This Article examines 4,668 Oklahoma homicide cases with an identified suspect that occurred during a twenty-three year period between January 1, 1990, and December 31, 2012. Among these, we identified 153 cases that ended with a death sentence. Overall we found that while the defendant's race did not correlate with a death sentence, there was a strong correlation with the race of the victim, with cases with white victims significantly more likely to end with a death sentence than cases with non-white victims. Homicides with female victims were also more likely to result in a death sentence than other cases. We then examined whether the homicide included multiple victims and/or additional felony circumstances, and coded each case to indicate whether it included zero, one, or two of these "additional legally relevant factors. " Using logistic regression analysis, where the effects of each predictor variable can be isolated, the data indicate that 1) having a white female victim, 2) having a white male victim, 3) having a female victim from a minority race or ethnicity, 4) having one additional legally relevant factor, and 5) having two additional legally relevant factors present are statistically significant predictors of a death sentence. Overall, the data show that the odds of a death sentence for those with white female victims are 9.59 times higher than in cases with minority male victims. The odds of a death sentence for those with white male victims are 3.22 times higher than the odds of a death sentence with minority male victims. Finally, the odds of a death sentence for those with minority female victims are 8.68 times higher than the odds of a death sentence with minority male victims. All these race/gender effects are net of our two control variables (multiple murder victims and the presence of additional felony circumstances).
TABLE OF CONTENTS INTRODUCTION 734 I. PREVIOUS RESEARCH 738 II. METHODOLOGY 741 A. Homicide Data Set 741 B. Death Row Data Set 743 III. RESULTS 746 A. Frequencies and Cross-Tabulations 746 B. Multiple Logistic Regression Analysis 748 CONCLUSION 750 APPENDIX 751 INTRODUCTION
In the first sixteen years of the twenty-first century, we have seen several indicators that the use of the death penalty is in sharp decline in the United States. (4) According to the Death Penalty Information Center, between 1996 and 2000 an annual average of 275 new prisoners arrived on America's death rows, but by 2016 this figure decreased to thirty (a drop of almost 90%). (5) In addition, the average number of executions per year has also fallen over 50% since the last five years of the twentieth century, from seventy-four executions between 1996-2000, to thirty-three in the last five years, 2012-2016. (6) In just the past ten years, seven states have abolished the death penalty, (7) including Delaware in 2016, the most recent. (8) Four more states--Washington, Oregon, Colorado and Pennsylvania--have seen their governors impose formal or informal moratoria on executions. (9) A September 2016 poll by the Pew Research Center found that slightly less than half of Americans (49%) supported the death penalty, the lowest level of support in more than forty years. (10) A 2015 poll by Quinnipiac indicates that more Americans (48%) now prefer a sentence of life imprisonment without parole (which is available in all death penalty jurisdictions) to a death sentence (43%). (11)
The research reported in this article focuses on Oklahoma, a state that hosted 112 executions between 1976 and 2016, a tally second only to Texas. (12) Even in Oklahoma, a November 2015 poll found that the majority of the population (52%) prefer a sentence of life plus restitution rather than the alternative of the death penalty. (13) A second poll taken in July 2016 found that 53% of the "likely voters" in the state prefer life sentences without parole and mandatory restitution instead of the death penalty. (14) These results, coupled with national data, (15) clearly indicate a changing climate around death penalty debates: apparently more Americans now prefer long prison terms rather than the death penalty.
One reason for the decline in support for, and the use of, the death penalty is growing concern that the penalty is not reserved for the worst of the worst. (16) In a nationwide Gallup Poll taken in October 2015, 41% of the respondents expressed the belief that the death penalty was being applied unfairly, and a 2009 Gallup Poll found that 59% of the respondents believed that an innocent person had been executed within the preceding five years. (17) This concern is undoubtedly on the minds of many Oklahomans, where ten inmates have been released from its death row since 1972 because of (at least) serious doubts about guilt. (18)
In this article, we examine another question related to the contention that the death penalty is reserved for the worst of the worst: the possibility that the race of the defendant and/or victim affects who ends up on death row. To do so, we will study all of the homicides that occurred in Oklahoma from January 1, 1990 through December 31, 2012 and compare those cases with the subset that resulted in the imposition of a death sentence.
Oklahoma is home to some 3.75 million citizens, of whom 72% are white, with the African-American, Native American, and Hispanic population each constituting about 7 to 8% of the population. (19) Racial and ethnic minorities are over-represented among those on death row, which housed forty-six men and one woman as of July 1, 2016 (twenty-three white, twenty African-American, two Native American, two Hispanic). (20) However, to say that African-Americans are 8% of the Oklahoma population and 42.6% of those on death row (20/47) does not tell us much because the number of African-Americans convicted of homicide, or especially the most aggravated homicides (and not the number in the state's population), is the relevant denominator.
Between 1972 and March 1, 2017, Oklahoma conducted 112 executions (with the first occurring in 1990), which ranks second among U.S. states behind Texas and gives Oklahoma the highest per capita execution rate in the U.S. (21)
Of the 112 executed inmates, sixty-seven were white (60%), thirty-five African-American, six Native American, two Asian, one Hispanic, and one whose race was classified as "Other." (22) These 112 people were put to death for killing 116 victims. Eighty-three of the 112 executed inmates were convicted of killing at least one white victim (74.1%), nineteen at least one African-American victim, seven at least one Asian victim, five at least one Hispanic victim, one at least one Native American victim, and one who killed two people whose races are classified as "Other" (both the assailant and his two victims were Iraqi). (23)
Concerns about the impact of the defendant's and/or victim's race on death penalty decisions have a long history in the U.S. Soon after the 1976 decision in Gregg v. Georgia that breathed new life into death penalty statutes by reaffirming the constitutionality of capital punishment, (24) researchers, led by the late University of Iowa legal scholar David Baldus, began to study the possible relationships between race and the death penalty. The most comprehensive study by Baldus and his team focused on Georgia. (25) The twenty-eight race studies conducted prior to 1990, including the Baldus study, were reviewed by the U.S. government's General Accounting Office in 1990, which produced a report concluding that in 82% of the twenty-eight studies reviewed, "race of victim was found to influence the likelihood of being charged with capital murder or receiving the death penalty." (26)
In 2003, Baldus and George Woodworth in effect updated and expanded the GAO Report, reviewing eighteen studies on race and death sentencing that had been published or released after 1990. (27) Their conclusions are worthy of a lengthy quote:
Overall, their results indicate that the patterns documented in the GAO study persist. Specifically, on the issue of race-of-victim discrimination, there is a consistent pattern of white-victim disparities across the systems for which we have data. However, they are not apparent in all jurisdictions nor at all stages of the charging and sentencing processes in which they do occur. On the issue of race-of-defendant discrimination in the system, with few exceptions the pre-1990 pattern of minimal minority-defendant disparities persists, although in some states [African-American] defendants in white-victim cases are at higher risk of being charged capitally and sentenced to death than are all other cases with different defendant/victim racial combinations. (28) Overall, Baldus and Woodworth concluded that the studies displayed four clear patterns: (1) with few exceptions, the defendant's race is not a significant correlate of death sentencing, (2) primarily because of prosecutorial charging decisions, those who kill whites are significantly more likely than others to be sentenced to death, (3) African-American defendants with white victims are especially likely to be treated more punitively, and (4) counties with large numbers of cases with African-American defendants or white victims show especially strong death sentencing rates for African-American defendants and those with white victims. (29)
Professor Baldus passed away in 2011, but one of his students, Catherine Grosso, has taken the reigns and assembled a team that has continued Baldus's work. (30) Among the Grosso team's publications is one that recently updated the Baldus literature review. (31) Published in 2014, the Grosso team identified thirty-six studies that were completed after the 1990 GAO Report. (32) Their review identified four patterns:
* Four of the studies did not discover any race effects.
* Four found independent effects of the race of the defendant (that is, effects that remained after...