Implementing LS/CMI Across Global Contexts: Localization Without Compromise
Key takeaways
- LS/CMI is grounded in the Risk–Need–Responsivity (RNR) framework, linking assessment to intervention through structured identification of criminogenic risk, need, and responsivity considerations.
- Global use depends on disciplined localization, not informal adaptation, to maintain its core structure while aligning with local systems.
- Localization takes multiple forms, including system integration, language and cultural adaptation, and local validation.
- Predictive validity varies by context, reinforcing the importance of local norms, validation, and training to support effective use.
Risk/needs assessments like the Level of Service/Case Management Inventory™ (LS/CMI™) are not limited by geography—they are shaped by context. The challenge, of course, is not proving that these tools work, but ensuring they can be applied locally without diluting the science behind them. Throughout correctional systems, success depends on striking a careful balance: maintaining a consistent, evidence-based framework while adapting implementation to fit local practices, populations, and structures.
LS/CMI is one of the most widely used risk/needs assessments in correctional practice. Grounded in the Risk–Need–Responsivity (RNR) framework and the Central Eight Risk/Need domains history of antisocial behavior, substance abuse, antisocial associates, etc.), LS/CMI provides a general, empirically supported model of criminal behavior validated across diverse populations and settings. Its design links assessment to intervention by structuring how practitioners move from information gathering (e.g., identifying criminogenic needs and responsivity factors) to case planning.
A substantial body of research on the Level of Service (LS) suite of instruments has demonstrated statistically significant prediction of general and violent recidivism across diverse samples.¹ At the same time, this work shows that predictive strength can vary across geographic settings. Effects tend to be strongest in jurisdictions where the LS tools were developed and normed, with somewhat smaller though still meaningful effects observed elsewhere. Take the validation studies in Canada, where the assessment tool was originally developed. The reports show the strongest prediction of recidivism outcomes, while U.S. samples show slightly smaller (though still significant) effects.¹ These findings suggest that LS/CMI captures broadly applicable dimensions of criminal behavior, while also highlighting the importance of localization (i.e., governance that aligns with local law, language, policy, and practice) in shaping how those dimensions are expressed and assessed.
International use of LS/CMI is therefore not solely about whether the tool works across jurisdictions. A risk/needs assessment that cannot adapt to local workflows may struggle to gain worldwide adoption; one that adapts too freely may drift from the science and fidelity. Practitioners, agencies, implementation teams, trainers, researchers, and policymakers are successful when they consider local systems, practices, and populations. Essentially, effective use depends on maintaining the integrity of the underlying framework while adapting to the surrounding structures within which the tool is applied.
The examples below illustrate three distinct aspects of this process: system-level integration, language and cultural adaptation, and local validation.
Scotland: Embedding LS/CMI Within System-Level Practice
The implementation of LS/CMI in Scotland provides a clear example of localization through system-level alignment. Following national adoption between 2010 and 2012, LS/CMI was incorporated across community- and prison-based social work services.² Oversight is provided by the Risk Management Authority, which has positioned the method as a standardized approach to assessment and case management within the justice system.³
Implementation was designed at the system level rather than at the individual tool level. The LS/CMI was embedded within existing workflows and supported through a coordinated infrastructure. This included a national electronic IT system for recording and sharing assessment and case management information, structured training provision, and formal quality assurance processes (e.g., LS/CMI Change Control portal for submitting change requests that don’t undermine LS/CMI’s scientific properties).
Adaptation has been managed through governance rather than informal modification. Processes are in place to review proposed changes, with the explicit aim of preserving the assessment’s theoretical structure, item content, and scoring procedures. In this respect, localization has focused on ensuring consistency of use across settings, rather than altering what LS/CMI measures.
This approach reflects a broader principle. Fidelity is not limited to maintaining item wording or scoring rules. It also depends on how consistently the tool is implemented, the quality of training provided to practitioners, and the presence of mechanisms to monitor and support its use over time.
Key lesson:
Localization is most effective when it focuses on how the tool is used within a system, rather than altering what the tool measures.
Quebec and France: Preserving Measurement Through Language and Culture
A second dimension of localization concerns language and culture. In contexts where LS/CMI is used across different linguistic and cultural settings, translation alone is insufficient. The central issue is whether the assessment continues to capture the same constructs in a way that is meaningful within the local context.
Research on the French adaptation of LS/CMI in Quebec illustrates how this challenge has been addressed. The translation followed a structured cross-cultural process designed to preserve both the wording and the intended meaning of items.⁴ The resulting version retained all risk/need items and the original scoring algorithm while adapting the language for use in a French-speaking population. The study reported strong reliability and acceptable predictive validity, indicating that the assessment performed in a manner broadly consistent with the original instrument.
Evidence from France provides a complementary perspective. A study of LS/CMI use among 128 French justice-involved individuals demonstrated acceptable levels of inter-rater reliability, internal consistency, and convergent validity as further confirmation of its psychometric properties.⁵ These findings suggest that, when carefully adapted and evaluated, the instrument can be applied in ways that are consistent with its underlying structure.
At the same time, these examples emphasize that translation and validation serve different purposes. Translation addresses how the assessment is communicated. Validation addresses how it performs. Both support meaningful implementation in ways that responsibly account for local and cultural nuances.
Key lesson:
Localization must address not just language, but also measurement integrity, ensuring that the constructs captured by the tool remain stable across contexts.
New Zealand: Examining Performance Beyond North America
In jurisdictions that differ from the contexts in which LS/CMI was developed, local validation plays a particularly important role. It provides an empirical basis for understanding how the tool operates within a specific correctional system.
A recent study conducted in New Zealand (NZ) examined the performance of LS/CMI among a sample size of 772 men under the management of the New Zealand Department of Corrections.⁶ Over a one-year follow-up period, the findings indicated that domain and total scores demonstrated low-to-moderate discriminative accuracy and added predictive value beyond an existing static risk measure.
The study also considered subgroup performance, reporting no statistically significant differences in predictive accuracy between NZ Māori and NZ European individuals. These results provide evidence that LS/CMI can function in a broadly comparable way across groups within this setting, while also highlighting the importance of examining contextual factors, including assessor practices, that may influence outcomes.
From an implementation perspective, such studies contribute to an ongoing process of evaluation. They support decisions about training, interpretation, and, where necessary, recalibration or norming. They do not imply that the tool operates identically across contexts, but rather that its performance can be examined and supported within those contexts.
Key lesson:
Local validation does not replace the core framework: it tests and supports its use within a specific context.
Localization as a Structured Process
Across these examples, a consistent pattern emerges: LS/CMI is not adapted through changes to its core structure. Instead, localization occurs through adjustments to the systems within which the assessment is used.
In practice, this involves aligning implementation with local legal frameworks, integrating the tool into existing workflows, and ensuring that practitioners are trained to apply it consistently. Where necessary, translation and cultural adaptation are undertaken to preserve meaning, and validation studies are used to examine how the tool performs in specific contexts.
This approach allows for variation in how LS/CMI is applied while maintaining a stable underlying framework. It also reflects the broader purpose of the assessment. Although predictive validity remains an important consideration, LS/CMI is designed as a comprehensive assessment and case management tool. Its structure supports not only the classification of risk, but also the identification of needs, the development of intervention plans, and the monitoring of progress over time.
Taken together, these features help explain why LS/CMI has been adopted across a range of jurisdictions. At its core, LS/CMI remains anchored in a general theory of crime that continues to provide a stable and generalizable foundation for understanding and addressing risk and needs across jurisdictions. Its framework appears sufficiently general to apply across settings, while implementation can be tailored to local systems through structured and controlled processes.
Consistency Without Compromise
LS/CMI demonstrates that it is possible to apply a structured, evidence-based assessment across diverse jurisdictions without fragmenting it through unstructured adaptation.
Its effectiveness lies in the balance between:
- A generalizable framework, grounded in the RNR model and the Central Eight Risk/Need domains, which together provide a structured and empirically supported basis for understanding and addressing risk and need.
- A disciplined approach to localization, ensuring that implementation aligns with local systems while preserving measurement integrity
When agencies treat LS/CMI as part of a broader system of practice, supported by training, governance, and ongoing evaluation, it can provide a consistent foundation for assessment and case management across settings, from custody to community supervision.
Ultimately, the lesson is not that LS/CMI works the same way everywhere. It can work effectively in different places when implemented thoughtfully, systematically, and with attention to local context.
Learn more about LS/CMI and how it supports structured, evidence-based case management.
Still have questions? Get in touch with our team to discuss how LS/CMI can be implemented within your local environment.
References
¹ Olver, M. E., Stockdale, K. C., & Wormith, J. S. (2014). Thirty years of research on the Level of Service scales: A meta-analytic examination of predictive accuracy and sources of variability. Psychological Assessment, 26(1), 156–176. https://doi.org/10.1037/a0035080
² Care Inspectorate. (2026). Prison-based social work review: Phase 2 report. https://ci-web-media-production-bucket.s3.eu-west-2.amazonaws.com/assets/strategic-reports/prison-based-social-work-review-phase-2-report.pdf
³ Risk Management Authority. (n.d.). LS/CMI. https://www.rma.scot/support/ls-cmi/
⁴ Giguère, G., Hamidi, N., Bourassa, C., & Brouillette-Alarie, S. (2026). Assessing cultural bias in recidivism risk tools: Differential item functioning and predictive validity of the ORAC-PCQ and LS/CMI among Indigenous and non-Indigenous men in Quebec’s correctional system. Race and Justice. https://doi.org/10.1177/21533687261419431
⁵ Bertsch, I., Guay, J.-P., Réveillère, C., Telle, É., Douceron, H., Dubuisson, M., & Pham, T. H. (2023). An investigation of reliability and validity of the LS/CMI with French offenders. L’Encéphale, 49(5), 460–465. https://doi.org/10.1016/j.encep.2022.05.003
⁶ Botha, R., Davies, S. T., & Nichols, T. (2026). Psychometric properties of the LS/CMI among men under correctional supervision in New Zealand. Criminal Justice and Behavior. https://doi.org/10.1177/00938548261428691
