Genetic Contributions to Criminal Behaviour

Initial twin/adoption studies classified someone as “criminal” only if they had an official conviction.
Key limitation:
- One can commit crimes without ever being detected or prosecuted.
- Underestimates true prevalence ➜ systematically lowers concordance figures.
Concordance rate = the proportion of twin pairs in which both twins meet the diagnostic/behavioural criterion.

Researchers now gather richer data to avoid the “undetected offender” problem:
- Anonymous self-report questionnaires covering frequency & variety of illegal acts.
- Collateral reports from peers, parents, teachers.
- Standardised psychological/psychiatric instruments.
Advantages:
- Captures covert offences.
- Allows dimensional (how much?) rather than binary (yes/no) scoring.
- Yields multiple, cross-validated perspectives on each participant.

With dimensional data, geneticists estimate heritability ( $h^{2}$ ): the proportion of variance in a trait attributable to genetic differences within a specific population and time.
Computed via twin-study formulas that compare monozygotic (MZ) and dizygotic (DZ) intraclass correlations (details skipped in lecture).
Benefits:
- Produces a single, interpretable number that factors in both similarity & variability.
- Separates genetic influence from shared/non-shared environmental effects.

Sample: $89$ monozygotic twin pairs & $47$ dizygotic twin pairs.
Instruments employed:
- Psychopathic Personality Inventory.
- Additional behaviour/antisocial checklists.
Core findings:
- MZ pairs: all behavioural correlations were positive & statistically significant.
- DZ pairs: correlations were weaker and failed to reach significance.
- Interpretation: higher genetic overlap (100 % in MZ vs ≈50 % in DZ) corresponded to higher behavioural similarity, implicating genetic factors.

Reported heritability coefficients for “criminal or antisocial behaviour” typically fall between $0.24$ and $0.56$ .
- Lower end (≈ $24\%$ ): genetics play a modest but non-trivial role.
- Upper end (≈ $56\%$ ): a little over half of observed variability linked to inherited factors.
Remaining variance is attributable to:
- Shared environment (family socioeconomic status, neighborhood, parenting style).
- Non-shared environment (unique peer groups, individual life events, measurement error).

Non-determinism: A heritability of $0.56$ does not mean an individual is 56 % predestined to offend; environmental interventions can still be highly effective.
Policy considerations:
- Early-life prevention programs may need to account for both genetic susceptibility and environmental context.
- Risk-assessment tools should avoid genetic essentialism.
Ethical issues:
- Potential misuse of genetic information (e.g., discrimination, eugenic rhetoric).
- Importance of framing results as population-level statistics, not individual fate.
Research trajectory:
- Incorporating molecular genetics (GWAS) to pinpoint specific alleles.
- Exploring gene-environment interaction (G×E) models to clarify how social context moderates genetic risk.