What did key reviews (e.g., Moffitt 1993, Herrnstein & Murray 1994) conclude about IQ and criminality?
This fact-check may be outdated. Consider refreshing it to get the most current information.
Executive summary
Terrie Moffitt’s 1993 developmental taxonomy argued that the IQ–delinquency link is heterogeneous across offender groups and that pooling hides stronger associations for life‑course‑persistent offenders, while Herrnstein and Murray reported a clear inverse association between measured cognitive ability and criminal involvement and used large national data to argue cognitive differences help explain social outcomes; both findings have been influential but also vigorously contested on methods, causation, and policy implications [1] [2] [3] [4].
1. Moffitt : a developmental partition that reframes the IQ–crime story
Moffitt’s 1993 developmental taxonomy did not assert a simple, uniform IQ–crime causal chain but instead distinguished two main offender types — adolescent‑onset (largely transient) and life‑course‑persistent (early‑onset, stable antisocial behavior) — and argued that pooling those groups obscures how cognitive deficits more strongly mark the small subset of life‑course‑persistent offenders, thereby changing how researchers interpret aggregate IQ–crime correlations [1].
2. Herrnstein & Murray : large samples, inverse association, and controversial policy inferences
Using national datasets (notably the NLSY/AFQT as an intelligence proxy), Herrnstein and Murray reported a statistically significant inverse association between measured cognitive ability and criminal involvement and argued cognitive ability explains a substantial share of class‑structured outcomes; they extended these empirical claims into policy recommendations that emphasized cognitive endowments over socioeconomic remedies, a stance that drew sharp criticism [2] [3] [4] [5].
3. What the quantitative reviews actually found: modest inverse correlations and complexity
Across multiple reviews and studies the IQ–crime correlation repeatedly appears negative and modest in size (roughly r ≈ −.20 in classic summaries), with many studies finding lower average IQ among offenders and particular deficits in verbal ability for many delinquent groups; at the same time research documented linear and non‑linear functional forms and variation by crime type, measurement, and sample [6] [7] [8] [9].
4. Methodological caveats that temper strong causal claims
Critics and subsequent analyses highlighted several serious limitations in drawing causal inferences: measurement choices (e.g., use of AFQT as an IQ proxy), the possibility that lower IQ raises detection or processing likelihood rather than offending per se, confounding with socioeconomic status and school failure, sample age/cohort issues in older datasets like the NLSY, and the analytical problem Moffitt flagged that pooling distinct offender trajectories masks heterogeneity — all of which mean the association is robust but its causal interpretation is contested [2] [3] [4] [8] [1].
5. Consensus and continuing debates: association acknowledged, mechanisms debated, policy contested
The mainstream literature accepts a robust association between lower cognitive ability and greater antisocial behavior while vigorously debating mechanisms — direct cognitive causes, mediated effects through schooling/self‑control/SES, differential detection, and developmental trajectories — and criticizing policy prescriptions that treat IQ as destiny; critics pointed out that Herrnstein & Murray’s emphasis on genetic explanations and policy shifts toward cognitive‑focused remedies reflects normative and political choices as much as empirical ones [10] [8] [5] [11].