Modeling Matters: Comparing the Presumptive Sentence Versus Base Offense Level Approaches for Estimating Racial/Ethnic Effects on Federal Sentencing
ObjectivesSince 2000, sentencing scholars have commonly controlled for racial/ethnic differences in underlying criminal conduct using presumptive sentence. In recent years, the presumptive sentence approach has been critiqued for (amongst other things) filtering out racial/ethnic disparities that ac...
| Main Author: | |
|---|---|
| Contributors: | |
| Format: | Electronic Article |
| Language: | English |
| Published: |
2024
|
| In: |
Journal of quantitative criminology
Year: 2024, Volume: 40, Issue: 2, Pages: 395-420 |
| Online Access: |
Volltext (lizenzpflichtig) |
| Journals Online & Print: | Drawer...
|
| Check availability: | HBZ Gateway |
| Keywords: |
| Summary: | ObjectivesSince 2000, sentencing scholars have commonly controlled for racial/ethnic differences in underlying criminal conduct using presumptive sentence. In recent years, the presumptive sentence approach has been critiqued for (amongst other things) filtering out racial/ethnic disparities that accumulate pre-sentencing. To circumvent this concern, a small but growing body of literature has begun to employ the base offense level approach. The goal of this study is to analyze the implications of using the presumptive sentence versus base offense level approach to isolate racial/ethnic effects on federal sentencing outcomes.MethodsUsing data from the United States Sentencing Commission (2018-2020), this analysis tracks racial/ethnic differences throughout the pre-sentence process (base offense level to final offense level). Subsequently, we compare racial/ethnic effects obtained in a series of multi-level multivariate regression models using both the presumptive sentence and base offense level approaches.ResultsFindings indicate that the two approaches provide vastly different starting points for racial/ethnic differences in underlying criminal conduct and, therefore, different conclusions about how race/ethnicity matters in sentencing. Most notably, Hispanic defendants are advantaged relative to Whites when accounting for racial/ethnic differences in base offense level but disadvantaged relative to Whites when accounting for racial/ethnic differences in presumptive sentence.ConclusionsFindings suggest that the presumptive sentence approach filters out important racial/ethnic differences in the pre-sentence process and that the two modeling approaches are not interchangeable. Results clearly indicate that modeling matters in sentencing research, and future research should pay close attention to their baselines for between-group differences in relevant conduct. |
|---|---|
| ISSN: | 1573-7799 |
| DOI: | 10.1007/s10940-023-09573-0 |