Gene-Environment Interactions and Crime

Gene-environment interactions and crime examines the specific statistical and theoretical framework through which genetic and environmental factors combine non-additively to shape antisocial behavior risk, a concept that functions as biosocial criminology’s central organizing theoretical commitment and that this article treats in considerably greater technical depth than the brief theoretical overview provided in the companion genetics and criminal behavior article. This article, situated within Biosocial Criminology and the broader Criminology Theories silo, examines the formal distinction between gene-environment correlation and interaction, the statistical approaches researchers use to model these effects, the MAOA-maltreatment finding that became this research area’s paradigmatic case study, additional documented interactions beyond MAOA, the competing theoretical models proposed to explain interaction patterns, and gene-environment interaction’s foundational role in biosocial criminology’s broader theoretical architecture.

Gene-environment interaction’s technical precision distinguishes it from the more general probabilistic and interactive theoretical vocabulary examined throughout this category’s other biosocial articles, since interaction carries a specific statistical meaning, that a genetic variant’s effect on an outcome differs systematically across levels of an environmental variable, distinct from gene-environment correlation’s different statistical meaning, that genetic and environmental exposures are themselves statistically associated with one another. Conflating these two distinct statistical phenomena represents a common source of confusion in both popular and occasionally scholarly discussion of biosocial criminology, making the formal distinction this article develops analytically important beyond its specific empirical applications.

This article proceeds through six sections: the formal distinction between gene-environment correlation and interaction, the statistical modeling approaches researchers use to test these effects, the MAOA-maltreatment case study that became this field’s paradigmatic finding, additional documented gene-environment interactions beyond MAOA, the competing theoretical models explaining interaction patterns, and this concept’s foundational significance for biosocial criminology’s broader theoretical architecture.

Distinguishing Correlation from Interaction

Gene-Environment Correlation Mechanisms

Gene-environment correlation describes the nonrandom statistical association between an individual’s genotype and the environment that individual experiences, a phenomenon behavioral geneticists have decomposed into three mechanisms: passive correlation, in which genetically related parents provide both genes and a correlated rearing environment to offspring; evocative correlation, in which an individual’s genetically influenced characteristics evoke particular responses from others; and active correlation, in which individuals select or construct environments correlated with their own genetic predispositions (Scarr & McCartney, 1983). Sandra Scarr and Kathleen McCartney’s foundational theoretical framework proposed that these three correlation types shift in relative importance across development, with passive correlation dominating in early childhood when parents control most environmental exposure, and active correlation becoming increasingly important as individuals gain autonomy to select their own environments during adolescence and adulthood.

Gene-environment correlation’s methodological significance lies in its capacity to generate spurious apparent environmental effects, since a correlation between an environmental exposure and a behavioral outcome might reflect genetically influenced selection into that environment rather than any genuine causal environmental effect, a confound that adoption and twin designs, examined in dedicated companion articles, were specifically developed to address by statistically separating genetic and environmental transmission pathways (Jaffee & Price, 2007). Sara Jaffee and Thalia Price’s review of gene-environment correlation research specifically identified evocative correlation as particularly relevant to understanding antisocial behavior’s developmental origins, since genetically influenced child temperamental characteristics, including early irritability, may evoke harsher parental discipline that observational research might otherwise misattribute to purely environmental parenting effects.

Gene-Environment Interaction Defined

Gene-environment interaction, formally distinct from correlation, describes circumstances in which a genetic variant’s statistical effect on a behavioral outcome differs systematically depending on environmental exposure level, such that the same genetic variant produces different behavioral consequences under different environmental conditions, a pattern requiring formal statistical interaction testing to distinguish from either genetic or environmental main effects considered independently (Moffitt, Caspi, & Rutter, 2006). Terrie Moffitt and colleagues’ influential methodological clarification specifically emphasized that gene-environment interaction requires demonstrating that genetic effects vary across environmental levels using appropriate statistical interaction terms, distinguishing genuine interaction from the simpler and more common finding that both genetic and environmental main effects independently predict an outcome without any interactive relationship between them.

This formal distinction carries substantial practical importance because popular and even some scholarly discussion frequently uses “gene-environment interaction” loosely to describe any research examining both genetic and environmental factors, when the technical concept specifically requires demonstrating a statistical interaction effect beyond simple additive main effects, a precision this article maintains throughout its subsequent discussion of specific documented interaction findings.

Table 1. Gene-Environment Correlation Versus Interaction

Concept	Statistical Question	Example	Methodological Approach
Passive rGE	Do genes and environment co-occur due to shared family origin?	Antisocial parents provide genes and higher-conflict household	Adoption designs separate genetic and rearing transmission
Evocative rGE	Does genotype evoke particular environmental responses?	Irritable temperament evokes harsher discipline	Longitudinal designs tracking child-evoked parenting change
Active rGE	Do individuals select environments matching genotype?	Impulsive adolescents select delinquent peer groups	Designs tracking environmental selection over development
Gene-Environment Interaction (GxE)	Does a genetic effect differ across environmental levels?	MAOA variant predicts antisocial outcomes only after maltreatment	Formal statistical interaction term in regression models

Statistical Approaches to Modeling GxE

Candidate Gene by Environment Designs

Candidate gene by environment designs, the methodological approach underlying the MAOA-maltreatment research examined later in this article, test whether a specific pre-selected candidate gene’s association with a behavioral outcome varies across levels of a specific measured environmental exposure, typically implemented through regression models including a formal genotype-by-environment interaction term alongside genetic and environmental main effects (Caspi et al., 2002). This candidate gene by environment approach inherited the broader candidate gene methodology’s limitations examined in the companion molecular genetics article, including modest sample sizes and the replication difficulties that prompted the field’s subsequent methodological reform.

Danielle Dick and colleagues’ methodological review specifically identified statistical power as candidate gene by environment research’s central limitation, since detecting genuine interaction effects statistically requires considerably larger samples than detecting comparable main effects, meaning that many candidate gene by environment studies, including some influential early MAOA research, were likely underpowered to reliably detect the interaction effects they reported (Dick et al., 2015).

Genome-Wide by Environment Interaction Models

Contemporary gene-environment interaction research increasingly employs genome-wide by environment interaction designs, extending the genome-wide association methodology examined in the companion molecular genetics article to test interaction effects across the full genome rather than within a single pre-selected candidate gene, an approach requiring even larger sample sizes than standard genome-wide association studies given the additional statistical power that interaction detection demands (Dick et al., 2015). This genome-wide interaction approach remains methodologically less mature than either standard genome-wide association studies or candidate gene by environment designs, reflecting the field’s ongoing development of adequately powered methodology for this particularly demanding analytical task.

Polygenic risk score by environment interaction designs, examined in the companion molecular genetics article’s discussion of polygenic scoring, offer an intermediate methodological approach, testing whether a composite polygenic risk score’s association with antisocial behavior varies across environmental conditions, an approach that aggregates genome-wide genetic information into a single interaction term while avoiding some of the full genome-wide interaction approach’s most demanding statistical power requirements (Tielbeek et al., 2017).

The MAOA-Maltreatment Paradigm Case

Caspi et al. and the Original Finding

Avshalom Caspi and colleagues’ 2002 study, examined briefly in the companion heritability and molecular genetics articles but treated here as this field’s central paradigmatic case, found that a functional polymorphism affecting monoamine oxidase A activity predicted antisocial outcomes specifically among individuals who had experienced childhood maltreatment, with minimal association among individuals who had not been maltreated, a finding published in Science that rapidly became behavioral genetics’ most frequently cited illustration of gene-environment interaction across psychiatric and criminological research alike (Caspi et al., 2002). This finding’s paradigmatic status reflects several converging factors: its publication in an exceptionally high-profile journal, its intuitively graspable interactive pattern illustrated through a simple crossover graph showing divergent outcomes by genotype specifically among maltreated individuals, and its arrival at a moment when behavioral genetics broadly sought a compelling illustration of interaction moving beyond simple heritability estimation.

The study’s specific finding, that low-activity MAOA genotype combined with childhood maltreatment predicted substantially higher antisocial outcomes while low-activity genotype without maltreatment predicted outcomes comparable to high-activity genotype regardless of maltreatment history, provided precisely the clean interactive pattern that behavioral genetic theory had proposed but that few empirical studies had previously demonstrated with comparable clarity (Caspi et al., 2002).

Replication, Meta-Analysis, and Ongoing Debate

Subsequent replication attempts targeting the Caspi MAOA-maltreatment finding produced a genuinely mixed pattern, with some studies replicating the basic interactive pattern and others failing to detect a comparable effect, prompting meta-analytic synthesis attempting to resolve this inconsistent replication record (Kim-Cohen, Caspi, Taylor, Williams, Newcombe, Craig, & Moffitt, 2006; Byrd & Manuck, 2014). Amber Byrd and Stephen Manuck’s influential meta-analysis, synthesizing the accumulated MAOA-maltreatment interaction literature, found a statistically significant but modest combined effect, smaller and less consistent than Caspi’s original finding, illustrating the broader pattern of effect size attenuation with replication that characterizes much candidate gene by environment research examined in the companion molecular genetics article.

This mixed replication record has generated substantial methodological debate regarding the original finding’s reliability, with some researchers emphasizing the meta-analytic confirmation of a genuine, if modest, effect, while others emphasize the considerable heterogeneity across studies as evidence that the original finding’s specific magnitude, and potentially its underlying mechanism, remain less certain than the finding’s paradigmatic textbook status might suggest (Byrd & Manuck, 2014).

Table 2. The MAOA-Maltreatment Finding Across Replication Attempts

Study Characteristic	Caspi et al. (2002) Original	Subsequent Replication Pattern
Sample Size	Approximately 1,000	Highly variable across replication attempts
Effect Size	Large, clean interactive pattern	Meta-analytically smaller and more variable
Maltreatment Measurement	Retrospective and prospective combined measures	Varies considerably across studies
Antisocial Outcome Measures	Composite index including conviction and diagnosis	Varies considerably across studies
Meta-Analytic Conclusion	N/A (original finding)	Significant but modest combined effect (Byrd & Manuck, 2014)

Additional Documented Gene-Environment Interactions

Serotonin Transporter and Stress Reactivity

The serotonin transporter gene-linked polymorphic region, examined initially within depression research before extending into criminological application, has shown interaction with early life stress in predicting antisocial and aggressive outcomes in some studies, with the short allele variant associated with heightened stress reactivity showing stronger association with antisocial behavior specifically among individuals with substantial early adversity exposure (Reif, Rösler, Freitag, Schneider, Eujen, Kissling, Wenzler, Jacob, Retz-Junginger, Thome, Lesch, & Retz, 2007). This serotonin transporter research parallels the MAOA-maltreatment pattern in its basic interactive structure, though this research literature has faced comparable replication challenges to those documented for MAOA, reflecting the broader candidate gene by environment replication difficulties examined throughout this article and the companion molecular genetics article.

Andreas Reif and colleagues’ research examining this interaction specifically within a criminal offender sample found that serotonin transporter genotype interacted with childhood adversity to predict violent offending, extending the basic gene-environment interaction pattern beyond MAOA to a second neurotransmitter system gene, though this replication across genetic systems has not resolved the underlying methodological concerns regarding candidate gene by environment research’s general statistical power limitations.

Dopamine System Genes and Parenting Quality

Dopamine receptor gene variants, examined in the companion molecular genetics article’s discussion of candidate gene research, have shown interaction with parenting quality and family environment in predicting externalizing and antisocial behavior in some research, with genetic variants associated with reduced dopaminergic function showing stronger association with antisocial outcomes specifically among children experiencing inadequate parental monitoring or harsh discipline (Bakermans-Kranenburg & van IJzendoorn, 2011). Marian Bakermans-Kranenburg and Marinus van IJzendoorn’s meta-analytic synthesis of dopamine-related gene-environment interaction research found evidence broadly consistent with differential susceptibility theory, examined in the companion genetics article, since dopamine-related genetic variants associated with poorer outcomes under adverse parenting also showed evidence of enhanced benefit from especially supportive and sensitive parenting, extending the differential susceptibility pattern beyond its original documentation to this additional neurotransmitter system.

This dopamine-parenting interaction research illustrates gene-environment interaction’s extension beyond the maltreatment-focused MAOA and serotonin transporter research examined earlier in this article toward more graduated environmental measures, including parenting quality variation within the normal range rather than only severe maltreatment specifically, broadening this research tradition’s environmental scope considerably beyond its original focus on extreme adversity.

Theoretical Models of Interaction

Diathesis-Stress Versus Differential Susceptibility

The diathesis-stress model, examined in the companion genetics article, proposes that genetic variants create vulnerability specifically to adverse environmental conditions, predicting worse outcomes under adversity without any corresponding benefit under favorable conditions, while differential susceptibility theory proposes that the same variants confer general environmental sensitivity, predicting both worse outcomes under adversity and better outcomes under favorable conditions relative to less genetically sensitive individuals (Belsky & Pluess, 2009). Distinguishing empirically between these two theoretical models requires research designs incorporating the full range of environmental quality, from severely adverse through highly supportive, since diathesis-stress and differential susceptibility models generate identical predictions within samples examining adversity alone but diverge specifically in their predictions regarding genetically sensitive individuals’ outcomes under unusually favorable environmental conditions.

Jay Belsky and Michael Pluess’s methodological argument for testing differential susceptibility specifically, rather than assuming diathesis-stress by default, has increasingly shaped contemporary gene-environment interaction research design within biosocial criminology, prompting researchers to incorporate positive environmental measures alongside adversity measures specifically to adjudicate between these competing theoretical models rather than assuming diathesis-stress’s vulnerability-only framing without adequate empirical testing (Belsky & Pluess, 2009).

The Bioecological Model of Development

Urie Bronfenbrenner’s bioecological model of human development, though developed originally within developmental psychology rather than criminology specifically, has substantially informed biosocial criminology’s theoretical framing of gene-environment interaction, proposing that individual development emerges through proximal processes, the individual’s active engagement with their immediate environment, operating within nested environmental systems ranging from immediate family and peer contexts through broader neighborhood, cultural, and historical contexts (Bronfenbrenner & Ceci, 1994). This bioecological framework provides gene-environment interaction research a broader developmental systems theoretical grounding beyond the narrower statistical interaction concept examined throughout this article, situating specific findings including MAOA-maltreatment interaction within a more comprehensive account of how genetic predispositions interact with the full, nested range of environmental systems that shape human development.

This bioecological grounding has proven particularly valuable for connecting gene-environment interaction research to the neighborhood and community-level sociological factors examined in the companion twin studies article’s discussion of neighborhood-moderated heritability, since Bronfenbrenner’s nested systems framework explicitly incorporates the broader social and community contexts that purely dyadic parent-child interaction research, including much of the MAOA and serotonin transporter research examined earlier in this article, does not directly address.

Implications for Biosocial Criminological Theory

Why GxE Became the Paradigm’s Central Concept

Gene-environment interaction’s centrality to biosocial criminology’s theoretical identity, examined throughout the companion genetics article, reflects this concept’s unique capacity to simultaneously validate genetic influence’s empirical reality while explicitly rejecting genetic determinism, providing biosocial criminology’s founders precisely the theoretical tool needed to overcome criminology’s historical resistance to biological explanation examined in the companion heritability article (Walsh & Beaver, 2009). No other single concept within biosocial criminology’s theoretical vocabulary accomplishes this dual task as directly as gene-environment interaction, which demonstrates empirically, through findings including MAOA-maltreatment, that genetic effects are neither absent nor deterministic but rather conditionally expressed depending on environmental circumstance.

This theoretical centrality explains why gene-environment interaction research, despite the replication concerns examined throughout this article regarding specific findings including MAOA, retains such prominent placement within biosocial criminology education and theoretical writing, since the concept’s theoretical and rhetorical function within the paradigm’s broader disciplinary project operates somewhat independently of any single specific finding’s replication status.

Remaining Challenges and Future Directions

Gene-environment interaction research continues facing the statistical power and replication challenges examined throughout this article, with contemporary methodological reform, including genome-wide by environment interaction designs and pre-registration requirements examined in the companion molecular genetics article, representing the field’s ongoing effort to establish gene-environment interaction findings on more statistically secure footing than the candidate gene by environment era achieved (Duncan & Keller, 2011). This continuing methodological development suggests that gene-environment interaction’s empirical evidence base will likely continue evolving considerably even as the concept’s theoretical centrality to biosocial criminology remains stable.

Future gene-environment interaction research increasingly incorporates epigenetic mechanisms, examined in a dedicated companion article, that provide additional molecular pathways through which environmental exposure might alter genetic expression without changing underlying DNA sequence, potentially offering mechanistic explanation for how environmental factors specifically modify genetic influence at the biological level that purely statistical interaction modeling cannot directly address.

Conclusion

Gene-environment interactions and crime, examined at the technical depth this article has provided, reveals a theoretically central but empirically still-developing research area, formally distinct from gene-environment correlation, that has provided biosocial criminology its most important theoretical tool for reconciling genetic influence’s empirical reality with explicit rejection of genetic determinism. The MAOA-maltreatment finding’s paradigmatic status, despite its genuinely mixed replication record, illustrates how a specific empirical finding can achieve lasting theoretical significance within a discipline’s self-understanding independent of its precise ongoing empirical validation, while the differential susceptibility and bioecological theoretical models examined throughout this article demonstrate the concept’s continuing theoretical development beyond its original diathesis-stress formulation.

Understanding gene-environment interaction’s formal statistical meaning, its specific empirical evidence base, and its foundational theoretical role within biosocial criminology together provide essential context for appreciating why this concept, more than any other single idea examined throughout this category, defines contemporary biosocial criminology’s distinctive theoretical identity and its considered response to criminology’s longer history of oscillating between deterministic biological and purely environmental explanation.

References

Bakermans-Kranenburg, M. J., & van IJzendoorn, M. H. (2011). Differential susceptibility to rearing environment depending on dopamine-related genes: New evidence and a meta-analysis. Development and Psychopathology, 23(1), 39–52. https://doi.org/10.1017/S0954579410000635
Belsky, J., & Pluess, M. (2009). Beyond diathesis stress: Differential susceptibility to environmental influences. Psychological Bulletin, 135(6), 885–908. https://doi.org/10.1037/a0017376
Bronfenbrenner, U., & Ceci, S. J. (1994). Nature-nurture reconceptualized in developmental perspective: A bioecological model. Psychological Review, 101(4), 568–586. https://doi.org/10.1037/0033-295X.101.4.568
Byrd, A. L., & Manuck, S. B. (2014). MAOA, childhood maltreatment, and antisocial behavior: Meta-analysis of a gene-environment interaction. Biological Psychiatry, 75(1), 9–17. https://doi.org/10.1016/j.biopsych.2013.05.004
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
Dick, D. M., Agrawal, A., Keller, M. C., Adkins, A., Aliev, F., Monroe, S., Hewitt, J. K., Kendler, K. S., & Sher, K. J. (2015). Candidate gene-environment interaction research: Reflections and recommendations. Perspectives on Psychological Science, 10(1), 37–59. https://doi.org/10.1177/1745691614556682
Duncan, L. E., & Keller, M. C. (2011). A critical review of the first 10 years of candidate gene-by-environment interaction research in psychiatry. American Journal of Psychiatry, 168(10), 1041–1049. https://doi.org/10.1176/appi.ajp.2011.11020191
Jaffee, S. R., & Price, T. S. (2007). Gene-environment correlations: A review of the evidence and implications for prevention of mental illness. Molecular Psychiatry, 12(5), 432–442. https://doi.org/10.1038/sj.mp.4001950
Kim-Cohen, J., Caspi, A., Taylor, A., Williams, B., Newcombe, R., Craig, I. W., & Moffitt, T. E. (2006). MAOA, maltreatment, and gene-environment interaction predicting children’s mental health: New evidence and a meta-analysis. Molecular Psychiatry, 11(10), 903–913. https://doi.org/10.1038/sj.mp.4001851
Moffitt, T. E., Caspi, A., & Rutter, M. (2006). Measured gene-environment interactions in psychopathology: Concepts, research strategies, and implications for research, intervention, and public understanding of genetics. Perspectives on Psychological Science, 1(1), 5–27. https://doi.org/10.1111/j.1745-6916.2006.00002.x
Reif, A., Rösler, M., Freitag, C. M., Schneider, M., Eujen, A., Kissling, C., Wenzler, D., Jacob, C. P., Retz-Junginger, P., Thome, J., Lesch, K. P., & Retz, W. (2007). Nature and nurture predispose to violent behavior: Serotonergic genes and adverse childhood environment. Neuropsychopharmacology, 32(11), 2375–2383. https://doi.org/10.1038/sj.npp.1301359
Scarr, S., & McCartney, K. (1983). How people make their own environments: A theory of genotype-environment effects. Child Development, 54(2), 424–435.
Tielbeek, J. J., Johansson, A., Polderman, T. J. C., Rautiainen, M.-R., Jansen, P., Taylor, M., et al. (2017). Genome-wide association studies of a broad spectrum of antisocial behavior. JAMA Psychiatry, 74(12), 1242–1250. https://doi.org/10.1001/jamapsychiatry.2017.3069
Walsh, A., & Beaver, K. M. (2009). Biosocial criminology: New directions in theory and research. Routledge.