Criminal Behavior

Abstract

Criminal behavior is a form of behavior that violates the criminal law, studied in cognitive and clinical psychology as the product of individual differences, developmental history, situational decision making, and neurobiology rather than of a single cause. Its most durable organizing finding is developmental: a small group of life-course-persistent offenders, whose antisocial conduct begins early and is rooted in neurocognitive deficits and adverse rearing, is distinct from a far larger group of adolescence-limited offenders whose law-breaking is common, transient, and tied to the gap between biological and social maturity (Moffitt, 1993). Alongside this taxonomy sit personality, low-self-control, rational-choice, and neurocriminological accounts, each carving the legal category differently. The field now couples these explanatory traditions to actuarial risk prediction and to rehabilitation programs whose effects on reoffending can be measured in randomized trials (Beaudry, Yu, Perry, & Fazel, 2021).

Keywords: antisocial behavior, offender decision making, developmental taxonomy

Why some people break the law and most do not is among the oldest questions in behavioral science, and the modern answer is deliberately plural. Criminal behavior is not one phenomenon with one explanation; it is a legal category that collects conduct arising from many interacting sources. A teenager who shoplifts once, an adult with a lifelong record of violence, and a first-time drunk driver occupy the same statute but not the same causal pathway. Cognitive psychology contributes by asking what in perception, judgment, self-control, and learning distinguishes those pathways, and by testing which distinctions actually predict who offends, who persists, and who can be helped to stop.

What Criminal Behavior Is

Criminal behavior is conduct prohibited by the criminal law and subject to its sanction. That definition is legal, not psychological, and the distinction matters: the same act may be a crime in one jurisdiction or era and not another, so the category is defined from the outside rather than by any shared mental signature. MeSH classifies the descriptor as a form of behavior, and psychology treats it as a heterogeneous outcome to be explained by more basic processes rather than as a unitary trait. What the scientific literature studies under this heading is therefore the set of individual, developmental, situational, and biological factors that raise or lower the probability of law-breaking, and the degree to which those factors generalize across offense types.

Because the category is legal, the first task of a psychological account is to carve it into kinds that do behave lawfully. The most influential carving is developmental. Moffitt (1993) argued that the well-known rise and fall of offending across adolescence hides two populations: a numerous group whose antisocial behavior is confined to the teenage years, and a small group whose antisocial behavior is stable from early childhood into adulthood. Other traditions carve the category differently. Eysenck (1996) located the predisposition in personality, arguing that people high in psychoticism, extraversion, and neuroticism condition poorly to social sanction and so acquire a weaker internalized conscience. Gottfredson and Hirschi (1990) reduced the common element across offenses to a single stable trait, low self-control, established early by parenting and manifest thereafter in impulsive, risk-seeking, short-sighted conduct of which crime is only one expression.

Types of Criminal Behavior

Within the Medical Subject Headings vocabulary, the descriptor Criminal Behavior is a form of behavior and heads a small set of narrower descriptors. These do not partition offending into a clean taxonomy of psychological kinds; MeSH is an indexing vocabulary, so its subtypes name the specific law-breaking categories the biomedical literature most often indexes rather than branches of one underlying mechanism. The two direct subtypes below are better read as salient instances of criminal behavior than as an exhaustive classification of it.

Developmental Trajectories

The single most replicated fact about crime is the age-crime curve: offending rises sharply in early adolescence, peaks in the late teens, and falls steeply through the twenties. Moffitt (1993) argued that this aggregate curve is the sum of two very different trajectories. Life-course-persistent offenders are few, begin showing conduct problems in early childhood, and carry neurocognitive deficits, especially in verbal and executive function, that interact with disadvantaged rearing to produce stable antisocial behavior across the lifespan. Adolescence-limited offenders are many, begin and end their law-breaking in the teenage years, and offend because a maturity gap between biological adulthood and the delayed grant of adult status makes delinquency briefly functional. The two groups look alike at the peak of the curve but differ sharply before and after it.

The persistent group is small, so it barely shifts the teenage peak, which the far more numerous adolescence-limited group dominates. Yet because persistent offending does not desist, almost all offending that remains in adulthood comes from it — which is why the same aggregate curve hides two very different populations.

The taxonomy has held up and been refined. Reviewing twenty-five years of evidence, Moffitt (2018) confirmed that the early-onset group shows the predicted long-run burden of poor health, violence, and economic difficulty, while also recognizing a third, smaller pattern of adult-onset offending the original model did not anticipate. The developmental account also absorbs the gene-environment work: Caspi and colleagues (2002) found that a functional polymorphism in the MAOA gene moderated the effect of childhood maltreatment on later antisocial behavior, so that maltreated children with the low-activity variant were far more likely to develop conduct problems, an early and influential demonstration that neither genes nor adversity acts alone. Social cognition supplies a complementary mechanism: Dodge and Coie (1987) showed that aggressive children disproportionately interpret ambiguous social cues as hostile, a biased reading of intent that makes a retaliatory, reactive aggression feel justified.

The Decision to Offend

Developmental and trait accounts explain standing differences in risk; they do not by themselves explain why an at-risk person offends here and not there, now and not then. The rational-choice perspective fills that gap by treating each offense as a situated decision. Cornish and Clarke (1986) modeled the offender as a reasoning agent who weighs the anticipated rewards of a specific crime against its perceived costs, including the probability and severity of punishment, under the constraints of time, information, and ability. The model is deliberately modest about rationality: the appraisal is bounded and offense-specific, not a global cost-benefit calculus, which is why situational prevention that raises the effort or the perceived risk of a particular crime can deter it without changing anything about the offender.

Raising certainty discounts the whole reward and scales the cost at once, so a small rise in the chance of being caught moves the net value far more than an equal proportional rise in severity. That asymmetry is the formal reason the certainty of punishment deters more reliably than its harshness.

A robust empirical regularity falls out of this framework: the certainty of sanction deters more reliably than its severity. A modest penalty that is very likely to follow suppresses offending better than a harsh penalty that rarely does, because the expected cost is dominated by the probability term when offenders discount the future steeply. The rational-choice account also meshes with the control tradition: Gottfredson and Hirschi (1990) argued that people low in self-control systematically underweight the delayed, uncertain costs of crime relative to its immediate rewards, so the same situational calculus yields different decisions depending on a stable individual difference in how steeply future consequences are discounted.

Prediction, Risk, and the Brain

If offending has identifiable causes, it should be partly predictable, and the applied science of crime is largely about prediction and its uses. Structured actuarial instruments that tally empirically validated risk factors predict violence and reoffending more accurately than unstructured clinical judgment. Reviewing seventy-three samples covering nearly 25,000 people, Fazel and colleagues (2012) found that such instruments identify low-risk individuals well but are considerably weaker at confirming that a high-risk flag will be followed by an actual offense, a limit that bears directly on how their output may fairly be used. The risk-need-responsivity model turns prediction into intervention: Andrews, Bonta, and Wormith (2006) distinguished static risk markers from dynamic criminogenic needs, changeable factors such as antisocial attitudes and associates, and argued that treatment reduces reoffending most when its intensity matches risk and its targets are those dynamic needs.

The Big Four carry the greatest predictive weight; the remaining four add moderately. A higher tier means a higher group-level probability of reoffending, not a verdict on any individual. Crucially, every factor here except criminal history is a dynamic criminogenic need — a changeable target that well-matched treatment can move.

Prediction also runs through individual clinical constructs and through the brain. Psychopathy, a configuration of callousness, shallow affect, and impulsive antisociality measured by Hare's checklist, is among the stronger single predictors of serious and persistent violence (Hare & Neumann, 2008). Neurocriminology seeks the biological substrate of these patterns: Glenn and Raine (2014) reviewed evidence relating antisocial and violent behavior to structural and functional differences in prefrontal cortex, amygdala, and related circuitry, and Raine (2019) proposed a neuromoral theory in which impaired function across a prefrontal-limbic moral network underlies the affective and regulatory deficits seen in antisocial and psychopathic individuals. The forensic application of profiling, by contrast, has fared poorly: a meta-analysis of forty years of research found little evidence that offender profiles predict an offender's characteristics with useful accuracy (Fox & Farrington, 2018).

Worked Example

Consider the rational-choice calculus for a property offense with an anticipated gain of $1,000. Suppose the offender perceives a 10% chance of being caught and, if caught, a sanction whose subjective cost is $8,000. Treating the decision as an expected-value appraisal, the expected payoff of offending is the probable gain minus the probable cost: E = (1 − 0.10) × 1000 − 0.10 × 8000 = 900 − 800 = +100. The expected value is positive, so the model predicts the offense proceeds.

Now change only the certainty of sanction, doubling the perceived probability of capture to 20% while leaving the severity untouched: E = (1 − 0.20) × 1000 − 0.20 × 8000 = 800 − 1600 = −800. The expected value is now strongly negative and the model predicts desistance. Contrast this with doubling severity instead, raising the sanction cost to $16,000 while holding capture at the original 10%: E = 900 − 0.10 × 16000 = 900 − 1600 = −700. Both changes deter, but note that the certainty increase and a severity increase of comparable deterrent effect here required doubling different quantities, and in populations that discount the future steeply the probability term dominates, which is the formal reason certainty of punishment deters more reliably than severity (Cornish & Clarke, 1986).

Discussion

The study of criminal behavior is unusual among cognitive-science topics in that its theories must coexist rather than compete for a single victor. Developmental taxonomy, personality, control, rational choice, and neurocriminology each capture a real part of the variance, and the honest synthesis is biopsychosocial: genetic and neural vulnerabilities are expressed through cognitive and personality differences, which interact with developmental adversity and are triggered or restrained by immediate situations. The gene-environment interaction work (Caspi et al., 2002) is the template for that integration, showing that the effect of adversity depends on biology and vice versa, so no single level is privileged.

The field also carries an unusual ethical load, because its findings feed directly into decisions that deprive people of liberty. Risk instruments predict group probabilities, not individual destinies, and their asymmetric accuracy, good at identifying low risk, weaker at confirming high risk (Fazel et al., 2012), means that acting on a high-risk flag as though it were a certainty will often be wrong. Neurocriminological evidence raises the parallel worry that a neural correlate of antisociality can be read either as grounds for treatment or as grounds for indefinite control. The rehabilitative tradition pushes back by insisting that criminogenic needs are changeable and that intervention, not prediction alone, is the point: Ward and Brown (2004) argued that the good lives model, which builds an offender's capacity to secure legitimate goods, complements risk management by giving desistance something to move toward rather than only something to avoid.

Current Directions

The most consequential recent advance is that the effect of rehabilitation on reoffending is now estimated from randomized evidence rather than observational comparison. Beaudry, Yu, Perry, and Fazel (2021) meta-analyzed randomized controlled trials of psychological interventions delivered in prison and found a modest but genuine reduction in recidivism, while highlighting how few well-controlled trials exist and how much heterogeneity they contain, a candid appraisal that sets the agenda for better trials rather than louder claims. The developmental program has likewise matured: Moffitt's (2018) long-run follow-up both confirmed the heavy adult burden borne by early-onset offenders and complicated the two-group picture with evidence for adult-onset pathways, so the taxonomy is being revised from within rather than abandoned.

Neurocriminology is the third active front. Raine's (2019) neuromoral theory is an attempt to move from a catalogue of neural correlates toward a mechanistic account of how a specific moral-processing network fails, which, if it holds, would connect the biological findings to the social-cognitive ones, such as the hostile attribution bias (Dodge & Coie, 1987), within a single circuit-level story. Across all three fronts the direction of travel is the same: from describing who offends toward specifying the mechanisms that generate offending and the interventions that measurably change them.

Common Misconceptions

Criminal behavior is a single trait people either have or lack.

It is a legal category that collects causally distinct pathways. The developmental taxonomy alone separates a small life-course-persistent group from a large adolescence-limited one, and the two differ in onset, mechanism, and outcome (Moffitt, 1993).

Harsher punishment is the most effective deterrent.

In the rational-choice model the expected cost of offending is dominated by the probability of sanction, not its magnitude, so the certainty of being caught deters more reliably than severity (Cornish & Clarke, 1986).

Genes or environment, one or the other, causes crime.

The effect of childhood maltreatment on later antisocial behavior depends on genotype: maltreated children with a low-activity MAOA variant are far more likely to develop conduct problems, a gene-environment interaction in which neither factor acts alone (Caspi et al., 2002).

Risk-assessment tools can tell which individual will reoffend.

They estimate group-level probabilities and are better at ruling out low risk than at confirming that a high-risk person will actually offend, so a high-risk flag is not an individual prediction (Fazel et al., 2012).

Glossary

Adolescence-limited offending.

Antisocial behavior confined to the teenage years, common and transient, attributed to the gap between biological and social maturity.

Age-crime curve.

The robust pattern in which recorded offending rises in early adolescence, peaks in the late teens, and declines through adulthood.

Antisocial behavior.

Conduct that violates the rights of others or major social norms; the broad psychological category of which much criminal behavior is an instance.

Criminogenic need.

A dynamic, changeable risk factor for reoffending, such as antisocial attitudes or associates, that rehabilitation can target.

Desistance.

The process of ceasing to offend, whether through maturation, changed circumstances, or intervention.

Deterrence.

The suppression of offending by the anticipated cost of sanction, more strongly driven by the perceived certainty of punishment than by its severity.

Gene-environment interaction.

A pattern in which the effect of an environmental factor on behavior depends on genotype, or vice versa, rather than the two acting independently.

Good lives model.

A rehabilitation framework that reduces reoffending by building an offender's capacity to achieve legitimate personal goods, complementing risk management.

Hostile attribution bias.

The tendency, pronounced in aggressive individuals, to interpret ambiguous social cues as signs of hostile intent.

Life-course-persistent offending.

Stable antisocial behavior beginning in early childhood and continuing into adulthood, linked to neurocognitive deficit and adverse rearing.

Neurocriminology.

The study of the neural and biological bases of antisocial and criminal behavior.

Psychopathy.

A clinical configuration of callousness, shallow affect, and impulsive antisociality measured by the Psychopathy Checklist and strongly predictive of serious violence.

Rational choice theory.

The view that an offense is a bounded, offense-specific decision weighing anticipated rewards against perceived risks and costs.

Recidivism.

The reoffending of a person who has already been sanctioned for a crime; a primary outcome measure for interventions.

Risk-need-responsivity (RNR).

A model of offender assessment and treatment matching intervention intensity to risk and targeting dynamic criminogenic needs.

Self-control theory.

The claim that a single stable trait of self-control, established in childhood, underlies the propensity to crime across its many forms.

Key Researchers

Kenneth A. Dodge

(living). William McDougall Distinguished Professor at the Sanford School of Public Policy, Duke University; originator of the social-information-processing model of aggression and the hostile attribution bias. Faculty page - Wikipedia

Hans J. Eysenck

(1916-1997). Psychologist at the Institute of Psychiatry, University of London; author of the personality theory of criminality linking offending to psychoticism, extraversion, and neuroticism. Wikipedia

David P. Farrington

(1944-2024). Professor of psychological criminology at the University of Cambridge; director of the Cambridge Study in Delinquent Development and a central figure in developmental and life-course criminology. ORCID - Wikipedia

Robert D. Hare

(living). Emeritus professor of psychology at the University of British Columbia; developer of the Psychopathy Checklist (PCL-R) and a leading authority on psychopathy. Faculty page - Wikipedia

Cesare Lombroso

(1835-1909). Italian criminologist at the University of Turin and founder of the positivist school; proponent of the now-discredited born-criminal theory, historically pivotal in making crime an object of empirical study. Wikipedia

Terrie E. Moffitt

(living). Nannerl O. Keohane University Professor at Duke University and professor at King's College London; author of the adolescence-limited and life-course-persistent developmental taxonomy of antisocial behavior. ORCID - Faculty page - Wikipedia

Adrian Raine

(living). Richard Perry University Professor of Criminology, Psychiatry, and Psychology at the University of Pennsylvania; founder of neurocriminology and author of the neuromoral theory of antisocial behavior. Faculty page - Wikipedia

Frequently Asked Questions

What is criminal behavior in psychology?

Criminal behavior is conduct that violates the criminal law. Psychology treats it not as a single trait but as a heterogeneous outcome explained by individual differences, developmental history, situational decisions, and neurobiology, and it carves the legal category into kinds, such as developmental trajectories, that behave more lawfully (Moffitt, 1993).

Are there different types of offenders?

Yes. The most influential distinction is developmental: a small group of life-course-persistent offenders whose antisocial behavior starts early and endures, and a large group of adolescence-limited offenders whose law-breaking is confined to the teenage years (Moffitt, 1993). The two differ in onset, cause, and long-term outcome.

Does punishment actually deter crime?

Within limits. The rational-choice model treats offending as a decision weighing reward against the probability and severity of sanction, and the evidence is that the certainty of being caught deters more reliably than the harshness of the penalty, because the expected cost is dominated by the probability term (Cornish & Clarke, 1986).

Is criminal behavior genetic?

Neither purely genetic nor purely environmental. A landmark study found that the effect of childhood maltreatment on later antisocial behavior depended on a gene regulating MAOA activity, so maltreated children with one variant were much more likely to offend, a gene-environment interaction in which neither acts alone (Caspi et al., 2002).

Can we predict who will reoffend?

Partly. Structured actuarial instruments predict reoffending better than clinical judgment, but they estimate group probabilities and are stronger at identifying low risk than at confirming that a high-risk person will actually offend (Fazel et al., 2012), so their output is a probability, not an individual verdict.

What is psychopathy, and how does it relate to crime?

Psychopathy is a clinical configuration of callousness, shallow affect, and impulsive antisociality, measured by Hare's Psychopathy Checklist, and it is among the stronger single predictors of serious and persistent violent offending (Hare & Neumann, 2008). It is a risk factor, not a synonym for criminality.

What is neurocriminology?

Neurocriminology studies the neural and biological bases of antisocial and criminal behavior. Reviews relate offending to differences in prefrontal cortex and limbic structures (Glenn & Raine, 2014), and the neuromoral theory proposes that impaired function across a prefrontal-limbic moral network underlies the deficits seen in antisocial and psychopathic individuals (Raine, 2019).

Do rehabilitation programs reduce reoffending?

The randomized evidence says yes, modestly. A meta-analysis of randomized controlled trials of psychological interventions in prison found a genuine but moderate reduction in recidivism, while noting how few rigorous trials exist (Beaudry, Yu, Perry, & Fazel, 2021). Programs work best when they match intensity to risk and target changeable criminogenic needs (Andrews, Bonta, & Wormith, 2006).

References

Andrews, D. A., Bonta, J., & Wormith, J. S. (2006). The recent past and near future of risk and/or need assessment. Crime & Delinquency, 52(1), 7-27. https://doi.org/10.1177/0011128705281756

Beaudry, G., Yu, R., Perry, A. E., & Fazel, S. (2021). Effectiveness of psychological interventions in prison to reduce recidivism: A systematic review and meta-analysis of randomised controlled trials. The Lancet Psychiatry, 8(9), 759-773. https://doi.org/10.1016/S2215-0366(21)00170-X

Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., Taylor, A., & Poulton, R. (2002). Role of genotype in the cycle of violence in maltreated children. Science, 297(5582), 851-854. https://doi.org/10.1126/science.1072290

Cornish, D. B., & Clarke, R. V. (Eds.). (1986). The reasoning criminal: Rational choice perspectives on offending. Springer-Verlag.

Dodge, K. A., & Coie, J. D. (1987). Social-information-processing factors in reactive and proactive aggression in children's peer groups. Journal of Personality and Social Psychology, 53(6), 1146-1158. https://doi.org/10.1037/0022-3514.53.6.1146

Eysenck, H. J. (1996). Personality and crime: Where do we stand? Psychology, Crime & Law, 2(3), 143-152. https://doi.org/10.1080/10683169608409773

Fazel, S., Singh, J. P., Doll, H., & Grann, M. (2012). Use of risk assessment instruments to predict violence and antisocial behaviour in 73 samples involving 24827 people: Systematic review and meta-analysis. BMJ, 345, e4692. https://doi.org/10.1136/bmj.e4692

Fox, B. H., & Farrington, D. P. (2018). What have we learned from offender profiling? A systematic review and meta-analysis of 40 years of research. Psychological Bulletin, 144(12), 1247-1274. https://doi.org/10.1037/bul0000170

Glenn, A. L., & Raine, A. (2014). Neurocriminology: Implications for the punishment, prediction and prevention of criminal behaviour. Nature Reviews Neuroscience, 15(1), 54-63. https://doi.org/10.1038/nrn3640

Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stanford University Press.

Hare, R. D., & Neumann, C. S. (2008). Psychopathy as a clinical and empirical construct. Annual Review of Clinical Psychology, 4, 217-246. https://doi.org/10.1146/annurev.clinpsy.3.022806.091452

Moffitt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100(4), 674-701. https://doi.org/10.1037/0033-295X.100.4.674

Moffitt, T. E. (2018). Male antisocial behaviour in adolescence and beyond. Nature Human Behaviour, 2(3), 177-186. https://doi.org/10.1038/s41562-018-0309-4

Raine, A. (2019). The neuromoral theory of antisocial, violent and psychopathic behavior. Psychiatry Research, 277, 64-69. https://doi.org/10.1016/j.psychres.2018.11.025

Ward, T., & Brown, M. (2004). The good lives model and conceptual issues in offender rehabilitation. Psychology, Crime & Law, 10(3), 243-257. https://doi.org/10.1080/10683160410001662744