The Problem of Uneven Contribution: Investigating the Implications of Place for System-Wide Estimates of Race/Ethnicity
ObjectivesThe analytic workhorse of modern sentencing research is the multi-level model—which gives more "empirical voice" to statistics coming from larger clusters. While an overall cluster balancing adjustment is certainly needed, it does not correct for subgroup unbalancing—a material i...
| Autores principales: | ; |
|---|---|
| Tipo de documento: | Electrónico Artículo |
| Lenguaje: | Inglés |
| Publicado: |
2025
|
| En: |
Journal of quantitative criminology
Año: 2025, Volumen: 41, Número: 2, Páginas: 215-239 |
| Acceso en línea: |
Volltext (lizenzpflichtig) |
| Journals Online & Print: | Gargar...
|
| Verificar disponibilidad: | HBZ Gateway |
| Palabras clave: |
| Sumario: | ObjectivesThe analytic workhorse of modern sentencing research is the multi-level model—which gives more "empirical voice" to statistics coming from larger clusters. While an overall cluster balancing adjustment is certainly needed, it does not correct for subgroup unbalancing—a material issue in the federal system where a few districts contribute an outsized number of minority defendants (i.e., the problem of uneven contribution). The goal of this study is to demonstrate how heavy saturation of racial/ethnic minority defendants in a select few districts obscures the generalizability of system-wide, multi-level, statistics.MethodsTo provide a more holistic understanding of the system-wide effects generated by the multi-level model, the current study supplies descriptive and multivariate analyses of racial/ethnic differences in federal sentencing outcomes ranging 2018 to 2021.ResultsOur main finding is that although the multi-level regression coefficient shows Hispanic federal defendants receive longer sentences than Whites (net of controls), most Hispanic federal defendants are sentenced in areas where they can expect to receive shorter sentences than Whites (net of controls). This seemingly incompatible finding exists for Black-White comparisons too—but to a lesser extent.ConclusionsA racially/ethnically clustered federal court system, an overall cluster size adjustment, and heterogenous statistical conclusion validity across districts, coalesce to deliver system-wide racial/ethnic effects reflecting the social world of the few. Seeing as people are sentenced, not places, the outsized voice of low minority contribution districts produces a misunderstanding of the scope and nature of racial/ethnic differences in sentencing. |
|---|---|
| ISSN: | 1573-7799 |
| DOI: | 10.1007/s10940-024-09597-0 |
