COMPAS vs LSI-R (Level of Service Inventory-Revised)

[COMPAS] excels at generating rapid, statistically derived risk scores from a large dataset of static and dynamic factors. Because it relies on a proprietary algorithm, it provides a level of consistency that eliminates inter-rater variability, making it a powerful tool for high-volume, initial screening. For example, its ability to process 137 items and output a decile risk score offers a standardized metric that is easily integrated into automated court docket systems.

[LSI-R] takes a fundamentally different approach by embedding risk assessment within a semi-structured interview guided by the Risk-Need-Responsivity (RNR) model. This results in a richer, more nuanced profile that identifies specific criminogenic needs—like substance abuse or antisocial patterns—directly linked to a case management plan. The trade-off is a significant investment in staff training and time, with a full assessment often taking 45-90 minutes to complete.

The key trade-off: If your priority is rapid, scalable triage for pretrial release decisions with minimal human resource expenditure, choose COMPAS. If you prioritize developing a dynamic, individualized rehabilitation roadmap that directly informs correctional programming intensity, choose LSI-R.

COMPAS excels at rapid, scalable risk scoring because it automates the prediction process using a proprietary algorithm. For example, it generates a decile risk score from 137 static and dynamic factors in seconds, providing a consistent, mathematically derived output that requires minimal staff training. This efficiency is critical for high-volume pretrial services where immediate, standardized risk flags are needed to inform initial detention decisions.

LSI-R takes a fundamentally different approach by embedding risk assessment within a structured, semi-structured interview conducted by a trained correctional professional. This results in a richer, more nuanced dataset that identifies specific criminogenic needs—such as substance abuse, education deficits, or antisocial attitudes—directly linked to the Risk-Need-Responsivity (RNR) model. The trade-off is significant: it requires 30-45 minutes of face-to-face time and ongoing staff certification, making it resource-intensive but highly actionable for individualized case management.

The key trade-off: If your priority is a fast, low-cost, and statistically consistent triage tool for high-volume sentencing or pretrial screening, choose COMPAS. If you prioritize a transparent, dynamic needs assessment that directly informs a sequenced rehabilitation plan and allows for clinical override, choose LSI-R. For agencies focused on reducing recidivism through targeted programming rather than just predicting it, the LSI-R's integration with the RNR framework provides a clear, evidence-based pathway that a black-box score cannot replicate.