Microsoft Word - Beckett & Evans FINAL.docx 2016 COLUMBIA JOURNAL OF RACE AND LAW 77 RACE, DEATH, AND JUSTICE: CAPITAL SENTENCING IN WASHINGTON STATE, 1981-2014 KATHERINE BECKETT* & HEATHER EVANS** This Article examines the role of race in the application of the death penalty in the wake of the Furman v. Georgia decision. Although contemporary death penalty statutes were designed to reduce arbitrariness and discrimination in capital sentencing, many studies indicate that race continues to play a significant role in determining which capital defendants live and which die in the post-Furman era. To date, however, no published study has examined the role of race in capital sentencing in Washington State, where the statutory framework effectively reduces the number of homicide cases that are eligible for capital punishment and prosecutorial discretion is therefore comparatively circumscribed. This Article assesses whether race influences the administration of capital punishment in Washington State, and if so, where in the process it matters. On the one hand, the narrowness of the statutory framework may effectively constrain prosecutorial discretion in ways that minimize the role of race. On the other hand, experimental research suggests that unconscious stereotypes that link Blacks to violence are widespread, and that jury selection and deliberations tend to amplify jurors’ implicit biases. We therefore hypothesize that race will matter most at the sentencing (as opposed to the filing) stage of the process. To test this, we analyze prosecutorial and jury decision-making in all Washington aggravated murder cases adjudicated since 1981 for which information is available. The results of statistical regression analyses support this hypothesis: although neither the race of the defendant nor the victim affect prosecutorial decision-making, jurors are more than four times more likely to impose a death sentence when the defendant is Black. These findings suggest that race plays a significant role in capital sentencing even where the statutory framework effectively narrows the pool of homicide cases that may result in the death penalty. I. INTRODUCTION ............................................................................................... 78 II. RACE AND THE DEATH PENALTY: PAST AND PRESENT ................................. 84 A. Race And The Administration Of Capital Punishment In The United States ................................................................................................... 84 B. The Role Of Implicit Racial Bias ....................................................... 87 III. DATA, METHODS AND ANALYTIC STRATEGY .................................................. 89 A. Data And Analytic Strategy ................................................................. 89 B. Statistical Methods .............................................................................. 91 IV. PRELIMINARY FINDINGS: DESCRIPTIVE STATISTICS ...................................... 94 V. REGRESSION RESULTS ..................................................................................... 97 A. Factors Influencing Prosecutorial Filing Decisions In Aggravated Murder Cases ...................................................................................... 98 B. Factors Influencing The Imposition Of Death Sentences In Aggravated Murder Cases ................................................................. 100 78 Race, Death, and Justice Vol. 6:2 VI. DISCUSSION AND CONCLUSION .................................................................... 103 VII. APPENDIX A: AGGRAVATING FACTORS ......................................................... 107 VIII. APPENDIX B: TRIAL REPORTS EXCLUDED FROM THE ANALYSES ............... 109 IX. APPENDIX C: MEASUREMENT OF VARIABLES ................................................ 110 X. APPENDIX D: MODELING PROSECUTORIAL FILING DECISIONS ................... 112 XI. APPENDIX E: MODELING SENTENCING OUTCOMES IN CAPITAL SENTENCING .................................................................................................. 114 I. INTRODUCTION Although the number of executions taking place in the United States has declined considerably in recent years,1 capital punishment remains shrouded in controversy.2 Concerns about “the ultimate sanction” include the high cost of its administration, the apparent arbitrariness of its application, the possibility that available techniques cause considerable pain and suffering, and evidence that the system is “fraught with error.”3 The role of race in capital sentencing is also the subject of much discussion and debate.4 Indeed, although contemporary death penalty statutes were ostensibly designed to reduce arbitrariness and discrimination in capital sentencing,5 researchers have nonetheless found that race and other extra-legal factors continue to play a significant role in determining which capital defendants live and which die in the *Katherine Beckett, PhD 1994, University of California at Los Angeles; B.A. 1986, University of California at San Diego. ** Heather Evans, M.A. 2008, University of Washington; B.A. 2005, University of Washington. The authors would like to thank Lila Silverstein and Neil Fox for their assistance with data collection, Chase Beauclair for his excellent research assistance, and the editorial staff of the Columbia Journal of Race and Law for their helpful editing contributions. 1 The Death Penalty in 2013: Year End Report, Death Penalty Information Center, http://deathpenaltyinfo.org/documents/YearEnd2013.pdf. 2 See generally, CHARLES J. OGLETREE, JR. & AUSTIN SARAT, FROM LYNCH MOBS TO THE KILLING STATE: RACE AND THE DEATH PENALTY IN AMERICA (2006). Capital punishment is also controversial in Washington State. In 2014, Governor Jay Inslee expressed deep concerns about capital punishment and declared a moratorium on it. See Jennifer Sullivan & Andrew Garber, Inslee Halts Executions, SEATTLE TIMES (February 12, 2014), http://www.seattletimes.com/seattle-news/inslee-halts-executions-impact-on-current-cases-may-be-minimal/. In response, state prosecutors have asked lawmakers to introduce a referendum that will enable voters to determine the fate of capital punishment. See Associated Press, Washington Prosecutors Want Public Vote on Death Penalty, SEATTLE TIMES (November 13, 2015), http://www.seattletimes.com/seattle-news/crime/washington-prosecutors-want-death-penalty- referendum/. 3 See, e.g., James Liebman, Jeffrey Fagan & Valerie West, A Broken System: Error Rates in Capital Cases, 1973-1995, http://www2.law.columbia.edu/instructionalservices/liebman/; Justin Marceau, Sam Kamin & Wanda Foglia, Death Eligibility in Colorado: Many are Called, Few are Chosen, 84 U. COLO. L. REV. 1069 (2013). Regarding pain and suffering, see Larry Greenemeier, Cruel and Unusual?: Is Capital Punishment by Lethal Injection Quick and Painless?, SCIENTIFIC AMERICAN (October 27, 2010), http://www.scientificamerican.com/article/capital-punishment-by-lethal-injection/. 4 See, e.g., American Civil Liberties Union, Race and the Death Penalty, https://www.aclu.org/race-and-death- penalty. See also, STUART BANNER, THE DEATH PENALTY: AN AMERICAN HISTORY (2009); Craig Haney, Condemning the Other in Death Penalty Trials: Biographical Racism, Structural Mitigation, and the Empathic Divide, 53 DEPAUL L. REV. 1557 (2004); Charles J. Ogletree, Jr., Black Man’s Burden: Race and the Death Penalty in America, 81 OR. L. REV. 15, 18 (2002); Touré, Put to Death for Being Black: New Hope Against Judicial System Bias, TIME.com (May 03, 2012), http://ideas.time.com/2012/05/03/put-to-death-for-being-black-new-hope-against-judicial-system-bias/. 5 Stephen B. Bright, Discrimination, Death and Denial: The Tolerance of Racial Discrimination in the Infliction of the Death Penalty, 35 SANTA CLARA L. REV. 433, 433–34 (1995). 2016 COLUMBIA JOURNAL OF RACE AND LAW 79 post-Furman era.6 In particular, there is strong evidence that the race of murder victims influences the administration of the death penalty: many studies find that defendants accused of killing Whites are significantly more likely to be sentenced to death than similarly situated defendants accused of killing Blacks.7 About a third of the studies investigating capital sentencing processes since 1990 also find that the race of the defendant continues to impact outcomes in capital cases,8 even as overt and intentional forms of racism decline.9 To date, however, no published study has examined the role of race in capital sentencing in Washington State, where the death penalty is now a potential outcome in a very small proportion of all homicide cases. Washington State’s current death penalty statute was enacted in 1981 and notably limits the proportion of homicide cases in which the death penalty is a possible outcome.10 Under the Revised Code of Washington (“RCW”), Chapter 10.95, the death penalty may only be sought if the defendant is convicted of first-degree murder and at least one of fourteen aggravating circumstances is found to exist.11 By contrast, many death penalty states define a larger category of cases as death-eligible. As a result, these states tend to have larger death row populations and have executed a comparatively large number of defendants than is the case in Washington State. This variation is depicted in Table 1.12 6 Furman v. Georgia, 408 U.S. 238 (1972). For an overview, see David C. Baldus & George Woodworth, Race Discrimination and the Death Penalty, in JAMES R. ACKER, ROBERT M. BOHM & CHARLES S. LANIER, AMERICA’S EXPERIMENT WITH CAPITAL PUNISHMENT: REFLECTIONS ON THE PAST, PRESENT, AND FUTURE OF THE ULTIMATE PENAL SANCTION, 519–26 (2nd ed. 2003). See also, U.S. GOV’T ACCOUNTABILITY OFF., GGD-90-57, DEATH PENALTY SENTENCING: RESEARCH INDICATES PATTERN OF RACIAL DISPARITIES (1990); SAMUEL WALKER, CASSIA SPOHN & MIRIAM DELONE, THE COLOR OF JUSTICE: RACE, ETHNICITY AND CRIME IN AMERICA (4th ed. 2006); JAMIE L. FLEXON, RACIAL DISPARITIES IN CAPITAL SENTENCING: PREJUDICE AND DISCRIMINATION IN THE JURY ROOM (2012). 7 See Baldus & Woodworth, supra note 6. See also U.S. GOV’T ACCOUNTABILITY OFF., U.S. GOV’T ACCOUNTABILITY OFF., supra note 6, at 5; WALKER, SPOHN & DELONE, supra note 6; FLEXON, supra note 6. 8 U.S. GOV’T ACCOUNTABILITY OFF., supra note 6. 9 Lawrence Bobo, James R. Kluegel & Ryan A. Smith, Laissez-Faire Racism: The Crystallization of a ‘Kinder, Gentler’ Antiblack Ideology, in RACIAL ATTITUDES IN THE 1990S: CONTINUITY AND CHANGE (1997); Adam R. Pearson, John F. Dovidio & Samuel L. Gaertner, The Nature of Contemporary Prejudice: Insights from Aversive Racism, 3 SOC. & PERSONALITY PSYCHOL. COMPASS 314, 315 (2009). 10 Wash. Rev. Code § 10.95.030 (2015). 11 See Appendix A for a list of the aggravating factors that differentiate aggravated homicide from non- aggravated homicide in Washington State. 12 TRACY L. SNELL, BUREAU OF JUSTICE STATISTICS, CAPITAL PUNISHMENT, 2013 – STATISTICAL TABLES (2014); Jeffrey L. Kirchmeier, Casting a Wider Net: Another Decade of Legislative Expansion of the Death Penalty in the United States, 34 PEPP. L. REV. 1 (2006). Statutory information is current as of December 31, 2013 and is taken from SNELL (2014). The number of aggravators is taken from SNELL (2014), and where this information was not provided, from Kirchmeier (2006); the number of death row inmates and executions was taken from Death Row U.S.A., NAACP LEGAL DEFENSE AND EDUCATIONAL FUND, INC., http://www.deathpenaltyinfo.org/documents/DRUSAWinter2016.pdf. 80 Race, Death, and Justice Vol. 6:2 Table 1. State-level Variation in Death Penalty Statutes and the Imposition of Death State Qualifying Offense(s) Statutory Aggravators Death Row Inmates Executions Since 1976 Least Restrictive: Murder Plus Other Offenses are Death Eligible Arkansas Capital murder; treason 10 35 27 California First-degree murder with special circumstances; sabotage; train wrecking causing death; treason; perjury in a capital case causing execution of an innocent person; fatal assault by prisoner serving life 22 746 13 Colorado First-degree murder with at least one aggravating circumstance; first-degree kidnapping resulting in death; treason 17 3 1 Florida First-degree murder; felony murder; capital drug trafficking; capital sexual battery 15 401 90 Georgia Aggravated murder; rape; armed robbery or kidnapping with injury or ransom when the victim dies; aircraft high jacking; treason 11 82 57 Idaho First-degree murder with aggravating factors; first-degree kidnapping; perjury resulting in the execution of an innocent person 11 11 3 Kentucky Capital murder with at least one aggravating circumstance; capital kidnapping 8 34 3 Louisiana First-degree murder; treason 12 84 28 Mississippi Capital murder; airplane piracy 8 48 21 Montana Capital murder with at least one aggravating circumstance; aggravated kidnapping; felony murder; capital sexual intercourse 11 2 3 Wyoming First-degree murder; murder during the commission of sexual assault, sexual abuse of a minor, arson, robbery, burglary, escape, resisting arrest, kidnapping, or abuse of a minor under 16 12 1 1 Moderately Restrictive: Non-Aggravated Murder, Any Murder, and/or Homicide are Death Eligible Indiana Murder with at least one aggravating circumstance 16 14 20 Missouri First-degree murder 17 33 83 New Hampshire Murder committed in the course of rape, kidnapping, drug crimes, or home invasion; killing of a police officer, judge or prosecutor; murder for hire; murder by an inmate while serving a life without parole sentence 10 1 0 Texas Criminal homicide with at least one aggravating circumstance 9 271 524 Most Restrictive: Only Aggravated First Degree Murder Cases are Death Eligible Alabama Intentional murder with at least one aggravating factor 10 201 56 Arizona First-degree murder, including felony murder, with at least one aggravating factor 14 122 37 Connecticut* Capital felony murder 8 12 1 2016 COLUMBIA JOURNAL OF RACE AND LAW 81 Table 1. State-level Variation in Death Penalty Statutes and the Imposition of Death State Qualifying Offense(s) Statutory Aggravators Death Row Inmates Executions Since 1976 Delaware First-degree murder with at least one aggravating circumstance 22* 17 16 Kansas Capital murder with at least one aggravating circumstance 8 9 0 Nebraska First-degree murder with at least one aggravating circumstance 9* 11 3 Nevada First-degree murder with at least one aggravating circumstance 15* 81 12 New Mexico* First-degree murder with at least one aggravating circumstance 7 2 1 New York First-degree murder with at least one aggravating circumstance 13 0 0 North Carolina First-degree murder with at least one aggravating circumstance 11 157 43 Ohio First-degree murder with at least one aggravating circumstance 10 145 53 Oklahoma First-degree murder with at least one aggravating circumstance 8 48 112 Oregon First-degree murder with at least one aggravating circumstance 12* 36 2 Pennsylvania First-degree murder with at least one aggravating circumstance 18 184 3 South Carolina First-degree murder with at least one aggravating circumstance 12 44 43 South Dakota First-degree murder with at least one aggravating circumstance 10 3 3 Tennessee First-degree murder with at least one aggravating circumstance 17 73 6 Utah First-degree murder with at least one aggravating circumstance 19 9 7 Virginia First-degree murder with at least one aggravating circumstance 15 8 110 Washington First-degree murder with at least one aggravating circumstance 14 9 5 In the first, least restrictive category of states, state law authorizes capital punishment for offenses other than first-degree murder and/or in all first-degree murder cases. In Florida, for example, conviction of first-degree murder, felony murder, capital drug trafficking, and capital sexual battery may result in a death sentence. A second, moderately restrictive group of states limit capital punishment to murder cases but, unlike Washington State, treat non-aggravated and/or non-first-degree murder cases as death-eligible. In Missouri, for example, all first-degree homicides are death-eligible (i.e., no aggravating circumstances are 82 Race, Death, and Justice Vol. 6:2 required). Some of these states (such as Colorado) also define first-degree murder broadly, to include non- intentional forms of homicide such as felony murder and death caused by “extreme indifference.”13 The third, most restrictive group of states, which includes Washington, limits capital punishment to aggravated and first-degree murders. That is, prosecutors must prove beyond a reasonable doubt that the defendant committed first-degree murder and that one or more statutorily defined aggravators exist. Among these states, the number of statutory aggravators varies. In general, the more aggravating circumstances, the larger the share of murder cases that are likely to qualify as “aggravated” murder.14 Unlike Washington, many of the states in this category have expanded the number of aggravating circumstances or otherwise broadened their death penalty statute in recent years.15 In Pennsylvania, for instance, eighteen aggravated circumstances render a homicide death-eligible; in Tennessee, the statute now identifies seventeen aggravating circumstances.16 However, it is not just the number of aggravators that increases the number of death-eligible cases; some aggravators have broader applicability than others. For example, although Colorado’s statute identifies only a slightly larger number (seventeen) of aggravators than does Washington State law (fourteen), a recent study found that over ninety percent of Colorado’s first-degree murder cases met the statutory definition of aggravated murder and involved death-eligible defendants.17 By contrast, just 13.3% of Washington State’s first-degree murder defendants were convicted of aggravated murder and were death-eligible.18 As a result of its relatively narrow statutory framework, the number of executions and the size of the death row population are comparatively small in Washington. At present, nine men are on death row in Washington. Only seven of the thirty-three states with prisoners awaiting execution have smaller death row populations,19 and only eleven have conducted fewer executions since 1976.20 The fact that Washington’s statutory framework limits the applicability and use of capital punishment is of great importance: statutory schemes that do not sufficiently narrow the class of cases that are death- eligible create the potential for a high degree of arbitrariness and discrimination in the administration of capital punishment.21 In a number of cases adjudicated since Furman,22 the Court has ruled that capital sentencing statutes avoid the related problems of over-inclusiveness and arbitrariness if they “genuinely narrow the class of persons eligible for the death penalty.”23 As Marceau concludes, “the constitutionally 13 Marceau, Kamin & Foglia, supra note 3, at 1087. 14 See Kirchmeier, supra note 12. 15 See infra notes 67–69. 16 See SNELL, supra note 12. 17 Marceau, Kamin & Foglia, supra note 3. 18 This figure was calculated as follows: First, we obtained data regarding the number of aggravated and non- aggravated first-degree murder convictions from 1999-2013 from the Washington State Sentencing Guidelines Commission. See Statistical Summary of Adult Felony Sentencing, 1999-2013, WASHINGTON STATE CASELOAD FORECAST COUNCIL, http://www.cfc.wa.gov/CriminalJustice_ADU_SEN.htm. We then subtracted the number of aggravated murder convictions involving non-death eligible defendants from the total number of aggravated murder convictions. Defendants who were juveniles at the time of the conviction or subject to an extradition agreement that precluded the death penalty were considered ineligible for the death penalty. Finally, we calculated the percent of all first degree murder convictions that involved death-eligible defendants. 19 Death Row Inmates by State, DEATH PENALTY INFORMATION CENTER, http://www.deathpenaltyinfo.org/death-row-inmates-state-and-size-death-row-year (last updated Jan. 1, 2016). 20 Facts About the Death Penalty, DEATH PENALTY INFORMATION CENTER, http://www.deathpenaltyinfo.org/documents/FactSheet.pdf (last updated April 18, 2016). 21 Marceau, Kamin & Foglia, supra note 3, at 1094. 22 Furman v. Georgia, 408 U.S. 238 (1972). 23 Zant v. United States, 462 U.S. 862, 865 (1983). See also Marceau, Kamin & Foglia, supra note 3, at 1081–82 (discussing the narrowing of the class of persons eligible for the death penalty). 2016 COLUMBIA JOURNAL OF RACE AND LAW 83 required narrowing [of death-eligibility] must occur at the legislative level in order to limit the unchecked discretion of prosecutors in deciding whom to prosecute under a statute.”24 Because Washington State’s statutory framework does meaningfully narrow the class of homicide cases that are death-eligible, analysis of the role of race in capital cases in Washington is especially instructive: evidence that race matters even where discretion is comparatively constrained would suggest that narrowing the death penalty statute will not necessarily eliminate the role of race in the adjudication of capital cases. Despite its comparatively limited use, the administration of capital punishment in Washington State remains controversial.25 In part, this is because federal courts have overturned eight of eleven capital cases after defendants lost their appeals.26 But the racial composition of persons on Washington’s death row is also controversial.27 At present, four of the nine (forty-four percent) men on death row are Black, despite the fact that the percentage of African Americans in the state population has hovered between three to four percent for decades.28 This Article assesses whether race influences the administration of capital punishment in Washington State, and if so, where in the process it matters. Recent studies highlight the importance of analyzing prosecutorial and jury decision-making separately in order to specify which, if any, decision-making processes are influenced by race.29 This is especially important given Washington’s restrictive statutory framework, which dramatically limits the range of murder cases in which prosecutors may elect to seek death. The following analyses therefore explore the impact of race on prosecutorial decisions to file death notices and, separately, on juries’ decisions30 to impose capital punishment in aggravated murder cases in which death notices have been filed. Specifically, this Article examines whether prosecutors are more likely to seek, and juries more likely to impose, the death penalty in cases involving Black defendants. This Article also analyzes whether the race of the victim influences prosecutorial and/or jury decision-making in capital cases adjudicated in Washington State. Part II of this Article provides an historical overview of the role of race in the administration of capital punishment and the results of empirical studies investigating this relationship. 24 Marceau, Kamin & Foglia, supra note 3, at 1083. 25See Sullivan & Garber, supra note 2. See Associated Press, supra note 2. 26 As of 2000, the federal courts had overturned seven of eight cases upheld by the Washington State Supreme Court. These cases included Mak v. Blodgett, 970 F.2d 614 (9th Cir. 1992), cert. denied, 507 U.S. 951 (1993); Harris by and through Ramseyer v. Blodgett, 853 F. Supp. 1239 (W. D. Wash. 1994), aff’d, 64 F.3d 1432 (9th Cir. 1995); Rupe v. Wood, 93 F.3d 1434 (9th Cir. 1996), cert. denied, 519 U.S. 1142 (1997); Jeffries v. Wood, 114 F.3d 1484 (9th Cir. 1997); Rice v. Wood, C89-568T (W.D. Wash. 1997); Lord v. Wood, 184 F.3d 1083 (9th Cir. 1999); Benn v. Wood, 2000 WL 1031361 (W.D. Wash. 2000). The one exception was Campbell v. Wood, 18 F.3d 662 (9th Cir. 1994) (en banc). See ACLU of Washington, Sentenced to Death: A Report on Washington Supreme Court Rulings in Capital Cases 2 (2000). Since 2000, federal courts have overturned one of three death sentences upheld by the Washington State Supreme Court. In 2002, the Ninth Circuit reversed the death sentence in Pirtle v. Morgan, 313 F.3d 1160 (9th Cir. 2002). In 2007, however, the United States Supreme Court affirmed Cal Brown’s sentence. See Uttecht v. Brown, 551 U.S. 1 (2007). Jonathan Gentry also lost his federal appeal in the Ninth Circuit. See Gentry v. Sinclair, 705 F.3d 884 (9th Cir. 2012). 27 See, for example, the recent statement unanimously adopted by all nine Seattle councilmembers. Steve Militich, Seattle City Leaders Urge State Legislators to Abolish Death Penalty, SEATTLE TIMES (Jan. 26, 2015), http://blogs.seattletimes.com/today/2015/01/seattle-city-leaders-urge-state-legislators-to-abolish-death-penalty/. See also Mishi Faruqee, Facing Race and the Death Penalty, ACLU OF WASHINGTON BLOG (Dec. 11, 2012), https://aclu- wa.org/blog/facing-race-and-death-penalty. 28 Figures are current as of January 1, 2016, and are available through the Death Penalty Information Center. Current Death Row Populations by Race, DEATH PENALTY INFORMATION CENTER, http://www.deathpenaltyinfo.org/race- death-row-inmates-executed-1976#deathrowpop (last updated Jan. 1, 2016). 29 See DAVID C. BALDUS & GEORGE WOODWORTH, AMERICA’S EXPERIMENT WITH CAPITAL PUNISHMENT: REFLECTIONS ON THE PAST, PRESENT, AND FUTURE OF THE ULTIMATE PENAL SANCTION (2nd ed. 2003). 30 If a defendant waives his or her right to a jury trial, a judge may impose a death sentence in cases in which a death notice has been filed. As a practical matter, however, juries almost always decide whether to impose a sentence of death. We therefore link sentencing decisions to jury decision-making throughout this Article. 84 Race, Death, and Justice Vol. 6:2 Part III describes the data and methods used in the statistical analyses presented here. Part IV provides preliminary, descriptive results in order to provide the reader with a clear understanding of the broad geographic and demographic patterns that characterize the administration of capital punishment in Washington State. Part V presents the results of a series of statistical regression analyses designed to identify the unique impact of victim and defendant race over and above other relevant factors. Part VI offers a discussion of the empirical findings and their significance. II. RACE AND THE DEATH PENALTY: PAST AND PRESENT Historically, the use of capital punishment in the United States was bound up with various racialized systems of control, including extra-legal violence. As legal scholar Charles Ogletree, Jr. puts it, “the racially disproportionate application of the death penalty can be seen as being in historical continuity with the long and sordid history of lynching in this country.”31 Although it is tempting to imagine this continuity solely in historical terms, numerous studies indicate that race has continued to influence the administration of capital punishment in locales across the country since its reinstatement in the late 1970s and early 1980s.32 Some of these studies analyze data regarding the administration of capital punishment from particular jurisdictions within the United States.33 Others use experimental methods to investigate how the race of hypothetical defendants, victims, and/or jurors impact mock jurors’ deliberations and sentencing decisions.34 In what follows, this Article summarizes the results of these two bodies of research. A. Race And The Administration Of Capital Punishment In The United States Numerous studies analyze whether race has impacted the (actual) administration of capital punishment in United States since its reinstatement by the Supreme Court in the late 1970s.35 Importantly, most of these studies have been conducted in states that utilize capital punishment far more than Washington State.36 Research shows that race continues to permeate the capital sentencing process despite the adoption of procedures designed to eliminate that possibility. This appears to have been the case in the years immediately following the Furman decision37 and in more recent decades as well. However, while most studies focusing on the more recent period continue to find that the race of the victim influences capital outcomes, only about one-third of these studies find race-of-defendant effects. A meta-analysis of studies published prior to 1990 conducted by the United States Government Accountability Office (“GAO”) found “a pattern of evidence indicating racial disparities in the charging, sentencing and imposition of the death penalty after the Furman decision.”38 Studies published during this period consistently reported that defendants convicted of killing Whites were more likely to be sentenced to death than other defendants, over and above any differences in case characteristics.39 Indeed, this finding was “remarkably consistent across data sets, states, data collection methods, and analytic techniques”; it was also found to exist at all stages of the criminal justice process.40 Moreover, more than half of the studies reviewed by the GAO indicated that the race of the defendant also significantly impacted the likelihood that defendants were charged with a capital offense and sentenced to death prior to 1990. In three-fourths of these studies, 31 See Ogletree, Jr., supra note 4. See also OGLETREE, JR. & SARAT, supra note 2. 32 See supra notes 7–8. See infra notes 49–56, 59, 66–69, 72. 33 See supra notes 7–8. See infra notes 49–56, 59. 34 See infra notes 66–69, 72. 35 Gregg v. Georgia, 428 U.S. 153 (1976). 36 See infra notes 50–65, 66–73. 37 Furman v. Georgia, 408 U.S. 238 (1972). 38 U.S. GOV’T ACCOUNTABILITY OFF., supra note 6, at 5. 39 U.S. GOV’T ACCOUNTABILITY OFF., supra note 6, at 5. 40 U.S. GOV’T ACCOUNTABILITY OFF., supra note 6, at 5. 2016 COLUMBIA JOURNAL OF RACE AND LAW 85 Black defendants were significantly more likely to face a death sentence than similarly situated White defendants.41 More recent studies report similar findings.42 In particular, studies analyzing more recent time periods fairly consistently report that victim-race and numerous other legal and extra-legal factors continue to influence the administration of capital punishment.43 Specifically, most studies find that defendants convicted of killing Whites are significantly more likely to receive a death sentence than others, even after controlling for a wide range of legal and extra-legal factors that may also influence outcomes in capital cases.44 For example, Songer and Unah analyzed capital sentencing in South Carolina in the 1990s and found that prosecutors were significantly more likely to seek death in cases involving White victims.45 Similarly, Barnes, Sloss, and Thaman analyzed the imposition of the death penalty in cases adjudicated in Missouri between 1997 and 2001, and report that defendants accused of killing Whites were significantly more likely to be sentenced to death than other defendants after controlling for other relevant factors.46 Radelet and Pierce analyzed the factors that predict the imposition of death sentences in eligible murder cases adjudicated in North Carolina between 1980 and 2007, and found that defendants accused of killing Whites were more likely to be sentenced to death than similarly situated others.47 Numerous other studies have reached similar conclusions.48 Several recent studies also found that the race of the defendant influences outcomes in capital cases, with Black defendants more likely to be sentenced to death than similarly situated White defendants.49 For example, Baldus reports that in cases adjudicated in Philadelphia between 1983 and 1993, Black defendants (and defendants accused of killing people who were not Black) were significantly more likely to be sentenced to death than similarly situated others.50 Baldus similarly reports that Black defendants in the United States military system were more likely to be sentenced to death than non-Black defendants even after controlling for relevant legal factors.51 Another recent study analyzing data from Arkansas found that Black defendants with White victims were significantly more likely to be sentenced to death than Black defendants with non- 41 U.S. GOV’T ACCOUNTABILITY OFF., supra note 6, at 6. 42 In a meta-analysis of the literature published in 2003, Baldus and Woodworth found that “in [eighty-three percent] (25/30) of the jurisdictions with relevant data, there is some evidence of race-of-victim disparities adversely affecting defendants whose victims are White, and in [thirty-three percent] (10/30) of these jurisdictions, there is some evidence of race-of-defendant disparities adversely affecting Black defendants.” See Baldus & Woodworth, supra note 6, at 519. 43 See Baldus & Woodworth, supra note 6, at 519 44 See Baldus & Woodworth, supra note 6, at 520–21. 45 Michael J. Songer & Issac Unah, The Effect of Race, Gender and Location on Prosecutorial Decisions to Seek the Death Penalty in South Carolina, 58 S.C. L. REV. 161 (2006). 46 Katherine Y. Barnes, David L. Sloss & Stephen C. Thaman, Life and Death Decisions: Prosecutorial Discretion and Capital Punishment in Missouri 58 (Ariz. Legal Studies, Discussion Paper No. 08-03, 2008). 47 Michael L. Radelet & Glenn L. Pierce, Race and Death Sentencing in North Carolina, 1980-2007, 89 N. C. L. REV. 2119, 2140–42 (2011). 48 See David C. Baldus, Catherine M. Grosso, George G. Woodworth & Richard Newell, Racial Discrimination in the Administration of the Death Penalty: The Experience of the United States Armed Forces (1984-2005), 101 J. CRIM. L. & CRIMINOLOGY 1227 (2011); Glenn L. Pierce & Michael L. Radelet, Death Sentencing in East Baton Rouge Parish, 1990-2008, 71 LA. L. REV 647 (2011); Glenn L. Pierce & Michael L. Radelet, The Impact of Legally Inappropriate Factors on Death Sentencing for California Homicides, 1990-1999, 46 SANTA CLARA L. REV. 1 (2005); Scott Phillips, Continued Racial Disparities in the Capital of Capital Punishment: The Rosenthal Era, 50 HOUS. L. REV. 131 (2012). 49 Baldus & Woodworth, supra note 6, at 519. 50 David C. Baldus, George Woodworth, David Zuckerman, Neil A. Weiner & Barbara Broffitt, Racial Discrimination and the Death Penalty in the Post-Furman Era: An Empirical and Legal Overview, with Recent Findings from Philadelphia, 83 CORNELL L. REV. 1638 (1998). 51 Baldus, Grosso, Woodworth & Newell, supra note 50. 86 Race, Death, and Justice Vol. 6:2 White victims and all White defendants.52 Even more recently, Donohue found that in Connecticut, minority defendants accused of killing White victims were substantially more likely to be charged and sentenced to death than other similarly situated defendants.53 Many of these studies further suggest that race may matter the most at the sentencing phase of capital trials. For example, Baldus examined capital cases in Philadelphia and found that Black defendants were significantly more likely to be sentenced to death after controlling for a host of other relevant factors.54 In fact, it is conceivable that the absence of race-of-defendant effects in some published studies is a consequence of the failure to analyze prosecutorial and jury decision-making processes separately. Studies also indicate that other extra-legal factors influence the administration of capital punishment.55 For example, some researchers have found that defendants convicted of killing women or children, and those who used a knife, are more likely to receive the ultimate sanction.56 The type of location also appears to matter, with defendants sentenced in rural and suburban areas more likely to be sentenced to death than their urban counterparts.57 In sum, there is substantial evidence that race and other extra-legal factors have continued to impact capital sentencing processes in locales across the country: most studies report that the race of the victim has a significant impact on capital case outcomes, and some, though not all, find that the race of the defendant also influences the administration of capital punishment. Studies that analyze prosecutorial and jury decision- making separately often find that race matters most at the sentencing phase of the capital process.58 Evidence that race continues to matter in capital cases challenges the widespread belief that we are, in the post-Furman era, 59 “post-racial.” Although overt, conscious, and intentional racism has diminished considerably in recent years, a number of studies show that both structural racism—racially unequal outcomes that flow from facially neutral institutional arrangements, policies or practices—and implicit (i.e., unconscious) racial bias persist.60 Indeed, numerous experimental studies of implicit bias show that race affects perception and decision-making even in the absence of overt racial animus or antipathy. Below, we provide a brief overview of experimental studies examining the impact of implicit bias on jury deliberations and capital sentencing. 52 David C. Baldus, Julie Brain, Neil A. Weiner & George Woodworth, Evidence of Racial Discrimination in the Use of the Death Penalty: A Story from Southwest Arkansas (1990-2005) with Special Reference to the Case of Death Row Inmate Frank Williams, Jr., 76 TENN. L. REV. 555 (2009). 53 John J. Donohue III, An Empirical Evaluation of the Connecticut Death Penalty System since 1973: Are there Unlawful Racial, Gender and Geographic Disparities?, Stanford Law School, Working Paper No. 464 (2014), http://ssrn.com/abstract=2470082. 54 Baldus, Woodworth, Zuckerman, Weiner & Broffitt, supra note 51, at 1714. 55 Donohue III, supra note 53, at 696. 56 See Songer & Unah, supra note 45, at 191–97. 57 See Baldus & Woodworth, supra note 6, at 520; Donohue III, supra note 53, at 52–57. 58 Baldus, Woodworth, Zuckerman, Weiner & Broffitt, supra note 50, at 1715–22; Isaac Unah, Empirical Analysis of Race and the Process of Capital Punishment in North Carolina, 2011 MICH. ST. L. REV. 609, 646–48 (2011). 59 Furman v. Georgia, 408 U.S. 238 (1972). 60 Bobo, Kluegel & Smith, supra note 9. 2016 COLUMBIA JOURNAL OF RACE AND LAW 87 B. The Role Of Implicit Racial Bias Researchers refer to the unconscious impact of race as “implicit bias” in order to differentiate it from conscious racial animus.61 A significant and growing body of research suggests that implicit biases are pervasive, even among individuals who do not openly express biased views.62 For example, experimental studies show that stereotypes such as the association between Blackness and violence are widespread: the mere visual presence of a Black man increases the likelihood that observers will think about the concepts with which Black men are stereotypically associated (e.g., violence), interpret ambiguous behavior as aggressive, and miscategorize ambiguous objects as weapons. 63 Moreover, the association between Blackness and violence is bi-directional: images of Blackness bring violence and criminality to mind, while discussions of violence conjure images of Blackness in the minds of many.64 These studies provide compelling evidence that the unconscious association between Blackness and violence is widespread and influences how people perceive and interpret behavior, objects and social situations. With respect to capital sentencing, numerous studies show that implicit racial bias shapes the identification and processing of death-eligible cases. For example, researchers using experimental methods to examine implicit biases among jury-eligible citizens in six leading death penalty states found that many citizens harbored implicit racial stereotypes about Blacks and placed more value on the lives of Whites.65 Moreover, the more mock jurors showed implicit racial bias, the more likely they were to sentence Black defendants to death. 66 Similarly, experimental studies show that death-qualified jurors who viewed a simulated California capital trial were significantly more likely to recommend death when the video depicted the defendant as Black rather than White.67 Experimental studies further suggest that the association between Blacks and violence is mediated by an unconscious but widespread association between Black human beings and animals.68 This dehumanizing association appears to influence basic cognitive processes and significantly alter judgments in criminal justice contexts.69 For example, researchers conducting an archival study of capital cases found that news stories about Black defendants convicted of capital crimes are significantly more likely to contain ape-relevant language than news stories about White capital defendants. Moreover, defendants depicted in more ape-like ways are more likely to be executed than others even after controlling for relevant legal factors.70 In follow- 61 Robert J. Sampson & Stephen W. Raudenbush, Seeing Disorder: Neighborhood Stigma and the Social Construction of “Broken Windows”, 67 SOC. PSYCHOL. Q. 319, 320–21 (2004). See also Lincoln Quillian, Does Unconscious Racism Exist?, 71 SOC. PSYCHOL. Q. 6, 7 (2008). 62 See Task Force on Race and the Criminal Justice System, PRELIMINARY REPORT ON RACE AND WASHINGTON’S CRIMINAL JUSTICE SYSTEM, 17–21 (2011), http://www.law.washington.edu/About/RaceTaskForce/preliminary_report_race_criminal_justice_030111.pdf. 63 B. Keith Payne, Prejudice and Perception: The Role of Automatic and Controlled Processes in Misperceiving a Weapon, 81 J. PERS. & SOC. PSYCHOL. 181, 190 (2001) (discussing the miscategorization of objects for weapons); Joshua Correll, Bernadette Park, Charles M. Judd & Bernd Wittenbrink, The Police Officer’s Dilemma: Using Ethnicity to Disambiguate Potentially Threatening Individuals, 83 J. PERS. & SOC. PSYCHOL. 1314, 1324–28 (2002). 64 Jennifer L. Eberhardt, Phillip Atiba Goff, Valerie J. Purdie & Paul G. Davies, Seeing Black: Race, Crime and Visual Processing, 87 J. PERS. & SOC. PSYCHOL. 876, 877 (2004). 65 Justin D. Levinson, Robert J. Smith & Danielle M. Young, Devaluing Death: An Empirical Study of Implicit Racial Bias on Jury-eligible Citizens in Six Death Penalty States, 89 N.Y.U. L. REV. 513 (2014). 66 Id. 67 Tara L. Mitchell, Ryann M. Haw, Jeffrey E. Pfeifer & Christian A. Meissner, Racial Bias in Mock Juror Decision- Making: A Meta-Analytic Review of Defendant Treatment, 29 LAW & HUMAN BEHAV. 621, 631 (2005). See also Samuel R. Sommers, Race and the Decision-Making of Juries, 12 LEGAL & CRIM. PSYCHOL. 171, 177 (2007). 68 Phillip Atiba Goff, Jennifer L. Eberhardt, Melissa J. Williams & Matthew Christian Jackson, Not Yet Human: Implicit Knowledge, Historical Dehumanization, and Contemporary Consequences, 94 J. PERS. & SOC. PSYCHOL. 292 (2008). 69 Id. at 303–04. 70 Id. 88 Race, Death, and Justice Vol. 6:2 up studies, researchers found that the degree to which Black defendants have a stereotypically Black appearance is an important predictor of the imposition of a death sentence in cases involving Black defendants and White victims.71 There is, then, ample evidence that implicit racial biases are widespread and affect decision-making in general and in capital cases specifically. But bias is not just a property of individuals; its expression can also be encouraged or suppressed by social context. 72 Recent scholarship indicates that jury selection and deliberation dynamics tend to encourage the expression of implicit racial bias and favor death sentences.73 For example, experimental studies indicate that mock jurors are more likely to shift their initial position from life to death than from death to life. Researchers attribute this pattern to the emotional strategies and tactics employed by pro-death jurors.74 White male jurors play an especially important role in encouraging such verdicts by utilizing “a panoply of powerful emotion-based tactics to sway others to their position in a manner that often contributes to racially biased outcomes.”75 This racial effect may also stem in part from the fact that jurors are less likely to give credence to mitigating evidence offered on behalf of Black defendants.76 In fact, findings from experimental studies suggest that evidence regarding mitigating circumstances that may be perceived as exculpatory for White defendants is often ignored or even interpreted as incriminating when defendants are Black.77 For example, although being raised in an abusive home is often interpreted as a mitigating circumstance for White defendants, the same background often works to pathologize Black defendants, casting violence as a way of life.78 Moreover, the process of death qualification may amplify implicit biases, and jury deliberations have been shown to exacerbate the tendency of mock White jurors to sentence Black defendants to death more frequently than White defendants.79 The fact that jury selection and deliberation processes encourage the expression of implicit biases and favor death sentences may help explain the fact that jurors are more likely to impose death sentences, and are more influenced by the age and race of the victim and defendant, than are judges.80 In short, a wide body of literature shows that implicit racial biases have a powerful impact on decision-making in both real and simulated capital cases, and are especially likely to be activated in the sentencing phase of capital trials. In the next section, this Article investigates the possibility that race may also affect capital case processing in Washington State, where death may only be imposed in the most serious, aggravated homicide cases and prosecutorial discretion is comparatively circumscribed. Based on our analysis of the nature of Washington’s statutory framework, which reduces the scope of prosecutorial discretion, as well as evidence that jury selection and deliberation processes tend to elicit and amplify implicit racial biases, 71 Jennifer L. Eberhardt, Paul G. Davies, Valerie J. Purdie-Vaughns & Sheri Lynn Johnson, Looking Deathworthy: Perceived Stereotypicality of Black Defendants Predicts Capital-Sentencing Outcomes, 17 PSYCHOLOGICAL SCIENCE 383, 384 (2006) (discussing a study in which participants were asked to rate the stereotypicality of Black faces based on a variety of facial features as well as skin tone and the correlation between these features and case outcomes is assessed). 72 Mona Lynch & Craig Haney, Emotion, Authority and Death: (Raced) Deliberations in Mock Capital Jury Deliberations, 40 LAW & SOC. INQUIRY 377 (2015). 73 Id. 74 Id. at 391–94. 75 Id. at 394–401. 76 Id. See also Mona Lynch & Craig Haney, Looking Across the Empathic Divide: Racialized Decision Making on the Capital Jury, 2011 MICH. ST. L. REV. 573 (2011). 77 Id. at 574. 78 Lynch & Haney, supra note 72, at 398. 79 Mona Lynch & Craig Haney, Capital Jury Deliberation: Effects on Death Sentencing, Comprehension, and Discrimination, 33 LAW & HUMAN BEHAV. 481, 491, 493 (2009); see also Lynch & Haney, supra note 72. 80 Radha Iyengar, Who’s the Fairest in the Land? Analysis of Judge and Jury Death Penalty Decisions, 54 J. L. & ECON. 693, 695–96, 708 (2011). 2016 COLUMBIA JOURNAL OF RACE AND LAW 89 this Article hypothesizes that race is most likely to matter at the sentencing rather than filing phase of the capital process. Before testing this hypothesis, this Article provides a brief overview of the data and methods. III. DATA, METHODS AND ANALYTIC STRATEGY A brief historical overview of the statutory history and framework in Washington State helps to contextualize our study. In 1854, the Territorial Legislature adopted the death penalty as an automatic penalty for anyone convicted of first-degree murder.81 In 1909, the legislature authorized the imposition of either a sentence of death or life imprisonment for persons convicted of that crime, thus rendering the death penalty non-mandatory.82 In the aftermath of the decision in Furman v. Georgia,83 the Washington legislature abolished the death penalty altogether, but then reinstated it in 1977 after adding additional procedures designed to reduce arbitrariness in its application.84 This statute was eventually declared unconstitutional because it specified that defendants who pled guilty would not receive a sentence of death, while defendants who exercised their right to a trial remained vulnerable to that sanction.85 Washington’s existing death penalty statute was enacted in 1981, and stipulates that only aggravated first-degree murder convictions can result in a death sentence.86 Life without the possibility of parole is the only other possible sentence in these aggravated murder cases.87 If prosecutors intend to seek a death sentence, they must file a Notice of a Special Sentencing Proceeding (sometimes referred to as a death notice). If such a notice is filed, and the defendant is convicted of aggravated first-degree murder, then a special sentencing proceeding takes place absent a legal development that precludes a special sentencing proceeding. Trial judges are required to file reports in all aggravated murder cases.88 These trial reports are required in order to facilitate proportionality review. Specifically, RCW 10.95.130(2)(b) mandates that the Court determine whether “the sentence of death is excessive or disproportionate to the penalty imposed in similar cases, considering both the crime and the defendant.”89 “Similar cases” means all cases resulting in one or more convictions for aggravated murder, regardless of whether a death sentence was sought or imposed.90 The purpose of this review “is to ensure that the sentence, in a particular case, is proportional to sentences given in similar cases, is not freakish, wanton or random; and is not based on race or other suspect classifications.”91 A. Data And Analytic Strategy This study analyzes data derived from trial reports pertaining to aggravated murder cases filed with the Washington State Supreme Court between December 1981 and May 2014—a total of 332 cases.92 It is important to note that trial reports are produced only in cases in which prosecutors establish that one or 81 L.G. Hellwig, Death Penalty in Washington: An Historical Perspective, 57 WASH. L. REV. 525 (1982). 82 Act of March 22, 1909, ch. 249, § 140, 1909 Wash. Sess. Laws 890, 930 (repealed 1975). 83 Furman v. Georgia, 408 U.S. 238 (1972). 84 Act of June 10, 1977, ch. 206, §§3–7, 1977 Wash. Sess. Laws 774, 776–79, invalidated by State v. Frampton, 627 P.2d. 922 (Wash. 1981). 85 Id. 86 Wash. State Bar Ass’n, Final Report of the Death Penalty Subcommittee to the Committee on Public Defense 5 (2006). 87 Id. 88 Id. at 6. 89 Wash. Rev. Code § 10.95.130 (2010). 90 Id. 91 State v. Cross, 132 P.3d 80, 104 (Wash. 2006). 92 Although there were 332 trial reports describing different special sentencing proceedings, we refer to a total of 331 cases because the hearing described in Trial Report 292 was not an aggravated murder case, Trial Reports 34 and 34A reflect two separate aggravated murder convictions for one defendant, and the defendant in the case described in Trial Report 16A was convicted prior to the adoption of the current sentencing statute. 90 Race, Death, and Justice Vol. 6:2 more aggravating circumstances apply. Because our data derive mainly from trial reports, we are unable to assess whether some homicide cases could have been determined to be aggravated (and therefore death- eligible) murder cases, but were not. Although it is conceivable that extra-legal factors such as race influence this process, we are unable to assess whether this is the case given the nature of the data analyzed here. Twenty-nine of the aggravated murder cases for which trial reports are available involved defendants who were under the age of eighteen at the time of the offense. These individuals were not eligible for the death penalty. In 1993, the Washington State Supreme Court determined that juveniles are ineligible for the death penalty.93 In this ruling, the Court also construed the statute to mean that the death penalty could never have been imposed upon juveniles. For these reasons, minors were excluded from the analyses presented here. Three adult defendants were also ineligible due to extradition agreements that precluded the imposition of a death sentence. In addition, two defendants were excluded because it was determined after the trial report was completed that, in one instance, the crime was committed before the effective date of the current statute and, in the second case, the defendant was not convicted of aggravated murder. (See Appendix B for the list of trial reports excluded from the analyses.) After these exclusions, the sample includes 298 aggravated first-degree murder cases involving death- eligible adult defendants. Prosecutors sought the death penalty in 29.1 percent of the cases involving death- eligible adults (87 out of 298), and juries imposed the death penalty in 42.7 percent of the cases in which death notices were filed and not later withdrawn (35 out of 82). Some of these death sentences were overturned on appeal. Of the 298 adults convicted of aggravated murder in Washington State between December 1981 and May 2014, five have been executed and another nine are currently on death row.94 The trial reports were coded according to a detailed coding protocol. 95 Two University of Washington students were trained to enter information from the trial reports into an Excel spreadsheet; these assistants cross-checked their work to ensure reliability. In other words, subjective judgments were not required. 96 For example, whether there was extensive publicity surrounding the case in question was determined by judges who checked either a yes or no box in response to this question; our data entry assistants did not make this judgment, but rather simply recorded whether the box was checked. Although the trial reports ask judges to supply information about a wide range of case, defendant, and victim characteristics, we discovered through the data entry process that many of the trial reports were incomplete. We were therefore unable to include a number of potentially relevant factors (such as defendant IQ and mental health status) in our analyses that may also influence the administration of capital punishment. Nevertheless, the coding process yielded a fairly comprehensive database that included information about numerous case, victim, and defendant characteristics. In addition, we supplemented the data gleaned from trial reports with information regarding victim demographics from newspaper articles and other case materials. We also compiled data regarding several county characteristics for inclusion in our analyses. Consistent with other recent studies of the role of race and other extra-legal factors in capital case processes, 93 State v. Furman, 858 P.2d 1092 (Wash. 1993). 94 See Washington State Department of Corrections, Capital Punishment in Washington State, http://www.doc.wa.gov/offenderinfo/capitalpunishment/. 95 This coding protocol was developed and implemented in consultation with appellate defense attorneys Lila Silverstein and Neil Fox. 96 The only exception was the mitigating circumstances variable, about which judges often hand-wrote notations that required some legal knowledge to interpret. For this reason, we relied on the legal expertise of Mr. Gregory’s attorneys in coding this variable. In addition, research assistants were asked to record whether judges indicated in words that a victim’s suffering was prolonged or allowed to endure over time. However, this variable was not included in the final analyses because sensitivity analysis revealed that it was consistently insignificant. 2016 COLUMBIA JOURNAL OF RACE AND LAW 91 we included several county-level variables, including measures of population density, demographic composition, and voting behavior, all of which were compiled from the United States Census Bureau. In addition, data regarding county revenues and expenditures were taken from the Office of Financial Management’s Washington State Data Book. We used the Bureau of Labor Statistics’ online inflation calculator to convert county revenue figures to constant 1981 dollars for meaningful comparison. 97 Appendix C provides a detailed description of the source and measurement for each variable. In the aggravated murder cases this Article analyzes, prosecutors may or may not have filed a death notice. If a death notice was filed and a special sentencing proceeding occurred, juries may (or may not) have imposed a sentence of death. The analyses presented here employ regression methods to assess the role of race in the two main stages of capital sentencing in Washington State. Specifically, this Article analyzes: a) prosecutorial decisions to file a death notice; and b) whether a death sentence was imposed in cases in which a special sentencing proceeding occurred. These regression analyses allow us to ascertain whether the race of the victim and/or defendant influence either prosecutorial decisions to file a death notice and/or decisions by juries to impose a sentence of death. Part IV provides descriptive information regarding the prevalence and distribution of death sentences in Washington State. Part IV begins by comparing the proportion of cases in which death notices were filed and death sentences were imposed at the county level in order to assess whether there is significant county-level variation in the seeking and imposition of death sentences. Next, Part IV compares the proportion of Black, White, and other defendants who were convicted of aggravated murder against whom prosecutors filed death notices, who were sentenced to death, and who have been executed or are currently on death row. Finally, Part IV compares the proportion of cases involving a Black defendant and White victim that resulted in a death sentence with the proportion of cases with different defendant-victim configurations in which a death sentence was sought or imposed. The results of these preliminary, descriptive analyses show that there is notable variation in the proportion of aggravated murder cases in which prosecutors seek, and juries impose, the death penalty at the county level. They also suggest that prosecutors filed death notices in a larger share of cases involving White than Black defendants. By contrast, a comparatively large proportion of Black defendants against whom death notices were filed were sentenced to death. It is important to note, however, that these descriptive results are suggestive rather than conclusive because they do not take into account the many legal factors that may influence prosecutorial and jury decision-making. To remedy this, Part V presents the results of statistical regression analyses that assess whether race impacts key outcomes when a full array of case characteristics are taken into account.98 B. Statistical Methods Regression is a statistical technique used to estimate the degree of correlation among variables included in a given model. Regression models include an outcome or dependent variable—in this case, a death notice or death sentence—as well as a number of factors (independent variables) that may affect the outcome. The results of the regression analysis reveal how much the outcome changes when any one of the independent variables is varied and the other independent variables are held constant. Regression analysis 97 Detailed information about the sources and measurement of the variables analyzed is presented in Appendix C. 98 A report completed in October 2014 presented several of these models. Three data errors have been corrected since that time. In addition, the case described in Trial Report 34A was included in the analysis. These corrections very slightly changed the size of individual regression coefficients. See Washington State Department of Corrections, Capital Punishment in Washington State, http://www.doc.wa.gov/offenderinfo/capitalpunishment/. 92 Race, Death, and Justice Vol. 6:2 thus allows researchers to identify the unique impact of each independent variable—in this case, the race of the defendant and victim—over and above the impact of the other variables included in the model. By convention, social scientists often identify statistical significance when the p-value is less than .05. This provides a rejection region of 2.5 percent at either end of the sampling distribution, leading us to reject the null hypothesis when results fall above or below those cut off points.99 However, when hypotheses are directional (e.g., the researcher expects covariates to increase and not decrease the probability of receiving the death penalty), the five percent rejection region is weighted at one side of the sampling distribution, leading us to reject the null hypothesis when the p-value is less than .10. Because few (if any) studies find that White defendants are more likely to be sentenced to death than are Black defendants, but many studies find that Black defendants are more likely to be sentenced to death than others, we hypothesize that the impact of a defendant’s race (Blackness) will increase, and not decrease, the odds that a death sentence is imposed. We therefore report the p-values of covariates that are statistically significant at both the .05 and .10 levels.100 Diagnostic tools were used to help identify the most appropriate regression models. When cases are nested in groups, such as counties, multilevel analysis is often used to isolate the statistical impact of the individual county. Fixed effects models are another common strategy to control for shared error among observations belonging to the same group. However, these methods are not appropriate in this case due to the unequal distribution of cases across counties. In the data analyzed here, twenty-eight counties are represented, but twenty-one counties have had fewer than ten death penalty cases since December 1981. Roughly ten percent of the aggravated murder cases occurred in counties with fewer than five trials. In fact, twelve counties have heard only one capital case since 1981. In these counties, then, there is little or no variation in the dependent variable. Multilevel analysis is therefore not appropriate for this analysis.101 Instead, we fitted logistic regression models, each with an outcome of 0 or 1, using Maximum Likelihood Estimate (“MLE”) procedures to estimate the probability of receiving a death notice or death sentence given a number of independent variables. In general, MLE estimates should be interpreted with caution for samples with fewer than 100 cases.102 As a precaution, we conducted careful sensitivity analyses, including and excluding case and county characteristics to gauge their impact on the pattern of results. We also conducted diagnostics to determine whether the results were unduly influenced by any outliers (individual cases that exhibit statistical leverage and/or influence).103 We present parsimonious models that 99 The null hypothesis asserts that there is no statistical association between two variables. ALAN AGRESTI & BARBARA FINLAY, STATISTICAL METHODS FOR THE SOCIAL SCIENCES 164–69 (3d ed. 1997). 100 Inclusion of the p ≤ .1 threshold is appropriate for several other reasons as well. First, it is arguable on ethical (and Constitutional) grounds that the risk of a falsely negative conclusion, which would be to conclude that race is not significant in the context of capital sentencing when in fact it is significant, is greater than the risk of a false positive, which is believing that the race of defendant matters when it does not. Second, other published studies of capital sentencing include an alpha level of .10. See David C. Baldus, Catherine M. Gross, George Woodworth and Richard Newell, Racial Discrimination in the Administration of the Death Penalty: The Experience of the United States Armed Forces (1984-2005), 101 JOURNAL OF CRIMINAL LAW & CRIMINOLOGY 1227 (2012); John Donahue III, Empirical Evaluation of the Connecticut Death Penalty System Since 1973: Are There Unlawful Racial, Gender and Geographic Disparities?, 11 JOURNAL OF EMPIRICAL LEGAL STUDIES 637 (2014). It is also worth noting that p-values are less important when populations rather than samples are analyzed, because the results are not used to draw inferences or generalize to other populations. See ALBERTO ABADIE, SUSAN ATHEY, GUIDO W. IMBENS & JEFFREY M. WOOLDRIDGE, FINITE POPULATION CAUSAL STANDARD ERRORS, Working Paper 20325, http://www.nber.org/papers/w20325 101 See generally TOM A. B. SNIJDERS & ROEL J. BOSKER, MULTILEVEL ANALYSIS: AN INTRODUCTION TO BASIC AND ADVANCED MULTILEVEL MODELING (1999). 102 See SCOTT J. LONG & JEREMY FREESE, REGRESSION MODELS FOR CATEGORICAL DEPENDENT VARIABLES USING STATA (2d ed. 2006). 103 Diagnostic tests revealed one potential outlier. However, removing this case from the analysis had no meaningful impact on the results, and it is therefore included in our analyses. 2016 COLUMBIA JOURNAL OF RACE AND LAW 93 include theoretically and substantively important variables and findings that are robust across various model specifications. As in similar studies conducted in other venues, two types of variables were included in the regression models: case characteristics, some of which we would expect to impact case outcomes, and extra- legal factors (such as race), which ideally would not. In the analysis of prosecutorial decision-making, we included case characteristics that have been found to be significant in some studies and would have been known to prosecutors early in the criminal process: the number of prior convictions possessed by the defendant; the number of victims; the number of aggravators alleged by prosecutors to exist (rather than found by juries); whether the defendant was suspected of also committing a sex crime in the course of the homicide; and whether the victim was a law enforcement officer. After assessing the role of case characteristics, we added several extra-legal factors to the models, as suggested by prior studies. In the analysis of prosecutorial discretion, these included: race of the defendant and the victim(s); whether the victim was female, a child, or a stranger; whether the defendant used a gun; and whether there was extensive publicity about the case. We also tested four county-level variables: the population density of the county in which the conviction occurred; the percentage of the county population that is Black; per capita county revenue; and the share of the county population that voted Republican in the most proximate Presidential election. In the analysis of jury decision-making, we included case characteristics that have statutory or other legal significance and would likely have been known by judges and jurors. These include the number of prior convictions possessed by the defendant; whether there were multiple victims; the nature of the defendant’s plea (guilty vs. not guilty);104 the number of aggravating circumstances found by the jury; the number of mitigating circumstances; the number of defenses offered; and whether the victim was held hostage. 105 Drawing on prior studies, we also tested the significance of a number of social factors, including the race, gender, and age of victims, and whether the victim was known to the defendant. County-level factors, including racial composition, political orientation, per capita county revenue, and population density were also tested. For each set of regression analyses, we present the results obtained when case characteristics and defendant race are included in the model. In the analysis of prosecutorial decision-making, we are able to simultaneously include a variety of other victim, case and county characteristics in the model as well. Unfortunately, the smaller number of cases means that not all of these factors can be included in the analysis of jury decision-making at once. Model testing, sometimes referred to as sensitivity analysis (in which the impact of these variables is tested in a variety of combinations), suggested that the only extra-legal factor that was consistently relevant to the outcome is the race of the defendant. However, given that many studies find that the race of the victim is significant, the results obtained when victim race is included in the model are also presented. Finally, we present the results obtained when a range of victim and county characteristics are included in the model in a step-wise manner. These results allow us to assess the degree to which outcomes in aggravated murder cases are influenced by race and other extra-legal factors over and above any differences in case characteristics. Diagnostic tests indicated that a handful of cases are univariate outliers with respect to the number of victims. We therefore measured the number of victims in terms of three categories: one victim; two to four 104 In the majority of cases, the same judge or jury served during both the guilt phase and the sentencing phase of the trial. In such cases, the judge or jury would have known whether the defendant pled guilty. In the small number of cases in which different decision-makers deliberated during the guilt and sentencing phases of the process, the nature of the defendant’s plea may not have been known to the jurors serving in the sentencing phase. 105 In these analyses, we treat evidence that the victim was held hostage as a measure of victim suffering. 94 Race, Death, and Justice Vol. 6:2 victims; or five or more victims. Diagnostics also showed that three variables were heavily skewed. These included the number of prior convictions, number of mitigating circumstances, and county per capita revenue. Logging these variables normalized their distribution and improved the model fit. The number of defenses and aggravators also showed some signs of skew, but after testing, the model fit was better (assessed by comparing log likelihood values, likelihood chi-squared ratio tests, and pseudo R2 scores) when these variables were not logged.106 Logged variables are identified in the tables with the standard notation (ln). IV. PRELIMINARY FINDINGS: DESCRIPTIVE STATISTICS The descriptive statistics presented below provide a preliminary overview of the distribution of efforts to obtain, and decisions to impose death sentences by county and across various groups of defendants. Table 2 shows the proportion of aggravated murder cases involving adult defendants in which prosecutors filed a death notice and in which a death sentence was imposed across Washington State counties. All counties in which five or more aggravated murder cases occurred between December 1981 and May 2014 are identified individually. We also include information about the average number of victims and aggravators present in the cases adjudicated in each county. As Table 2 makes evident, the proportion of aggravated murder cases in which prosecutors seek death varies notably. In Thurston County, prosecutors sought the death penalty in sixty-seven percent of all aggravated murder cases. By contrast, prosecutors in Okanogan County did not seek the death penalty in any of the eight cases that took place there. In larger counties with more aggravated murder cases, the proportion of cases in which prosecutors sought death also varied markedly, from a high of forty-eight percent in Kitsap County to a low of zero percent in Yakima County. The proportion of cases involving defendants who had special sentencing proceedings in which juries imposed a sentence of death also varies notably. Focusing on counties with more than one such case, the proportion ranges from a high of seventy-five percent in Clark County to twenty-five percent in Clark County. Moreover, it does not appear that these differences are a function of the number of victims or aggravating circumstances involved in the relevant cases. 106 See Appendix C for detailed information about variable measurement and transformation. 2016 COLUMBIA JOURNAL OF RACE AND LAW 95 Table 2. Proportion of Aggravated Murder Cases with Death-Eligible Defendants in which Death was Sought and Imposed, by County, December 1981 – May 2014 County Proportion of Aggravated Murder Cases in which Death Notices were Filed Proportion of Aggravated Murder Cases in which Death Notices were Filed (and Retained) and the Death Penalty was Imposed Average Number of Victims in Cases Adjudicated in the County Average Number of Aggravators in Cases Adjudicated in the County Thurston 67% (4/6) 50% (2/4) 1 2 Clallam 50% (3/6) 67% (2/3) 2 2 Kitsap 48% (10/21) 25% (2/8) 1 3 Pierce 46% (25/54) 44% (11/25) 2 2 Spokane 40% (8/20) 29% (2/7) 1 2 Snohomish 23% (7/31) 71% (5/7) 1 3 King 22% (16/72) 40% (6/15) 3 2 Clark 18% (4/22) 75% (3/4) 1 3 Whatcom 17% (1/6) 100% (1/1) 1 2 Benton 13% (1/8) 100% (1/1) 2 2 Cowlitz 13% (1/8) 0% (0/1) 1 2 Skagit 0% (0/5) NA (0/0) 1 3 Okanogan 0% (0/8) NA (0/0) 1 1 Yakima 0% (0/9) NA (0/0) 2 1 All Washington State Counties 29% (87/298) 43% (35/82) 2 2 Note: Counties with five or more aggravated murder cases are individually identified, but all counties are included in the totals. The figures shown in Table 2 thus suggest that the likelihood that prosecutors will seek and juries will impose death for a given aggravated murder defendant may depend in part on the county the case is adjudicated. Table 3 compares the proportion of Black, White and other death-eligible defendants against whom prosecutors filed a death notice and, among those who had a special sentencing proceeding, the share that received a death sentence. This table also shows the proportion of cases in which death sentences survived the appeals process for each racial group.107 The findings indicate that prosecutors sought death sentences in a larger proportion (thirty-three percent) of aggravated murder cases involving White defendants than they did in cases involving Black (twenty-five percent) or other (twenty percent) defendants. However, juries imposed death in forty percent of the cases involving White defendants, but sixty-four percent of the 107 This information was provided to us by attorneys Lila Silverstein and Neil Fox. 96 Race, Death, and Justice Vol. 6:2 cases involving Black defendants, in which prosecutors filed a death notice. Interestingly, the death penalty has been retained in a larger proportion of cases involving Black defendants (twenty-nine percent) than it has in cases involving White (fourteen percent) or other (ten percent) defendants.108 Table 3. Capital Sentence Outcomes among Death-Eligible Washington State Aggravated Murder Defendants, December 1981 – May 2014, by Race of Defendant White Black Other Race All Death Notice Filed 32.6% (62/190) 24.6% (14/57) 20% (10/50) 29.2% (87/298) Death Penalty Imposed 40.4% (23/57) 64.3% (9/14) 30% (3/10) 42.7% (35/82) Death Penalty Retained 14% (8/57) 28.6% (4/14) 10% (1/10) 15.9% (13/82) Note: Defendant race is unknown in one case; the category “All” therefore includes one case more than the sum of Whites (190), Blacks (57) and Other Race (50). Prosecutors filed death notices in 87 cases. In three of these cases, defendants were later determined to be ineligible for special sentencing proceedings (trial reports 68, 217, and 308). In two additional cases, defendants subsequently entered a stipulated guilty plea and a special sentencing hearing therefore did not occur (trial reports 152 and 153). These defendants are not included in the figures above (n=82). In light of research indicating that the race of victims often influences the likelihood that defendants receive the death penalty, Table 4 compares outcomes for Black and White defendants convicted of killing a single White victim versus a single Black victim. (This analysis is limited to cases involving just one victim mainly because it is difficult to categorize cases involving multiple victims with different racial identities). The results show that prosecutors sought death in a slightly larger share of cases involving White victims and either Black or White defendants (twenty-eight and twenty-nine percent, respectively) than in cases involving a Black defendant and Black victim (twenty percent). However, a death sentence was imposed in a substantially larger proportion of cases involving Black defendants than it was in cases involving White defendants—regardless of the race of the victim. For example, of cases involving a single White victim, death sentences were imposed in twenty-four percent of the cases involving White defendants but seventy-one percent of those involving Black defendants. These figures provide preliminary support for the hypothesis that the race of the defendant, but not the race of the victim, significantly impacts sentencing outcomes in capital cases. Interestingly, the death penalty has also been retained in a larger share (twenty-nine percent) of cases involving a Black defendant and White victim than in cases involving other racial configurations. 108 “Retained” in this context means the death sentence was not reversed by a higher court or was re-imposed after reversal of the original death sentence. 2016 COLUMBIA JOURNAL OF RACE AND LAW 97 Table 4. Capital Case Outcomes Among Death-Eligible Washington State Aggravated Murder Defendants, December 1981 – May 2014, by Race of Defendant and Race of Victim Defendant/Victim Race Death Notice Filed Death Penalty Imposed Death Penalty Retained Black Defendant/ White Victim 28% (7/25) 71.4% (5/7) 28.6% (2/7) Black Defendant/ Black Victim 20% (1/5) 100% (1/1) 0% (0/1) White Defendant/ White Victim 28.8% (34/118) 24.2% (8/33) 9.1% (3/33) White Defendant/ Black Victim 0% (0/0) 0% (0/0) 0% (0/0) Note: Figures include only Black and White “death eligible” defendants with one White or Black victim. In summary, the preliminary, descriptive findings presented above suggest that counties vary notably in their propensity to seek and impose death in aggravated murder cases. They also provide support for the hypothesis that the race of the defendant influenced decisions to impose (but not seek) the death penalty in aggravated murder cases adjudicated in Washington State from December 1981 to May 2014. Prosecutors filed death notices in a larger share of cases involving White (thirty-three percent) versus Black (twenty-five percent) defendants. By contrast, among cases in which special sentencing proceedings occurred, juries imposed death in forty percent of the cases involving White defendants, but sixty-four percent of the cases involving Black defendants. However, it is conceivable that this stark racial disparity is a function of case characteristics rather than of race itself. For example, if cases involving Black defendants have, on average, more aggravating circumstances or fewer mitigating circumstances than cases involving White defendants, this could explain why juries sentence Black defendants to death more frequently than they do White defendants. In the next section, this Article presents the results of regression analyses that control for these and other factors and then isolates the unique impact of race on case outcomes. V. REGRESSION RESULTS We present two sets of regression analyses. The first set analyzes the impact of case characteristics and a variety of social factors on prosecutors’ decisions to file a death notice.109 The second set identifies the case characteristics and social factors that influence sentencing decisions in capital cases in which a death notice was filed and not withdrawn. As noted previously, multivariate regression analysis identifies significant relationships between the independent variables included in the model and the outcome (dependent) variable. The coefficients provide a measure of the direction and strength of the correlation between each potential explanatory variable and the outcome being analyzed. The direction of the association (i.e., whether the coefficient has a negative or positive value) indicates whether the variable causes a decrease or an increase in the likelihood of receiving a death notice or a death sentence. Coefficients resulting from a logistic MLE model are presented as log-odds. In order to facilitate interpretation, we convert these to odds ratios. Odds ratios reveal how changes in the independent variable impact the odds of the outcome of interest. For example, the odds ratio for Black defendant presented in our first model of sentencing outcomes is 4.646, which means that juries are 4.6 times more likely to sentence a defendant to death when he/she is Black. 109 In three of these cases, death notices were filed but judges later ruled that the defendants were ineligible for a special sentencing proceeding. In two additional cases, defendants subsequently entered a stipulated guilty plea and a special sentencing hearing therefore did not occur. In the analyses of prosecutorial decision-making, we include all cases in which prosecutors filed a death notice. However, in the analyses of jury decision-making, we include only cases in which a death notice was filed and a special sentencing proceeding occurred. See Appendix B for trial report numbers. 98 Race, Death, and Justice Vol. 6:2 It is important to note that the results of these analyses identify which of the explanatory variables included in the model are significantly associated with the dependent variable holding all other variables included in the model constant. That is, regression analysis simultaneously takes a number of factors into consideration and identifies the unique impact of each variable on the outcome. If the regression results indicate that being Black is positively and significantly associated with being sentenced to death, this would mean that defendants who are Black are more likely to be sentenced to death after taking all other variables in the model into account. A. Factors Influencing Prosecutorial Filing Decisions In Aggravated Murder Cases The regression models presented below assess the extent to which a variety of case characteristics predict whether prosecutors filed a death notice in aggravated murder cases involving death-eligible defendants. These models include case characteristics that are evident in the early stages of criminal processing: the (logged) number of prior convictions; the number of victims; whether the defendant was also suspected of committing a sex crime; whether the victim was a law enforcement officer; and the number of aggravating circumstances alleged by prosecutors.110 We included the number of aggravating circumstances alleged by prosecutors because prosecutors do not yet know how many of these aggravating circumstances will be found by the jury. Because the defendant’s plea is sometimes entered after prosecutors have decided whether to seek death, it is not included as a potential predictor in this analysis. Table 5 shows the results that are obtained when the case characteristics identified above as well as a variety of defendant, victim, and county characteristics are included in the model. Note that the coefficients are log-odds. Negative values indicate that the predictor reduces the probability that prosecutors filed a death notice; positive coefficients indicate that the variable in question increased the probability that prosecutors filed a death notice. Odds ratios that are less than one mean that the variable in question reduces the likelihood that prosecutors filed a death notice; odds ratios that are greater than one mean that the variable of interest increased the probability that this occurred. In this model, cases are categorized as having either one or more than one victim.111 The results indicate that the number of prior convictions and aggravators, sex crime allegations, and law enforcement victims are significant after controlling for a variety of social factors. Specifically, each of these variables significantly increase the odds that prosecutors filed a death notice. For example, the odds ratio for alleged aggravating circumstances is 1.4, meaning that each additional aggravator increased the likelihood that prosecutors sought death by 1.4. Victim race (measured as all victims were White compared to cases in which victims were not all White) is marginally significant (p=.087). However, model testing (i.e., sensitivity analysis) indicates that this finding is not consistent across model variants.112 The results further indicate that the race of the defendant does not significantly impact prosecutorial decision-making. Victim- gender and age also appear to be irrelevant at this stage of the criminal process. 110 We also assessed whether prosecutors were more likely to file a death notice if the victim had been held hostage. The regression results indicated that this case characteristic did not have a significant impact on prosecutorial decision-making. Because this information was missing in twenty-one cases, including this variable in the model would have significantly reduced the number of cases analyzed. We therefore elected not to include it in the models presented here. 111 To present the most parsimonious model, we compare the effect of having one victim versus multiple victims in the second regression analysis. 112 For example, when “victim stranger” is omitted from the model above, “victim race” becomes non- significant (p-value = .199). Similarly, when “child victim” is omitted the model above, victim race becomes non- significant (p-value = .107). 2016 COLUMBIA JOURNAL OF RACE AND LAW 99 However, the results indicate that some extra-legal factors do influence prosecutorial decision- making. Specifically, whether a case received extensive publicity significantly impacts prosecutors’ decisions: prosecutors were more than three times more likely to seek death in cases characterized by extensive publicity than they were in cases that were not highly publicized (according to the judge). In addition, the size of the Black population in the county in which the case was adjudicated significantly impacts the likelihood that prosecutors will file a death notice in aggravated murder cases. These two findings are highly significant at a p-value < 0.01. The results also indicate that prosecutors in counties with relatively large Black populations are significantly more likely to file death notices than other prosecutors. This finding is consistent with a significant body of evidence indicating that demographic factors generally, and the size of the Black population specifically, have an important impact on criminal justice outcomes.113 Table 5. Impact of Case Characteristics and Extra-Legal Factors on Prosecutorial Decisions to File Death Notices in Aggravated Murder Cases, December 1981 – May 2014 N = 266 Death Notice Filed R2 = 0.2164 Variable Coefficient Exact P-value Odds Ratio Referent (Compared to) Case Characteristics Prior Convictions (ln) 0.169 0.015 1.184** 1 Victim -0.148 0.701 0.863 Multiple victims Alleged Aggravators 0.322 0.008 1.379*** Sex Crime 0.865 0.069 2.376* Sex crime not indicated Law Enforcement Officer 1.477 0.024 4.378** Non-police victim(s) Social Factors Black Defendant -0.548 0.237 0.578 Non-Black defendants Extensive Publicity 1.301 0.002 3.672*** No extensive publicity Victim Characteristics White Victim(s) 0.744 0.087 2.105* Non-White victims Female Victim(s) -0.017 0.965 0.983 Males/both sexes Stranger Victim(s) -0.259 0.466 0.771 Non-stranger victims Child Victim(s) 0.538 0.216 1.713 Adult victim(s) County Characteristics Percent Republican 0.002 0.928 1.002 Population Density -0.002 0.041 0.998** Percent Black 0.376 0.000 1.457*** Per Capita Revenue(ln) -0.139 0.770 0.870 * significant at α = .10 ** significant at α = .05 *** significant at α = .01 113 Many studies have found that the racial composition of the population is a significant predictor of enhanced penalty. See, e.g., Katherine Beckett & Bruce Western, Governing Social Marginality: Welfare, Incarceration, and the Transformation of State Policy, 3 PUNISHMENT & SOC’Y. 43 (2001); George S. Bridges & Robert D. Crutchfield, Law, Social Standing and Racial Disparities in Imprisonment, 66 SOC. FORCES 699 (1987-1988); Clayton Mosher, Predicting Drug Arrest Rates: Conflict and Social Disorganization Perspectives, 47 CRIME & DELINQUENCY 841 (2001); Cassia Spohn & David Holleran, The Imprisonment Penalty Paid by Young, Unemployed Black and Hispanic Male Offenders, 38 CRIMINOLOGY 281 (2000); Darrell Steffensmeier & Stephen Demuth, Ethnicity and Judges’ Sentencing Decisions: Hispanics-Black-White Comparisons, 39 CRIMINOLOGY 145 (2001). 100 Race, Death, and Justice Vol. 6:2 Note: In this model, thirty-two cases (10.8%) were missing data and were therefore dropped from the analysis. The finding that neither the race of the victim nor of the defendant impact prosecutorial decision making in (potentially) capital cases is consistent with our hypothesis that Washington’s statutory framework restricts prosecutorial discretion, thus reducing the likelihood that race will impact the exercise of prosecutorial discretion. B. Factors Influencing The Imposition Of Death Sentences In Aggravated Murder Cases Death notices were filed and special sentencing proceedings occurred in 82 cases involving adults charged with aggravated murder. Capital punishment was imposed in 35 (forty-three percent) of these cases. The next set of regression analyses identifies the factors that influence the decision to impose a sentence of death in these cases. Because these analyses only include cases in which prosecutors filed a death notice and a special sentencing proceeding occurred, the sample size is notably smaller than it was in the previous analyses (n = 82). As a result, the number of predictors that can be included in the models at any one time is limited.114 Below, we present the results obtained when only case characteristics and defendant race are included in the model. Although model testing consistently revealed the race of victim to be insignificant, we also present the results obtained when victim race is also included in this model because it has been found to be a significant predictor in many other studies. The model presented in Table 6 includes only case characteristics that would have been known by judges and jurors: the (logged) number of prior convictions; the number of victims (included here as a binary variable for one victim/multiple victims); the number of applied aggravators (as determined by the jury); the (logged) number of mitigating circumstances identified; the number of defenses offered; and whether the victim was held hostage. 115 It also includes the race of the defendant (comparing Black to non-Black defendants) and the race of the victim (comparing cases with one or more White victims against those in which victims were not exclusively White). These results indicate that the number of aggravators and defenses have a significant impact on sentencing decisions. Specifically, each additional aggravator increased the odds that a jury would impose death by 1.9, and each additional defense offered reduced the odds that a jury would impose death by 0.5. The race of the victim is not found to be a significant predictor of sentencing outcomes (p=.47). By contrast, the results indicate that a defendant’s race has a significant impact on sentencing outcomes in aggravated murder cases in which prosecutors file death notices. Specifically, Black defendants are 4.5 times more likely to be sentenced to death than similarly situated non-Black defendants, after controlling for all other variables included in the model (p = .053). 114 See Appendix Table E1 for complete descriptive statistics regarding the variables included. 115 The defendant’s plea was determined through model testing to be insignificant and, in order to accommodate the inclusion of defendant race, is not included in this model. We also tested the significance of a concomitant sex crime and whether the victim was a law enforcement officer. Neither of these factors was found to be a significant predictor of sentencing outcomes, and they are therefore not included in the model. 2016 COLUMBIA JOURNAL OF RACE AND LAW 101 Table 6. Impact of Case Characteristics, Defendant Race, and Victim Race on Capital Sentencing Outcomes in Death Eligible Cases, December 1981 – May 2014 N = 77 Death Penalty Imposed Pseudo R2 = .2153 Variable Coefficient Exact P-Value Odds Ratio Referent (Compared to) Prior Convictions (ln) -0.087 0.528 0.916 1 Victim -0.653 0.274 0.520 Multiple victims Applied Aggravators 0.646 0.013 1.908** Mitigating Circumstances (ln) -0.253 0.103 0.777 Defenses -0.737 0.050 0.478* Victim Held Hostage 0.746 0.206 2.108 Not held hostage Black Defendant 1.511 0.053 4.529* Non-Black Defendant White Victim -0.545 0.469 0.580 Non-White Victim * significant at α = .10 ** significant at α = .05 *** significant at α = .01 Note: In this model, five cases (6.1%) were missing data and were therefore dropped from the analysis. In order to compare Black defendants to White defendants (as well as defendants of other races to White defendants), we also ran a model that included three dummy variables for defendant race (measured as Black, White, or another race). Following conventional practice, we include two of these categories at a time, using the excluded category as a referent. The regression results in Table 7 below. Table 7. Impact of Case Characteristics and Defendant Race on Capital Sentencing Outcomes in Death Eligible Cases, December 1981 - May 2014 N= 77 Death Penalty Imposed Pseudo R2 = 0.2373 Variable Coefficient Exact P-Value Odds Ratio Referent (Compared to) Prior Convictions (ln) -0.095 0.498 0.909 1 Victim -0.720 0.223 0.487 Multiple victims Applied Aggravators 0.629 0.016 1.876** Mitigating Circumstances (ln) -0.263 0.086 0.769* Defenses -0.786 0.037 0.456** Victim Held Hostage 0.704 0.235 2.022 Not held hostage Black Defendant 1.557 0.045 4.743** White defendant Other Race Defendant -0.125 0.890 0.883 White defendant * significant at α = .10 ** significant at α = .05 *** significant at α = .01 Note: In this model, five cases (6.1%) were missing data and were therefore dropped from the analysis. Thus, when Black Defendants are compared to White Defendants (specified as the referent category), the log odds for Blacks are 1.56 (meaning that Black defendants are 4.7 times more likely that White defendants to receive a death sentence, p = 0.045). Other defendants do not statistically significantly differ from White defendants (p = 0.890). In Tables 8 and 9 below, we show the coefficients and associated p-values that are obtained under nine additional model specifications. These models include only those case characteristics that have been shown in the models presented above or during model testing be significant (i.e. the number of aggravating 102 Race, Death, and Justice Vol. 6:2 circumstances found by the jury; the (logged) number of mitigating circumstances, and the number of defenses offered); defendant race; and various victim and county characteristics. As previously noted, not all of these factors can be included simultaneously in the analysis of jury decision-making due to the (relatively small) number of cases. For this reason, each of the victim and county characteristics is tested separately (but in combination with significant case characteristics and defendant race). Table 8. Impact of Case Characteristics, Race of Defendant, and Victim Characteristics on Capital Sentencing Outcomes in Washington State Aggravated Murder Cases, December 1981 - May 2014 Death Penalty Imposed Model 1 Model 2 Model 3 Model 4 Model 5 Odds Ratio (P-Value) Odds Ratio (P-Value) Odds Ratio (P-Value) Odds Ratio (P-Value) Odds Ratio (P-Value) Aggravators 1.88*** (p = .007) 1.88*** (p = .008) 1.94 ** (p = .006) 1.88*** (p = .007) 1.93*** (p = .005) Mitigating Circumstances (ln) 0.85 (p = .228) 0.82 (p = .156) 0.84 (p = .204) 0.85 (p = .228) 0.88 (p = .365) Defenses 0.47** (p = .029) 0.48** (p = .033) 0.46** (p = .028) 0.47** (p = .036) 0.44** (p = .027) Black Defendant 4.79** (p = .030) 4.28 * (p = .049) 5.64** (p = .021) 4.79** (p = .030) 7.25** (p = .017) White Victim(s) 0.596 (p = .468) Female Victim(s) 0.49 (p = .194) Child Victims 1.07 (p = .919) Victim(s) Stranger 0.37 (p = .120) Pseudo R2 .2034 .2137 .2193 .2034 .2355 Probability > chi2 .0002 .0003 .0002 .0005 .0001 N 80 79 80 80 78 * significant at α = .10 ** significant at α = .05 *** significant at α = .01 Note: Figures shown are odds ratios; exact P-values are in parentheses. In this table, all significant results are bolded. The race of the defendant remains significant across all five of the models tested here. Specifically, the results indicate that Black defendants are between 4.3 and 7.3 times more likely to be sentenced to death than non-Black defendants controlling for the other variables included in the model. None of the victim characteristics tested in these models appear to be significant predictors of sentencing outcomes in capital cases in Washington State. Table 9 shows below the results that are obtained when a various county characteristics are included in the model. The results show that the race of the defendant remains significant across all five of the models tested here (with p-values ranging from .027 to .051). In these models, the results indicate that Black defendants are from 4.4 to 4.9 times more likely to be sentenced to death than non-Black defendants after controlling for the other factors included in the model. Only one of the county characteristics tested in these models appears to be a significant predictor of sentencing outcomes in capital cases in Washington State: the percent of the county population that voted Republican in the most recent Presidential election. Specifically, the results suggest that jurors in counties with more Republican voters are slightly less likely to impose death sentences after controlling for the other variables included in the model.116 Notably, the inclusion of this 116 In this case, the odds ratio is .94 which, because it is less than 1, suggests an inverse relationship. 2016 COLUMBIA JOURNAL OF RACE AND LAW 103 significant predictor in the model does not meaningfully reduce the significance and magnitude of the effect of the race of defendant in the sentencing phase of capital cases. Table 9. Impact of Case Characteristics, Race of Defendant, and County Characteristics on Capital Sentencing Outcomes in Washington State Aggravated Murder Cases, December 1981 - May 2014 Death Penalty Imposed Model 6 Model 7 Model 8 Model 9 Model 10 Odds Ratio (P-Value) Odds Ratio (P-Value) Odds Ratio (P-Value) Odds Ratio (P-Value) Odds Ratio (P-Value) Applied Aggravators 1.88 *** (p = .007) 1.94 *** (p = .007) 1.85 ** (p = .010) 1.89 *** (p = .007) 1.91 *** (p = .007) Mitigating Circumstances (ln) 0.85 (p = .228) 0.85 (p = .234) 0.82 (p = .151) 0.85 (p = .215) 0.86 (p = .273) Defenses 0.47 ** (p = .029) 0.46 ** (p = .027) 0.47 ** (p = .045) 0.47 ** (p = .029) 0.46 ** (p = .030) Black Defendant 4.79 ** (p = .030) 4.37 * (p = .051) 4.46 ** (p = .049) 4.71 ** (p = .033) 4.85 ** (p = .027) % Black in County at Year of Sentencing 1.05 (p = .659) % County Voted Republican 0.94 * (p = .062) Densely Populated at Year of Sentence 1.000 (p = .739) Per Capita Revenue in 1981 Real Dollars 0.45 (p = .378) Pseudo R2 .2034 .2051 .2369 .2044 .2111 Prob > chi2 .0002 .0004 .0001 .0005 .0003 N 80 80 80 80 80 * significant at α = .10 ** significant at α = .05 *** significant at α = .01 Note: Data shown are odds ratios; exact P-values are in parentheses. VI. DISCUSSION AND CONCLUSION This study is characterized by several limitations. First, our analyses rely largely upon trial reports, and trial reports only exist for cases that have been determined to involve aggravated, first-degree murder. As a result, we are unable to analyze whether and how extra-legal factors, perhaps including race, influence the process by which aggravating circumstances are demonstrated and affirmed. Second, the trial reports are incomplete. Although we were able to find some of the missing information regarding victim demographics in newspaper reports, this was not easily accomplished for other variables. As a result, we were unable to include some factors (such as a defendant’s IQ and mental health status) in our models that may also influence outcomes in capital cases. Third, some of the information that is treated in our analyses as a case characteristic may, in fact, be the result of extra-legal factors. For example, the number of aggravators affirmed by the court may in part reflect the quality of defendants’ legal representation. For this reason, our results likely underestimate the influence of extra-legal factors, including race, in capital cases. Finally, our analyses of sentencing decisions are limited by the small number of cases in these models, and hence by the small number of control variables that could be included in those models at once. We adopted a multi-faceted strategy for dealing with this challenge. First, we conducted diagnostics to identify any outliers that may have had undue leverage. Second, we engaged in sensitivity analysis, testing a variety of case characteristics to determine which of these influenced sentencing outcomes. The results indicate that although victims’ race does not significantly influence juries’ sentencing decisions, the race of defendant does. Specifically, the regression results consistently indicate that juries are 104 Race, Death, and Justice Vol. 6:2 more than four times as likely to impose a death sentence in cases involving Black defendants (after controlling for case characteristics). Indeed, the findings regarding the role of defendant race in the sentencing (but not filing) phase of aggravated murder cases in Washington State are robust (consistent) across numerous model specifications. Although the data set is small, the regression results are remarkably consistent, regularly indicating that Black defendants are more than four times more likely to receive a death sentence after controlling for relevant case characteristics across a variety of model specifications. The effect of the race of the defendant on sentencing outcomes remains consistent when Black defendants are compared to White defendants and when information about the race of the victim is included in the models. Similarly, inclusion of neither victim nor county characteristics in the regression models notably alters this effect. Together, the descriptive and statistical results support two main conclusions. The first pertains to intra-state variation in the propensity of prosecutors to seek, and juries to impose, death sentences. The descriptive statistics presented in Table 2 show there is significant variation in efforts to obtain death sentences and in decisions to impose them across counties in Washington State. Indeed, the proportion of cases in which prosecutors sought the death penalty in aggravated murder cases involving death-eligible adult defendants varies notably by county, from a high of sixty-seven percent in Thurston County to a low of zero percent in Okanogan County. The proportion of cases in which prosecutors sought death also varies markedly across larger counties with more aggravated murder cases, from a high of seventy-five percent in Clark County to a low of twenty-five percent in Kitsap County. Even in the context of a comparatively narrow statutory framework, it appears that prosecutorial discretion persists and is exercised in different ways across counties. The regression results provide additional support for this hypothesis. Neither the race of the victim nor the race of the defendant had a significant impact on prosecutorial decisions to seek death. These results are consistent with the descriptive data, which show that prosecutors filed death notices in a notably larger share of cases involving White versus Black defendants. However, our results indicate that two other extra- legal factors did significantly impact prosecutorial decisions regarding whether to file death notices: whether there was significant publicity about the case, and the size of the Black population in the county in which the case was adjudicated. Thus, it appears that county-level variation in the propensity to seek death is not strictly a function of legal factors or case characteristics. The fact that extensive publicity increases the likelihood that prosecutors will seek death is not too surprising given that county prosecutors are elected in Washington State, and publicity is likely to be seen as an indicator of public concern.117 The effect of the size of the county’s Black population on prosecutorial decisions to file death notices is more difficult to explain. Researchers have often included measures of the size of the Black or non-White population to assess the hypothesis that jurisdictions with larger minority populations are more punitive; evidence that racial composition matters is often interpreted in this manner.118 Although we cannot assess whether this interpretation is accurate, the findings clearly indicate that defendants in counties with larger Black populations are more likely to have a death notice filed against them. Moreover, this effect does not appear to stem from differences in population density, political orientation, or fiscal capacity, as measures of these factors were included in the regression analysis and thus controlled for in the model. In short, although neither the race of the victim nor the race of the defendant appear to affect prosecutorial decision-making, two other extra-legal factors (publicity and the size of the Black population in 117 In separate models, we assessed whether cases involving White victims or Black defendants were more likely to receive extensive publicity than other cases. The results did not indicate that this was the case when publicly known victim characteristics (number of victims, child victims, etc.) were taken into account. In fact, neither legal nor social characteristics available in these data provide robust predictors of whether a case receives extensive publicity; at most, just over five percent of the variation can be explained by these variables. 118 Act of March 22, 1909, ch. 249, § 140, 1909 Wash. Sess. Laws 890, 930 (repealed 1975). 2016 COLUMBIA JOURNAL OF RACE AND LAW 105 the county in which the case is adjudicated) do matter at this stage of the capital process and help to explain significant variation in death notice filing rates across Washington State counties. Second, with respect to jury decision-making, the findings indicate that several case characteristics were significant predictors of jury decisions to impose a sentence of death: the number of aggravating circumstances and the number of defenses. Surprisingly, the number of victims did not have a significant impact on jury (or prosecutorial) decision-making.119 Most dramatically, the findings indicate that juries were significantly more likely to impose a death sentence in cases involving Black defendants than they were in cases involving White or other defendants. Specifically, the regression results indicate that juries were more than four times more likely to impose a death sentence when the defendant was Black than in cases involving similarly situated non-Black defendants. When Black defendants were compared specifically to White defendants, the results were very similar. Nor did the inclusion of information about the race of the victim, other victim characteristics, or county characteristics notably reduce the race-of-defendant effect. Although the results presented here are based on analyses of a relatively small data set, they nonetheless consistently indicate that the race of the defendant has had a marked impact on sentencing in aggravated murder cases in Washington State since the adoption of the existing statutory framework. These findings are also consistent with the descriptive data, which show that juries imposed death sentences in forty percent of cases involving White defendants but 64 percent of cases involving Black defendants. This finding is somewhat unusual in that recent studies of the administration of capital punishment more commonly report race-of-victim effects than race-of defendant effects. On the other hand, our findings regarding the impact of defendant race are consistent with the results of some recent regression-based studies—especially those that analyze prosecutorial and jury-decision-making separately—that also find race-of-defendant effects.120 The finding that Washington State juries are more than four times more likely to impose death when the defendant is Black is also consistent with the results of numerous experimental studies, which indicate that implicit racial biases influence the propensity of mock jurors to vote for death, and that these biases are exacerbated by jury selection and deliberation processes.121 Most importantly, the findings presented here suggest that efforts to limit the role of race in capital sentencing by adopting a statutory framework that narrows the pool of homicide cases in which the death penalty may be imposed will not necessarily eliminate racialized decision-making in capital cases. Although it is possible that Washington’s comparatively narrow statutory framework helps to explain the absence of race- of-defendant or race-of-victim effects on prosecutorial decision-making in aggravated murder cases, some extra-legal factors, including the size of the counties’ Black population, do significantly impact prosecutorial decision-making in Washington State capital cases. Moreover, the narrowness of this statutory framework still leaves ample room for discretion and discrimination in jury decision-making. The regression results presented here consistently indicate that the race of the defendant is a significant predictor of a jury decision to sentence defendants to death rather than life. Indeed, across all of the models presented, the results indicate that juries were more than four times more likely to impose a death sentence when the defendant was Black (after controlling for case characteristics). It thus appears that statutory frameworks that effectively constrict the applicability of capital punishment do not do enough to reduce the role of race in capital 119 Although perhaps counter-intuitive, this finding is unsurprising given that some of Washington State’s most notorious serial and mass murderers—including Gary Ridgeway (who committed forty-eight murders), Ben Ng (who was convicted of killing thirteen people), and Kwan Fai Mak (who also committed thirteen murders)—were sentenced to life without parole rather than death. See WASHINGTON STATE BAR ASSOCIATION, FINAL REPORT OF THE DEATH PENALTY SUBCOMMITTEE OF THE COMMITTEE ON PUBLIC DEFENSE 4 (2006). 120 See, e.g., Baldus, Woodworth, Zuckerman, Weiner & Broffitt, supra note 51, at 1714; Baldus, Brain, Weiner & Woodworth, supra note 53; Donohue III, supra note 54; Songer & Unah, supra note 46. 121 Lynch & Haney, supra note 72; Lynch & Haney, supra note 79. 106 Race, Death, and Justice Vol. 6:2 sentencing. Despite our best efforts and intentions, our quest for color blindness appears to continue to elude us, even in matters of life and death. 2016 COLUMBIA JOURNAL OF RACE AND LAW 107 VII. APPENDIX A: AGGRAVATING FACTORS Under RCW 10.95.020, aggravating factors include the following: 1. The victim was a law enforcement officer, corrections officer, or a fire fighter who was performing his or her official duties at the time of the act resulting in death and the victim was known or reasonably should have been known by the person to be such at the time of the killing; 2. At the time of the act resulting in the death, the person was serving a term of imprisonment, had escaped, or was on authorized or unauthorized leave in or from a state facility or program for the incarceration or treatment of persons adjudicated guilty of crimes; 3. At the time of the act resulting in death, the person was in custody in a county or county-city jail as a consequence of having been adjudicated guilty of a felony; 4. The person committed the murder pursuant to an agreement that he or she would receive money or any other thing of value for committing the murder; 5. The person solicited another person to commit the murder and had paid or had agreed to pay money or any other thing of value for committing the murder; 6. The person committed the murder to obtain or maintain his or her membership or to advance his or her position in the hierarchy of an organization, association, or identifiable group; 7. The murder was committed during the course of or as a result of a shooting where the discharge of the firearm, as defined in RCW 9.41.010, is either from a motor vehicle or from the immediate area of a motor vehicle that was used to transport the shooter or the firearm, or both, to the scene of the discharge; 8. The victim was: (a) a judge; juror or former juror; prospective, current, or former witness in an adjudicative proceeding; prosecuting attorney; deputy prosecuting attorney; defense attorney; a member of the indeterminate sentence review board; or a probation or parole officer; and (b) the murder was related to the exercise of official duties performed or to be performed by the victim; 9. The person committed the murder to conceal the commission of a crime or to protect or conceal the identity of any person committing a crime, including, but specifically not limited to, any attempt to avoid prosecution as a persistent offender as defined in RCW 9.94A.030; 10. There was more than one victim and the murders were part of a common scheme or plan or the result of a single act of the person; 11. The murder was committed in the course of, in furtherance of, or in immediate flight from one of the following crimes: (a) robbery in the first or second degree; (b) rape in the first or second degree; (c) burglary in the first or second degree or residential burglary; (d) kidnapping in the first degree; or (e) arson in the first degree; 108 Race, Death, and Justice Vol. 6:2 12. The victim was regularly employed or self-employed as a news-reporter and the murder was committed to obstruct or hinder the investigative, research, or reporting activities of the victim; 13. At the time the person committed the murder, there existed a court order, issued in this or any other state, which prohibited the person from either contacting the victim, molesting the victim, or disturbing the peace of the victim, and the person had knowledge of the existence of that order; and 14. At the time the person committed the murder, the person and the victim were “family or household members” as that term is defined in RCW 10.99.020(1), and the person had previously engaged in a pattern or practice of three or more of the following crimes committed upon the victim within a five-year period, regardless of whether a conviction resulted: (a) harassment as defined in RCW 9A.46.020; or (b) any criminal assault. In addition, the following conditions must be met: 1. The jury affirmatively answers whether “having in mind the crime of which the defendant has been found guilty, are convinced beyond a reasonable doubt that there are not sufficient mitigating circumstances to merit leniency” at the conclusion of the special sentencing proceeding; and 2. The Washington Supreme Court conducts a proportionality review of a death sentence to determine: (a) whether there was sufficient evidence to justify the death sentence; (b) whether the defendant was mentally retarded; (c) whether it was brought on by passion or prejudice; and (d) whether the sentence was excessive or disproportionate. See RCW 10.95.60, RCW 10.95.70, and RCW 10.95.100. 2016 COLUMBIA JOURNAL OF RACE AND LAW 109 VIII. APPENDIX B: TRIAL REPORTS EXCLUDED FROM THE ANALYSES Cases described in trial reports numbers 1-332 (including 34A) were included in the analyses with the following exceptions: • One case was not included in the analysis because the defendant was not convicted of aggravated murder.122 • Once case was not included because the crime took place before the effective date of the current death penalty statute.123 • Three cases were not included in the analysis because defendants were ineligible for the death penalty due to extradition agreements.124 • Twenty-nine cases were not included because defendants were juveniles when they committed aggravated murder.125 • Three cases were not included in the jury decision-making analysis because although death notices were filed by prosecutors, legal rulings prevented a special sentencing proceeding.126 • Two cases were not included in the jury decision-making analysis because after a death notice was filed, the defendants entered a stipulated guilty plea and a special sentencing hearing therefore did not occur.127 122 TR No. 292. 123 TR No. 100. 124 TR Nos. 285, 286, 319. 125 TR Nos. 50, 61, 67, 70, 73, 110, 111, 122, 134, 139, 145, 149, 161, 170, 171, 189, 195, 196, 205, 206, 208, 209, 222, 223, 226, 246, 267, 270, 323. 126 TR Nos. 68, 217, 308. 127 TR Nos. 152, 153. 110 Race, Death, and Justice Vol. 6:2 IX. APPENDIX C: MEASUREMENT OF VARIABLES Appendix Table C. Variables and Measurement Variable Description Measurement Outcomes Death Notice Filed Death Notice Filed by Prosecutors Coded: 1=DN Filed; 0= ND Not Filed Death Penalty Imposed by Judge/Jury Sentenced entered as Death Coded: 1= Death; 0= Life Without Parole Predictors – Case Characteristics Number of Prior Convictions Total Number of Prior Convictions Number (logged) Number of Alleged Aggravators Total Number of Alleged Aggravators Number Number of Aggravators found by Jury Total Number of Applied Aggravators Number Mitigating Circumstances Total Number of Mitigating Circumstances Number (logged) Number of Defenses Offered Total Number of Defenses Number Plea Plea entered Coded: 1=Plead Guilty; 0= Plead Not Guilty Number of Victims Total Number of Victims 3 Coding Categories: 1 Victim; 2-4 Victims; 5 or more Victims; coded as 0/1 Victim Held Hostage If Victim was held hostage Coded: 1=Yes; 0= No Sex Crime If sex crime also suspected to have occurred Coded: 1=Yes; 0= No Victim Law Enforcement Officer Any victim a law enforcement officer Coded: 1=Yes; 0= No Predictors – Defendant Characteristics Defendant Race Defendant’s Race 3 Coding Categories: White; Black; Other Race Each coded as 0/1 Predictors – Victim Characteristics Victim Race Victims’ Race 4 Coding Categories: All Victims White; All Victims Black; All Victims Other Race; Victims of Multiple Races. Each coded as 0/1 Victim Sex Victims’ Sex 3 Coding Categories: All Victims Female; All Victims Male; Victims Mixed Sexes. Each coded as 0/1 Victim Stranger If defendant knew victim Coded: 1=Yes; 0=No Victim Child^ Any victim under age 18 Coded: 1=Yes; 0=No Predictors – County Characteristic Publicity Extensive publicity about the trial Coded: 1=Yes; 0=No Densely Populated County* Population Density > 150 people per square mile at year of sentencing Coded: 1=Yes; 0=No 2016 COLUMBIA JOURNAL OF RACE AND LAW 111 Percent Black in County at time of Arrest* Share of county population that is Black at arrest Proportion Percent Black in County at time of Sentencing* Share of county population that is Black at sentencing Proportion Percent White in County at time of Arrest* Share of county population that is White at arrest Proportion % White in County at time of Sentencing* Share of county population that is White at sentencing Proportion % Latino in County at time of Arrest* Share of county population that is Latino at arrest Proportion % Latino in County at time of Sentencing* Share of county population that is Latino at sentencing Proportion % Vote Republican* Percent of county population that voted Republican in most proximate Presidential election Proportion Per Capita Revenue at Year of Sentencing* Per capita revenue of county in Real Dollars (1981) Number (logged) Note: All indicators were taken from trial reports unless marked with a symbol. *These data were taken from the United States Census Bureau. ^These data were taken from both trial reports and newspaper reports of case. +Mitigating circumstances were coded by attorneys Lila Silverstein and Neil Fox. 112 Race, Death, and Justice Vol. 6:2 X. APPENDIX D: MODELING PROSECUTORIAL FILING DECISIONS Appendix Table D1. Descriptive Statistics: Prosecutorial Filing Decisions in Capital Sentencing, December 1981 – May 2014 N Min. Max Mean/ Proportion Standard Deviation Missing Percent Missing Death Notice Filed 298 0 1 .29 .455 0 0% Case characteristics Number of Priors 285 0 68 4.07 6.579 13 4.4% One Victim 298 0 1 .65 .479 0 0% Two to Four Victims 298 0 1 .33 .471 0 0% Five or More Victims 298 0 1 .02 .152 0 0% Alleged Aggravators 298 1 17 2.19 1.661 0 0% Sex Crime 298 0 1 .19 .394 0 0% Victim Police Officer 298 0 1 .05 .219 0 0% Defendant Characteristics Black Defendant 297 0 1 .19 .394 1 .004% White Defendant 297 0 1 .64 .481 1 .004% Other Race Defendant 297 0 1 .17 .375 1 .004% Victim Characteristics White Victim(s) 291 0 1 .75 .436 7 2.3% Female Victim(s) 298 0 1 .42 .494 0 0% Child Victim(s) 298 0 1 .17 .374 0 0% Victim(s) Stranger 294 0 1 .32 .467 4 1.3% County Characteristics % Black in County at Year of Arrest 296 .1 7.5 3.29 2.50 2 .01% % County Voted Republican 298 27.4 69.3 44.88 9.0 0 0% Densely Populated at Year of Arrest 298 0 1 .76 .427 0 0% Per Capita Revenue in 1981 Real Dollars 296 139.79 1395.33 328.63 165.89 2 .01% Publicity Was a Factor 286 0 1 .74 .439 12 4% 2016 COLUMBIA JOURNAL OF RACE AND LAW 113 Appendix Table D2. Correlation Matrix: Alleged Aggravators and Crime Characteristics Death Penalty Sought 1 Victim 2-4 Victims 5 or More Victims Alleged Aggravators Sex Crime Law Enforcement Officer Death Penalty Sought 1.0000 1 Victim -.0671 (.2479) 1.0000 2-4 Victims .0690 (.2353) -.9490 (.0000) 1.0000 5 or More Victims -.0021 (.9708) -.2103 (.0003) -.1086 (.0612) 1.0000 Alleged Aggravators .2210 (.0001) -.1325 (.0222) .1490 (.0100) -.0443 (.4461) 1.0000 Sex Crime .1756 (.0023) .1444 (.0126) -.1225 (.0345) -.0754 (.1941) .2690 (.0000) 1.0000 Law Enforcement Officer .1222 (.0349 ) .0413 (.4776) -.0305 (.6003) -.0357 (.5392) -.0631 (.2775) -.1120 (.0535) 1.0000 114 Race, Death, and Justice Vol. 6:2 XI. APPENDIX E: MODELING SENTENCING OUTCOMES IN CAPITAL SENTENCING Appendix Table E1. Descriptive Statistics: Capital Sentencing Outcomes in Death Eligible Cases, December 1981 – May 2014 N Min. Max Mean/ Proportion Standard Deviation Missing Percent Missing Death Notice Filed and not Withdrawn by Prosecutors or Judge 82 0 1 .43 .498 0 0% Case characteristics Number of Priors 80 0 23 4.24 4.192 2 2.4% One Victim 82 0 1 .61 .491 0 0% Two to Four Victims 82 0 1 .37 .485 0 0% Five or More Victims 82 0 1 .02 .155 0 0% Plead Guilty 82 0 1 .20 .401 0 0% Aggravators Found by Jury 82 0 12 2.27 1.905 0 0% Total Mitigating Circumstances 82 0 11 2.40 2..119 0 0% Number of Defenses 82 0 2 .21 .515 0 0% Defendant Characteristics Black Defendant 81 0 1 .17 .380 1 1.2% White Defendant 81 0 1 .70 .459 1 1.2% Other Race Defendant 81 0 1 .12 .331 1 1.2% Victim Characteristics Any Victim Held Hostage 81 0 1 .33 .474 1 1.2% White Victim(s) 80 0 1 .83 .382 2 2.4% Female Victim(s) 82 0 1 .51 .503 0 0% Child Victim(s) 82 0 1 .23 .425 0 0% Victim(s) Stranger 80 0 1 .31 .466 2 2.4% County Characteristics % Black in County at Year of Sentencing 82 .2 9.3 3.83 2.679 0 0% % County Voted Republican 82 27.4 62.4 45.17 8.476 0 0% Densely Populated at Year of Sentence 82 00 1 .85 .356 0 0% Per Capita Revenue in 1981 Real Dollars 82 174.93 769.03 306.78 135.79 0 0%