ABSTRACT
Studies find mixed results regarding the effects of race on criminal justice case processing, likely because they rarely account for the race of both the victim and offender, and few consider how prosecutorial case screening may influence later criminal justice stages. This study examines the impact of victim and defendant race on case screening, bail, and sentencing outcomes for 1,131 firearm offenses that occurred between 2015 and 2018 in St. Louis, MO. Regressions modeling the relationships between each outcome and victim and defendant race (estimated separately and as defendant-victim racial dyads) find that cases involving Black victims, alone and in combination with Black defendants, are more likely than others to be dismissed by prosecutors during case screening, whereas legally relevant factors affect bail and sentencing outcomes. The results suggest that disregarding initial gatekeeping stages of criminal justice case processing may lead to the mistaken conclusion that racial disparities do not exist.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1. For example, in 2017, the St. Louis Metropolitan Police Department (SLMPD) created a Gun Crimes Intelligence Center that uses recovered firearms and shell casings to investigate firearm-related crimes.
2. As Baumer and colleagues have noted (Baumer et al., Citation2000, p. 283), “A crime against an upstanding citizen may be perceived by judges and jurors as a greater harm, and thus as a more serious offense, than a crime against a disreputable person (e.g., Black Citation1989). In addition, sympathy for the victim within the community, and anger toward the defendant, are likely to be greater when the victim is someone who is considered reputable. If this is so, the demand for punishment may be weaker in cases involving victims who have engaged in disreputable conduct; judges and jurors may be less likely to indict or convict defendants in such cases.”
3. Other non-legal factors, such as whether victims had been drinking or had invited suspects into their residences prior to their assaults, have also been found to affect prosecutorial charging decisions (Beichner & Spohn, Citation2012).
4. Some recent work does highlight the importance of studying dyads (see, Donnelly & MacDonald, Citation2018; Johnson & Larroulet, Citation2019; Johnson, Citation2018; Kutateladze et al., Citation2016; Metcalfe & Chiricos, Citation2018; Stemen & Escobar, Citation2018; Stevenson, Citation2018).
5. See, (Bowers & Pierce, Citation1980; Paternoster, Citation1983; Sorensen & Wallace, Citation1999; Spohn et al., Citation2001; Spears & Spohn, Citation1996; Spohn & Spears, Citation1997; Ulmer et al., Citation2020; Vaughn, Citation2020b;; Worrall et al., Citation2006).
6. We include unauthorized use and possession offenses because they are important precursors to firearm violence. Further, while these offenses are felonies under Missouri law, they do not necessarily result in a prison sentence. Therefore, considerable variability exists in sentencing outcomes.
7. Unlawful use and unlawful possession of a firearm might appear to be crimes without victims. According to sections 571.101 to 571.121 of the Missouri code, however, crimes such as brandishing a weapon, discharging or shooting a firearm into a dwelling house, or shooting across a highway can have victims without being charged with assault on a person.
8. Five hundred and five arrests were officially charged, but data on the most serious Circuit Attorney charge were missing for two cases.
9. In our sample, the final bail amount starts at a minimum of $7,000 (prior to log-transforming the values). The sample has six cases in which bail was denied. These six cases were not included in the analysis for bail amounts. They were included in the final analysis for “incarceration,” as we include “days held pretrial” as a control (rather than bail amount) as prior research has suggested that pretrial detention and the amount of time held in pretrial detention are better predictors of sentence outcome than bail amount alone.
10. Bail decisions in the 22nd Circuit during the period of study were at the discretion of the presiding judge, and were not limited to a formalized bond schedule that was in place in the Eastern District Court of Missouri.
11. Some convicted defendants were sentenced to “shock” incarceration (typically 15–30 days), which were usually followed up by lengthy probation sentences. These cases are coded as 0.
12. “Mixed” gender and race refer to cases with both Black and White victims or both male and female victims.
13. Multicollinearity tests were run, with VIF’s ranging from 1.05 to 2.14. These values are well within the acceptable range, indicating no significant issues of multicollinearity.
14. The author(s) considered Heckman and two-part step models following Bushway et al. (Citation2007). However, there were no clear selection criteria and the current outcomes are not dependent on dichotomous outcomes to successfully control for selection (ie., offered bail followed by bail amount; sentenced to prison followed by prison sentence length). Therefore, the current models are not corrected for potential selection bias, and we believe any selection bias of cases that move to the bail or sentencing stage that may exist are shown in the significant effects predicting case refused.
15. The authors would like to note that the small sample sizes might produce inaccurate estimates related to the racial defendant/victim dyads (See, Berk et al., Citation2005 for a thorough discussion of small cell issues). We suggest that the results be interpreted with caution. However, it is important to note that the results for the models examining defendant/victim race independently (and without small cell issues) produce similar results to the dyad models, suggesting that the selection issue of small sample cells may not be a major concern in our models.