Indigenous-focused criminal justice programs are usually evaluated and studied without an adherence to, or acknowledgement of, Indigenous epistemologies, axiologies and ontologies. Indeed, many of the evaluations conducted of Australian Indigenous sentencing courts have relied on quantitative analyses of reoffending, finding little or no impact on recidivism, despite there being some evidence, derived mainly from qualitative analyses, that they have had an impact on strengthening informal social controls within Indigenous communities. This article uses published evaluations and impact studies of Indigenous sentencing courts as a case study to gain a better understanding of how these courts have been evaluated and researched, and how the methodological approaches used to study the courts may not properly capture the Indigenous-focused and community-building aims and goals of the programs. Employing a meta-review approach, the article examines why research that evaluates Indigenous-focused criminal justice programs should rethink how evaluations are framed and conducted when trying to determine 'what works'. 3 Cunneen and Rowe (2014) describe a Western positivist methodological approach as one that uses a scientific paradigm to separate, quantify, measure and statistically analyse individualised factors, while viewing these factors and the analysis as unrelated to the 'social, economic, cultural and political set of relations' within which they are embedded (2014:52). When researching questions related to the over-representation of Indigenous peoples within the criminal justice system, a Western positivist methodology also separates individualised factors from 'the historical experience of colonisation and dispossession ' (2014:52).