![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
Abstract
This study uses propensity score weighting to examine three key death penalty decisions in Pennsylvania from 2000–2010, focusing on the role of defendant and victim race: prosecutors’ decisions to seek the death penalty, prosecutors’ decisions to retract death filings, and decisions to sentence defendants to the death penalty. We collected data on 880 first degree murder convictions in 18 Pennsylvania counties, encompassing 87% of the state’s first-degree murder convictions. We do not find that black defendants, or black defendants who kill white victims specifically, are more likely to have the death penalty sought or imposed. Instead, we find that those who kill white victims, regardless of defendant race, are more likely to receive the death penalty. We further found that black defendants, and blacks who killed black victims, were more likely to have a death filing retracted by prosecutors. Finally, patterns of death penalty race disparity varied greatly depending on the county in which a case was prosecuted and sentenced.
Notes
1 According to the Pennsylvania Department of Corrections Execution List (Pennsylvania Department of Corrections, Persons Sentenced to Execution in Pennsylvania, December 1, 2016), there were two Asians (1.1%); seventeen Hispanics (9.7%); sixty-four Whites (36.6%) and ninety-two African-Americans (52.6%) under sentence of death in Pennsylvania. By contrast, in 2015, non-Hispanic Whites accounted for 77.4% of the state population, while the percentage African-Americans was 11.7%. U.S. Census Bureau, http://quickfacts.census.gov/qfd/states/42000.html as of July 1, 2015.
2 There are seven statutorily listed mitigating factors in Pennsylvania, and any other evidence of mitigation can also be introduced. Mitigating factors must be sustained by a preponderance of the evidence standard.
3 For example, see Donahue’s (2014) description of Connecticut’s statutory aggravating circumstances, and Paternoster et al.’s (Citation2004) description of Maryland’s.
4 This larger ongoing project focuses more broadly on between-county variation in homicide case processing and sentencing, and encompasses both quantitative and qualitative data collection. In connection with the qualitative dimension of this project, we have conducted interviews with prosecutors, judges, and defense attorneys that with substantial experience with first degree murder cases. Specifically, we have interviewed 11 District Attorneys and Assistant District Attorneys, eight judges, and six defense attorneys in four counties. These four counties are among those in the first degree murder data we analyze here. The interviews encompass questions about the charging decisions, the decision to seek the death penalty, the decision to retract the death penalty if sought, plea bargaining, and the sentencing process, among other topics.
5 While first and second degree murder are not covered by Pennsylvania’s sentencing guidelines, the PCS collects information on the sentencing of convictions for these offenses.
6 Due to the large number of Philadelphia cases, the Philadelphia District Attorney’s Office requested that we limited our data collection to cases charged from 2005–2010, as a condition of cooperation. This yielded 331 first degree murder convictions, and resulted in the omission of 250 cases prior to 2005.
7 The coauthors performed the coding of the cases in the 18 counties, along with a trained and supervised a team of student assistants. We pretested our field data collection codebook and strategy in two counties, and used this pretesting to create uniform coding rules. We attempted to minimize discretionary coder judgement as much as possible by coding the presence of case attributes (including aggravating and mitigating factors), victim characteristics, defense motions, evidence variables, etc. as present only if there was unambiguous positive mention or record of the factor in the police reports, docket transcripts, arraignments, court proceedings, death certificates, newspaper reports, etc. The field data collection codebook is available on request.
8 In our multivariate analyses, we include the small numbers of female defendants, but do not include gender as a control variable. In supplemental analyses, we omitted female defendants from the data, and results are nearly identical to those presented.
9 We conducted all propensity score weighting analyses using the “TEFFECTS” and “IPW” procedures in STATA statistical software, version 14. For documentation and a fuller explanations of propensity score analysis, see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3144483/ For details on propensity score methods procedures in STATA, see http://blog.stata.com/2015/07/07/introduction-to-treatment-effects-in-stata-part-1/ and http://www.stata.com/manuals13/te.pdf.
10 We compared the proportions of black and white defendants for whom prosecutors filed aggravating factors, and the proportions of black and white defendants for which coders found evidence of aggravating factors. Prosecutors appeared to file certain aggravating factors (i.e., history of violent felony convictions, and defendant created grave risk of death for another besides victim) for black defendants proportionally more frequently than white defendants. However, based on our independent coding of aggravating factors, we found that cases involving black defendants were proportionally more likely to exhibit the presence of aggravating factors, particularly the ones that prosecutors were more likely to file for blacks. Thus, where racial disproportionality existed in the filing of aggravating factors, this appeared to coincide with disproportionality in the presence of those aggravating factors in our independent coding.
11 In the case of multiple murder victims, the victim race/ethnicity variable indicates whether any of the victims were white, black, or Hispanic.
12 In supplemental models, we controlled for the county variables by including them in the propensity score model instead of in IPWR models. Results were substantively the same as those in the IPWR models that control for county in the tables we present.