes, suggesting that conscious beliefs capture only part of the attitudinal architecture sustaining abuse.

Crucially, the reviewers stress that these psychological drivers show small-to-moderate but reliable associations with perpetration and predict clinically relevant outcomes, including violence severity, treatment engagement and dropout, and recidivism. No single factor is sufficient, and the drivers rarely operate in isolation; instead they interact through shared mechanisms. This convergence leads the authors to argue that psychological explanations should be understood as complementary to, not competing with, broader ecological accounts of intimate partner violence. The practical payoff is a case for assessment-guided, responsivity-based interventions tailored to each perpetrator’s individual risk profile, consistent with the risk-need-responsivity model of offender rehabilitation.

Evidence for such tailoring is already accumulating. Randomized controlled trials of the Strength at Home program, which addresses trauma-related emotional dysregulation, have reduced partner violence in military and civilian samples. Integrated treatments combining cognitive behavioral therapy with substance abuse care have shown efficacy for substance-dependent offenders, and motivational enhancement therapy for alcohol problems among partner-violent men has produced promising results. Pilot trials of cognitive training designed for perpetrators have improved cognitive flexibility and reaction time, while individualized motivational plans within batterer intervention programs have improved completion rates. The Valencia group’s own randomized trial of integrated motivational strategies for perpetrators with substance use problems adds further support. Together, these findings suggest that the era of one-size-fits-all batterer programs may be giving way to a more precise, psychologically informed science of intervention, one that treats the wounded boy behind the violent man not as an excuse for abuse, but as a clinical target in the service of safety for women.

Subject of Research: Psychological drivers of men's perpetration of intimate partner violence against women

Article Title: Psychological drivers of men’s perpetration of intimate partner violence against women

Article References: Lila, M., Gracia, E., Expósito-Álvarez, C., & Pérez-Sabater, R. (2026). Psychological drivers of men’s perpetration of intimate partner violence against women. Nature Reviews Psychology. https://doi.org/10.1038/s44159-026-00616-4

Image Credits: AI Generated

DOI: 10.1038/s44159-026-00616-4

Keywords: intimate partner violence, perpetrators, childhood trauma, personality disorders, attachment insecurity, substance use, anger and hostility, empathy deficits, sexism, emotion dysregulation, recidivism, batterer intervention programs