Psychometric and conceptual challenges in dynamic risk factors for recidivism

Dynamic risk factors play a central role in correctional practice because they are theoretically malleable and empirically linked to recidivism. Using them to guide supervision, treatment, and discharge decisions requires addressing several psychometric, methodological, and conceptual challenges. This article synthesises guidance in six areas: (a) reliability (inter-rater and test–retest); (b) predictive validity, including discrimination and calibration; (c) regression to the mean and design strategies to mitigate it; (d) construct validity and measurement invariance across time; (e) sensitivity to meaningful change, including multi-wave modelling; and (f) causal inference, distinguishing risk markers from causal targets. Practical procedures, such as rater training, drift monitoring, structural equation modelling, Cox models with time-varying covariates, dynamic calibration, and landmarking, are outlined alongside applied examples. Collectively, these recommendations strengthen interpretability, fairness, and predictive performance, improving the evidentiary foundation for dynamic risk factor-informed assessment and case management.PRACTICE IMPACT STATEMENT Defensible use of dynamic risk factors requires multiple time-point assessments, model-based estimates of change, and transparent reporting of reliability, the standard error of measurement, and reliable-change thresholds. Establishing measurement invariance across time helps ensure that observed score changes reflect meaningful shifts in criminogenic need rather than measurement noise or instability. Together, these practices support clearer interpretation, fairer decision-making, and more accurate communication of changes in risk over time.