Psychometric Properties of a Risk Tool Across Indigenous Māori and European Samples in Aotearoa New Zealand: Measurement Invariance, Discrimination, and Calibration for Predicting Criminal Recidivism

Due to recent legal cases highlighting a lack of cross-ethnicity validity research using correctional risk assessment tools, we evaluated psychometric properties of Dynamic Risk Assessment for Offender Re-entry (DRAOR) scores across Māori (n = 1,812) and New Zealand (NZ) European samples (n = 1,211) in Aotearoa NZ. Using routine administrative data, our analyses suggested scoring properties were invariant across ethnicity for 15 of 19 items. Discrimination properties were also equivalent, but we observed a higher recidivism base rate among Māori participants, consistent with official statistics. Consequently, calibration analyses using a fixed follow-up (N = 372) demonstrated higher predicted recidivism rates for Māori participants at each DRAOR score. This suggests that Māori participants with similar levels of DRAOR-assessed need factors as NZ European participants experienced relatively greater continued justice contact. DRAOR users should prioritize delivering quality case management to clients, recognizing that both case-specific and systemic factors may underlie differential base rates.

We have written an article describing our analysis of ratings recorded by corrections staff employed by Ara Poutama Aotearoa (New Zealand Department of Corrections) as they provided supervision services to Māori people in Aotearoa New Zealand.We identify our ethnicity as European Caucasian, and our ancestry as descendants of settlers and immigrants who arrived in countries with established peoples.We do not identify as Māori.
It is clear to us that we cannot represent and it is not our intention that our words would represent the views of Māori people nor communicate the stories of Māori people.Our article is written from our personal viewpoints, formed by our own ethnic identities and experiences of the world, shaped by our advantages and the sum of the systemic treatment we have characteristically received from others and societal institutions throughout our lives.We are outsiders to Aotearoa New Zealand, and our viewpoints similarly do not represent the perspectives of NZ Europeans nor staff at Ara Poutama Aotearoa (New Zealand Department of Corrections).
We wrote this article because we regularly contribute to the knowledgebase about a correctional case management tool, the Dynamic Risk Assessment for Offender Re-entry (DRAOR), thereby supporting its use by corrections agencies.We write about DRAOR in ways that both promote evidence of its helpful features and clarify its limitations.From this stance, we believed it was inappropriate to turn a blind eye to any potential limitations related to cross-ethnicity validity, when we had the opportunity to examine these limitations through our access to DRAOR data.
We appreciate that our voice in the broader and important conversation about overrepresentation of Indigenous peoples within international justice systems is limited and can only ever offer a narrow contribution, given our lack of first-hand perspective.However, we offer this article in the hope that it may contribute to positive solutions for better understanding and reducing this clear problem of overrepresentation.Perhaps you will perceive our intentions as naïve, but (as researchers) we are prone to believing that when we generate specific information about problems, this assists the pursuit of solutions.
For our readers who identify as Māori, we hope that this article can be useful, providing specific information that might guide specific self-advocating and self-determining choices and actions.We also acknowledge the helpful transparency of Ara Poutama Aotearoa (New Zealand Department of Corrections), both in providing these data for our analysis and regularly publishing information that identifies the nature of the problem of Māori overrepresentation.This is atypical, and it greatly enhanced our learning and data interpretation to be able to easily find publicly available information about group recidivism base rates, for example.
As we wrote in the article, we strongly reject and disavow any potential application of our research that would (1) undermine self-determination among Māori people, (2) implement correctional practices that are not humane, client-focused, and strength-building, or (3) perpetuate existing inequalities.More broadly, we believe the goals of Māori autonomy, equality, and positive experiences when engaging in correctional matters and processes cannot be achieved unless any implemented practice is transparent and repeatedly subjected to rigorous evaluation.We view our article as one small piece of this transparent evaluation process and invite any communication or criticisms of our work that will help advance these goals.

Table S1
Statistically Significant Standardized Factor Loadings Resulting From a Three-Factor Configural Exploratory Structural Equation Model of DRAOR Items with a Sample of NZ European and Māori People on Parole Italicized values indicate p >.001.Bolded values indicate strongest loading for each item.DRAOR = Dynamic Risk Assessment for Offender Re-entry (Serin, 2007); NZ = New Zealand; Ʌ = Standardized factor loading estimate; SE = Standard error.a n = 1,211.b n = 1,812.All thresholds significant (p <.001) unless otherwise specified.DRAOR = Dynamic Risk Assessment for Offender Re-entry (Serin, 2007); NZ = New Zealand; SE = Standard error.a p = .007.b p = .372.