Substance Use and Criminal Offending: Associations, Mechanisms, and Limits
Author: Ukrainian Psychological Hub · Published: October 11, 2026 · Editorial Policy
Substance use and criminal offending overlap through several mechanisms that must be distinguished. Acute intoxication can change attention and judgment; sustained and expensive drug use can increase financial pressure for some people; illegal-market disputes can lead to violence; and shared developmental and social conditions can increase the likelihood of both substance use and offending. Criminal involvement may itself increase exposure to drugs. The Oxford Handbook's review of substance use and crime describes these pathways as heterogeneous and potentially reciprocal.
The strongest general conclusion is that no single relationship applies to every substance, offender, or offense. A person can use a psychoactive substance without committing any offense, commit an offense without using one, or have episodes where the two coincide for reasons that cannot be established from timing alone. A drug-positive test, an addiction diagnosis, and a criminal conviction answer different questions. Forensic explanation demands a precise mechanism and case-specific evidence.
What is the substance use–crime relationship?
Substance use means exposure to alcohol or other psychoactive drugs and can range from occasional consumption to a severe, persistent pattern. Intoxication means a substance is affecting a person in a particular period. A substance use disorder (SUD) is a clinical syndrome established using diagnostic criteria; it is not synonymous with use, a positive screen, withdrawal, or possession of a substance. Criminal offending refers to actions defined by the law of a jurisdiction. Arrest, conviction, incarceration, and self-reported offenses are distinct measures; one cannot automatically substitute for another.
The expression “drug-related crime” can refer to at least five phenomena: drug-law violations such as unlawful possession; crimes committed during intoxication; crimes committed to finance consumption; violence in illegal supply networks; and criminal acts statistically associated with substance use because of shared social or personal factors. Researchers need to state which meaning they intend. The U.S. National Institute of Justice's report on how much crime is drug-related demonstrates the sensitivity of estimates to attribution rules.
What does the research show?
The association is well documented, but causality varies by mechanism
Evidence comes from prison surveys, population cohorts, observational studies, laboratory experiments, treatment trials, and comparative policy evaluations. Survey evidence can identify co-occurrence but usually cannot prove what caused what. Longitudinal designs improve temporal inference but can leave unmeasured confounding. Randomized laboratory studies can isolate acute effects on ethically permissible proxies such as aggressive responses, while the behavior of interest in the justice system may be a much more complex real-world act. Treatment trials estimate outcomes in selected participants and with specific follow-up intervals.
A foundational framework is Goldstein's tripartite model, which distinguishes psychopharmacological, economic-compulsive, and systemic pathways linking drugs to violence. The model is a research framework for identifying potential connections; it is not a psychological diagnostic instrument or a method for determining legal guilt. Later NIJ work adds important questions about victimization, market context, reciprocal effects, and incomplete crime data.
The meaning and limits of U.S. prison statistics
In the U.S. Bureau of Justice Statistics' Survey of Prison Inmates, 2016, published in 2021, 31% of people in state prisons and 25% in federal prisons reported drinking alcohol at the time of the offense for which they were imprisoned. Drug use at the time of the offense was reported by 39% of state and 31% of federal prisoners. Among those eligible for a prior-year disorder measure, roughly 49% of state and 32% of federal prisoners met DSM-IV substance-use-disorder criteria.
These percentages describe a selected incarcerated population and rest on retrospective reports. They are not rates of criminality among substance users, probabilities that intoxication caused an offense, or estimates of future danger. They do not generalize directly to all arrested people, people convicted but not imprisoned, or community samples. The distinction is a conditional-probability problem: the prevalence of drug use among prisoners does not equal the probability of imprisonment among people who use drugs. The underlying BJS survey design makes the sampling frame explicit.
Mechanism 1: Acute psychopharmacological effects
Intoxication can change how people perceive provocation, allocate attention, interpret threats, regulate emotional responses, and evaluate consequences. The strongest controlled experimental evidence concerns alcohol. The relevant account is conditional: in settings containing provocation, aggressive cues, social conflict, or impaired situational judgment, alcohol can increase aggressive responding for some participants. A meta-analytic review of experimental explanations supported indirect cognitive mechanisms and found stronger effects of provocation in intoxicated participants.
In a meta-analysis of alcohol administration and aggression among women, the pooled experimental effect was small but statistically significant (d = 0.17). A 2025 meta-analysis of experimental studies on alcohol and sexual-aggression measures found a small pooled effect (g = 0.25). These outcomes are laboratory measures and behavioral proxies. They do not establish the probability of assault, sexual violence, or conviction for an identifiable individual. They also do not show that every instance of drinking creates aggression.
Alcohol's effects depend on dose, timing, drinking history, expectations, immediate social cues, prior behavior, available weapons, and opportunities for de-escalation. A research participant responding aggressively in an experimental task is not equivalent to a defendant who committed a legally defined offense. Even when intoxication contributed to an event, responsibility is a legal determination made under the relevant jurisdiction's rules rather than a direct implication of a psychological effect.
How evidence differs for stimulants, cannabis, opioids, and sedatives
Researchers should avoid treating all psychoactive drugs as one pharmacological category. A review comparing alcohol and stimulants found a clearer controlled experimental basis for an acute causal alcohol–aggression link than for cocaine or amphetamines. Observational stimulant–violence associations may be influenced by polydrug exposure, sleep deprivation, social networks, prior violence, market context, and psychiatric symptoms. Those issues require substance-specific evidence and an assessment of alternative explanations.
Cannabis use, cannabis-related impairment, possession offenses, and criminal violence are different outcomes. A change in legal status can alter possession arrests without changing interpersonal aggression. Opioids often involve distinct pathways related to drug markets, economic precarity, and treatment interruption; the existence of these pathways does not make opioid intoxication a universal direct cause of assault. Sedatives can impair cognition and coordination, but individual conclusions require details of dose, timing, combinations, medical use, and behavior.
A positive toxicology test may indicate exposure over a detection window rather than impairment at the moment of an event. In forensic work, laboratory analysis, observed behavior, dose information, symptom timing, witness reports, and the law's specific evidentiary requirements must be evaluated separately. Combining drugs can further obscure which exposure, if any, had a causal role.
Mechanism 2: Economic-compulsive offending
The economic-compulsive hypothesis addresses situations where the cost of a sustained pattern of drug use creates financial strain, and a person responds through income-generating crimes such as theft, burglary, fraud, or robbery. This pathway depends on substance cost, dependence severity, access to work and housing, available assistance, prior offending, and the alternatives the person perceives. It does not imply that people with addiction generally steal or commit violence.
A classic review of narcotics and crime reported increases in property offending around addiction onset in some populations, but also found that the first arrest often preceded addiction. Comparisons of periods before and during treatment sometimes showed reduced property offending. This pattern is compatible with substance use intensifying an existing trajectory for some individuals while being less relevant or unrelated for others. The interpretation depends on how offending, addiction, and treatment participation were measured.
Robbery and theft should remain distinct: robbery involves legal elements related to force or threat that require additional explanation. Similarly, illegal possession or supply is a different category from theft to finance consumption. A statutory offense directly defined by controlled-substance law tells us little by itself about the psychological causes of interpersonal violence.
Mechanism 3: Violence within illegal drug markets
The systemic pathway concerns how illegal markets operate when participants cannot routinely use lawful contracts or conventional dispute-resolution institutions. Territorial conflicts, debt collection, theft within distribution chains, retaliation, and coercive enforcement can contribute to violence even when the perpetrator is not intoxicated. This is a mechanism involving social organization, incentives, and access to weapons, not simply a drug's action on the nervous system.
A study of 1988 New York City homicides cataloged by NIJ identified systemic violence as the principal category among drug-related homicides in its selected cases. The study illustrates the model rather than supplying an estimate applicable to all years and places. Local trafficking organization, policing, market volatility, and firearms may substantially change the pattern across jurisdictions.
The same market conditions can increase victimization and coercion of consumers, sellers, residents, and bystanders. A person found in a drug-trading environment may be a victim or witness rather than an offender. Reliable classification requires reconstructing the event, the participant's role, and the mechanism supported by available records.
Mechanism 4: Shared risks, social context, and reciprocal pathways
Some predictors of substance use are also associated with offending: exposure to early adversity, unstable socioeconomic circumstances, school disengagement, victimization, certain impulsive behaviors, peer networks, and opportunities for rule breaking. These factors interact and vary across the life course. A risk factor is a group-level association, not a criminal motive, deterministic cause, or individual prediction. Individuals with these experiences follow many different trajectories.
A four-wave longitudinal study involving 1,760 adolescents examined paths through drug use, delinquent peers, criminal thinking, victimization, and later serious offending. Its results support the plausibility of temporally ordered social and cognitive pathways, but observational mediation cannot exclude all confounders or establish what would have happened to a specific participant under a different exposure. Cross-lagged and longitudinal patterns should be interpreted within the designs' assumptions.
Offending may also increase later substance use through peer exposure, stress, market involvement, or the effects of justice contact. Researchers therefore ask whether drug use came before the first offense, whether both escalated after another event, and whether changes in substance use track changes in offending within the same person. Periods of abstinence with continued offending, or periods of use without offending, are especially informative when evaluating simplistic explanations.
Different crimes, different explanatory requirements
Assault, threats, and homicide
In violent offenses, acute alcohol effects may interact with provocation, relationship history, weapons, and social setting, while illegal-market violence may follow a separate trajectory. Toxicology findings from a victim or suspect cannot identify which mechanism operated. Homicide attribution in particular requires accurate information on event sequence, prior interactions, and credible causal links. A substance-related setting does not automatically make an event psychopharmacologically caused.
Property offending and income generation
For theft, burglary, fraud, and some robberies, economic pressure from expensive dependent use can be relevant, but a longitudinal history and possible alternative motives remain essential. Financial instability and opportunity also affect people who do not use substances. Researchers must avoid classifying every offense committed by a person with SUD as “addiction driven.”
Drug-law offenses and impaired driving
Possession and trafficking offenses depend partly on how law defines prohibited behavior; changing law or enforcement priorities can change recorded drug offenses. Impaired driving requires a different evidence base involving substance effects on driving and specific legal standards. Blood concentrations, observed impairment, driving performance, and statutory thresholds have distinct roles. These rules vary by jurisdiction and cannot be inferred from a general study of drug use and crime.
Domestic abuse and sexual violence
Alcohol may elevate risk in some interpersonal contexts, but intoxication does not replace the investigation of coercive conduct, patterns of abuse, consent, victim safety, and legal elements. The experimental literature on sexual-aggression proxies can inform prevention research without explaining every criminal case. A victim's own substance use never supplies a psychological explanation that excuses another person's violence.
Victimization belongs in the model
Substance-related environments can expose people to theft, assault, exploitation, and coercion, and victimization can in turn influence substance use. Thus a sample with high drug involvement may reflect increased perpetration, victimization, police contact, or combinations of these events. The NIJ research agenda identifies victimization as a critical missing component in simplistic drug–crime accounts. A 2024 international review documents harms experienced from other people's alcohol consumption, reinforcing the need to consider affected persons beyond the consumer.
How can studies establish a causal connection?
A causal argument requires temporal order, a plausible mechanism, an appropriate counterfactual comparison, careful measurement, and evaluation of alternative explanations. Cross-sectional correlations cannot settle whether drugs influenced an offense, offending preceded drug use, or a shared factor influenced both. Longitudinal cohorts can test developmental ordering; within-person studies can compare higher- and lower-use periods; natural experiments may assess policy changes; randomized treatment trials can estimate intervention effects in the participants included. Every method has scope limitations.
Arrest, rearrest, reconviction, reincarceration, self-reported behavior, and violent incidents are not interchangeable outcomes. They reflect different combinations of conduct, reporting, policing, prosecution, and legal procedure. Similarly, a psychometric score is not a forensic conclusion, and a forensic conclusion is not a legal verdict. Study quality depends on transparent definitions, appropriate control variables, follow-up time, attrition, and the degree to which the results transport to another population.
An important statistical error is reversing conditional probabilities: “many incarcerated people reported drug use” tells us nothing directly about the proportion of drug users who will be incarcerated. Another is equating an average association with an individualized forecast. Prediction requires a validated model for the target population and outcome, known base rates, discrimination and calibration evidence, and accounting for false positives and false negatives. Even then, uncertainty remains substantial at individual level.
Substance use in forensic assessment and legal decisions
Forensic assessment should begin with the exact referral question. The clinician may evaluate intoxication history, substance-related symptoms, functioning, collateral records, timing of alleged conduct, and relevant treatment records. Toxicology can provide biological evidence with specific detection limitations. Screening scores reflect responses on an instrument; diagnosis reflects clinical criteria; risk estimates describe probabilistic associations; forensic opinions synthesize evidence for a legally relevant question. Competency, capacity, criminal responsibility, and guilt are governed by applicable law.
The APA Specialty Guidelines for Forensic Psychology call for competence, impartiality, attention to data quality, and disclosure of the limits of conclusions. A professional should not derive a legal conclusion solely from a SUD diagnosis, positive test, or statistical base rate. Instruments developed for general reoffending must not be repurposed as proof of violent intent or guilt. Cultural fairness, differential enforcement, validation samples, calibration, and protective factors become especially important when risk estimates influence liberty or services.
Mental health and criminality require separate analysis. A person may have a co-occurring psychiatric condition, substance use disorder, both, or neither; none of these categories establishes offending. The live English Hub article Schizophrenia and Violence: What Research Really Shows addresses diagnosis-specific evidence. This page retains ownership of the substance-use–offending interface rather than making broad claims about psychiatric diagnosis.
Does addiction treatment reduce criminal offending?
Treating substance use disorders has clinical and public health benefits. Whether an intervention reduces crime depends on the type of treatment, comparison group, participant characteristics, justice-system context, and definition of recidivism. Improved treatment retention, reduced use, fewer arrests, and fewer convictions are separate outcomes. Associations between completing treatment and lower reoffending can arise partly because people more able to complete a program have different starting circumstances.
Medication treatment for opioid use disorder
Methadone and buprenorphine have established clinical value. A 2023 Lancet Psychiatry systematic review and meta-analysis00095-0) compared these treatments on retention and other outcomes, including criminal-justice measures. A meta-analysis of medication treatment in prison and jail found that methadone increased engagement with community treatment and reduced illicit opioid use after release; the pooled randomized evidence did not demonstrate a recidivism reduction (OR 0.93, 95% CI 0.51–1.68). The absence of a clear rearrest effect does not negate these clinical benefits.
A randomized New England Journal of Medicine trial of extended-release naltrexone in selected justice-involved adults found lower opioid-relapse rates during treatment than usual care (43% versus 64%), while the difference waned after treatment stopped. Reincarceration was not significantly reduced. Selection of participants already abstinent at baseline and willing to use that treatment limits generalization to all people with opioid use disorder or all correctional settings.
After release from custody, overdose risk, disrupted care, housing challenges, and interrupted medications can coincide. SAMHSA guidance on reentry emphasizes continuity of medication, treatment, coverage, housing, and services. Such health and recovery benefits merit support independently of any unproven assertion about a specific recidivism outcome.
Drug courts, diversion, and integrated services
A meta-analysis covering 154 drug-court evaluations reported reductions in average recidivism for adult drug courts, with weaker or less consistent effects in some other court types. The commonly cited change equivalent to 50% versus 38% rearrest is an illustrative pooled comparison, not a guarantee for a jurisdiction. Evaluation rigor, participant selection, completion rates, eligibility, and sanctions can affect results. A Cochrane review of medication interventions for drug-using offenders also found small trial numbers and limits to generalization; all pharmacological interventions should not be treated as equivalent.
Effective responses may combine addiction care with housing, employment, social support, and coordinated reentry. Each component should be assessed for clinical benefits, well-being, victim protection, and lawful justice outcomes. Treatment needs should be identified clinically, and access to care should not depend on assuming a patient will commit a crime.
Prevention and public policy
Prevention must match the mechanism. Acute alcohol-related harm can be addressed through evidence-informed public health and environmental strategies; illegal-market violence requires attention to market organization and community safety; economic pressure can be addressed through treatment and material stability; developmental risks require longer-term support. The Community Preventive Services Task Force recommends regulating alcohol outlet density to reduce excessive consumption and related harm, while acknowledging limitations in the studies.
Arrest trends also reflect what is illegal and how law is enforced. A fall in possession arrests after decriminalization need not imply a decline in assaults, and a rise in trafficking prosecutions need not imply an equivalent rise in drug consumption. Policy claims should specify jurisdiction, legal change, intervention fidelity, comparison area, public-health outcome, violent-offense measure, and time frame. Cross-jurisdiction generalizations require caution.
What the evidence cannot establish
A clinical diagnosis is not an explanation of criminal behavior. A risk factor is not proof of causation. A relative risk from a population study is not an individual's probability of offending. A toxicology result is not proof of motive or intent. A risk-assessment score is not evidence of guilt, and an assessment instrument can have false positives, false negatives, biased inputs, and limited transportability across groups and jurisdictions. Finally, improvements in addiction-related health do not automatically imply statistically demonstrable reductions in criminal recidivism.
For ethical and scientific reasons, forensic conclusions should separate documented behavioral observations, clinical assessment, test results, probabilistic interpretations, competing hypotheses, and legal standards. This distinction protects public safety and ensures that people with substance use disorders are not stereotyped as criminals.
Frequently asked questions
Does substance abuse cause crime?
Substance use can contribute causally to particular offenses through acute effects, financial demands, or involvement in illegal markets. In other cases the association arises from shared conditions or reverse influence. The term “substance abuse” is broad and imprecise; use, intoxication, harmful patterns, and substance use disorder should be distinguished.
Is alcohol more connected to violence than other drugs?
Alcohol has comparatively strong experimental evidence for raising some forms of aggressive response under controlled conditions. That does not rank all drug classes by general criminality, and laboratory aggression is not equivalent to a particular real-world offense.
Do people with SUD inevitably commit crimes?
No. Clinical diagnosis has neither the purpose nor the accuracy of predicting individual criminal acts. Exposure, context, age, prior behavior, legal rules, treatment, and protective factors all affect observed associations.
Can drug use explain murder or assault?
Sometimes a substance-related mechanism may contribute, but a forensic explanation requires case-specific evidence and consideration of other mechanisms. A positive drug test or evidence of consumption near the event is insufficient by itself.
Does addiction cause theft?
In some cases costly dependent use contributes to income-generating offending. Many people with addiction do not steal, while property crime may also occur before dependence or continue after substance use changes. Evidence about timing and motive is essential.
Do drug tests establish criminal intent?
No. A test may provide evidence about exposure, subject to specimen-specific detection windows and procedures. Intent and the legal elements of an offense require separate evidence.
Does opioid treatment prevent rearrest?
Some programs may improve criminal-justice outcomes, but studies consistently show clearer effects on opioid use and care engagement than on rearrest or reincarceration. Results depend on treatment type, follow-up, and population.
What is the role of a forensic psychologist?
To evaluate a defined psycholegal question using appropriate methods, evidence, alternative explanations, and clear limitations. Legal responsibility and verdicts are determined under applicable law by authorized decision-makers.
Conclusion
The link between substance use and criminal offending is a network of causal possibilities, social conditions, and legal classifications. Alcohol-related intoxication, economic pressure, illegal markets, victimization, shared vulnerabilities, and reciprocal trajectories can each matter in different circumstances. Scientific evaluation proceeds by identifying the mechanism, defining the offense, comparing explanations, and stating the limits of inference. Clinical care, lawful investigation, and evidence-based prevention benefit from keeping these distinctions explicit.
Related Articles
References
Alcohol consumption and sexual aggression perpetration: A meta-analytic review. (2025). Peer-reviewed meta-analysis, PMID 41066300.
American Psychological Association. (2013). Specialty guidelines for forensic psychology.
Brownstein, H. H., Baxi, H. R. S., Goldstein, P. J., & Ryan, P. J. (1992). Relationship of drugs, drug trafficking, and drug traffickers to homicide. National Institute of Justice.
Bureau of Justice Statistics. (2021). Maruschak, L. M., Bronson, J., & Alper, M. Alcohol and drug use and treatment reported by prisoners: Survey of Prison Inmates, 2016. U.S. Department of Justice.
Caulkins, J. P., & Kleiman, M. A. R. (2014). How much crime is drug-related? National Institute of Justice.
Crane, C. A., et al. (2017). The proximal effects of acute alcohol use on female aggression. Psychology of Addictive Behaviors, 31, 21–26. https://doi.org/10.1037/adb0000244.
Goldstein, P. J. (1985). The drugs/violence nexus: A tripartite conceptual framework. Journal of Drug Issues, 15, 493–506. https://doi.org/10.1177/002204268501500406.
Intoxicated aggression: Do alcohol and stimulants cause dose-related aggression? A review. (2018). Peer-reviewed review, PMID 29941239.
Kilian, C., et al. (2024). National and regional prevalence of interpersonal violence from others' alcohol use: A systematic review and modelling study. PubMed.
Lee, J. D., et al. (2016). Extended-release naltrexone to prevent opioid relapse in criminal justice offenders. New England Journal of Medicine, 374, 1232–1242. https://doi.org/10.1056/NEJMoa1505409.
Medication treatment for opioid use in prison and jail settings: A meta-analysis and systematic review. (2019). Peer-reviewed meta-analysis.
Mitchell, O., Wilson, D. B., Eggers, A., & MacKenzie, D. L. (2012). Assessing the effectiveness of drug courts on recidivism. Journal of Criminal Justice, 40, 60–71. Journal page.
National Institute of Justice. (2003). Toward a drugs and crime research agenda for the 21st century.
Perry, A. E., et al. (2015). Pharmacological interventions for drug-using offenders. Cochrane Database of Systematic Reviews. PubMed.
Santo, T., et al. (2023). Buprenorphine versus methadone for the treatment of opioid dependence. The Lancet Psychiatry, 10, 386–402. https://doi.org/10.1016/S2215-0366(23)00095-000095-0).
Speckart, G., & Anglin, M. D. (1985). Narcotics and crime: An analysis of existing evidence for a causal relationship. Behavioral Sciences & the Law, 3, 259–282. https://doi.org/10.1002/bsl.2370030304.
Substance Abuse and Mental Health Services Administration. (2026). Intercept 4: Reentry.
The Community Preventive Services Task Force. Regulation of alcohol outlet density.
White, H. R. (2014). Substance use and crime. The Oxford Handbook of Substance Use and Substance Use Disorders, Volume 2. https://doi.org/10.1093/oxfordhb/9780199381708.013.004.