Skip to main content

Flowers Case


Death-Qualified Juries and the Flowers Trials

Compiled by Alice Chao, Ali Franz, Josh Howard, Molly Huffaker, Christina Lee, and Federico Wynter

Overview: what is a death-qualified jury?

A “death-qualified jury” is a jury pronounced fit to decide a capital case. Death-qualified juries are composed of jurors who are not strictly opposed to capital punishment but who also do not believe that the death penalty should be imposed in all cases of capital murder. While the death qualification process is designed to produce a fair and impartial jury, social science research suggests that it instead produces a jury that is biased against a capital defendant.

Constitutional Cases

The constitutionality of death-qualified juries was debated by the Supreme Court in two major cases.

Witherspoon v. Illinois, 391 U.S. 510 (1968)

Witherspoon was convicted of murder and sentenced to death by a jury in Illinois in 1960. At this time, there was an Illinois statute that allowed for the dismissal of any juror with “conscientious scruples” against capital punishment. The State of Illinois accordingly allowed the prosecution to eliminate any prospective jurors who were either opposed to capital punishment or doubtful about inflicting it. However, when doing so, the prosecution did not find out if most of the jurors who they eliminated would have actually voted against the death penalty in practice.
Consequently, Witherspoon appealed and alleged that the jurors’ dismissal violated both his Sixth and Fourteenth Amendment rights. He also argued that the jury chosen was more likely to be biased in favor of conviction than a jury chosen at random without regard to the jurors’ opinions about capital punishment. Witherspoon maintained that the type of juror who would be able to inflict the death penalty would also be more likely to ignore the defendant’s presumption of innocence, accept the prosecution’s story, and return a guilty verdict. Witherspoon produced social science evidence, detailed below, to support these claims.
Ultimately, the Court rejected Witherspoon’s data, but did hold that his death sentence was unconstitutional. Had the prosecution only eliminated jurors who have never considered the death penalty, the jury would have remained constitutional. However, by eliminating all jurors who were simply opposed to the death penalty in principle, “the State crossed the line of neutrality.” As the Court said in its opinion, “In its quest for a jury capable of imposing the death penalty, the State produced a jury uncommonly willing to condemn a man to die… The State of Illinois has stacked the deck against the petitioner.”
The term “Witherspoon-excludable,” a person who is opposed to capital punishment and therefore may be excluded as a juror, comes from this case. See Grigsby v. Mabry, 569 F.Supp. 1273, 1289 (E.D. Ark. 1983).
Lockhart v. McCree, 476 U.S. 162 (1986)
Eighteen years later, this question was revisited by the Supreme Court in Lockhart v. McCree. McCree was convicted of capital murder, and during his trial prospective jurors were excluded for expressing doubt about the death penalty. The Eighth Circuit Court of Appeals held that this removal violated McCree’s Sixth and Fourteenth Amendment rights, particularly because this juror exclusion deprived him of a jury selected from a representative cross-section of the community.
The U.S. Supreme Court, however, disagreed, stating that grouping jurors by their willingness to inflict the death penalty did not make them a group for fair cross-section purposes. This was an attribute that jurors could control, unlike race or sex; furthermore, not every juror who opposed the death penalty would be removed in all capital cases. According to the Court, the only requirement to protect McCree’s constitutional right to a fair trial was to ensure that the jury was selected from what the Court considered to be a fair cross-section of the community, was impartial and properly applied the law to the facts of the case.
McCree also introduced six studies, detailed below, arguing that because death qualification increases the likelihood of conviction, it undermines the impartiality of the jury. The Court gave these studies little weight.

Social Science in Witherspoon and Lockhart

In Witherspoon v. Illinois and Lockhart v. McCree, the Supreme Court rejected social science research with regards to the constitutionality of the use of death-qualified juries.  First, in Witherspoon, the Court rejected two studies investigating the effect of whether a jury member conscientiously opposed capital punishment on the likelihood that the jury member would find a defendant guilty.  A 1964 study by W.C. Wilson presented 187 college student participants with evidence in six capital cases and asked participants to return a verdict (Wilson, 1964).  Experimenters also asked participants whether they had “conscientious scruples” against the death penalty.  The study found that students who stated they had conscientious scruples against the death penalty gave guilty verdicts significantly less often than did students who said they had no such scruples. 

Similarly, in a study by F.J. Goldberg, experimenters presented 200 college student participants with evidence in simulated trials, asked them to determine guilt and penalty, and asked whether they had “conscientious scruples” against the death penalty (Goldberg, undated).  This study also found that students with scruples against the death penalty were more likely to find the defendant not guilty; however, unlike in the Wilson study, this difference was not statistically significant.

The Court in Witherspoon rejected these studies as too “tentative and fragmentary” to support an argument that a defendant’s chance of acquittal is significantly reduced when he faces a death-qualified jury and a general constitutional rule against death-qualified juries.  In light of the weaknesses of these studies, the Court’s rejection is perhaps justifiable. First, the participants of both studies were college students rather than actual jurors, giving the study low external validity and making it almost impossible to predict whether the results from the study would replicate in a real-life situation with an actual jury.  In addition, both studies were unpublished at the time that Witherspoon cited them in his brief.  Lastly, the fact that the difference between students who opposed the death penalty and students who did not was not statistically significant in Goldberg’s study casts doubt on Witherspoon’s argument that a death-qualified jury is significantly more likely to find a defendant guilty.

In Grigsby v. Mabry, the opinion of the court below in Lockhart, the court cited studies of death-qualified juries by Dr. Phoebe Ellsworth that were designed to address the same issues as the studies in Witherspoon using more sophisticated research techniques (Fitzgerald & Ellsworth, 1984).  The subjects of the studies were drawn from a panel of 288 jury-eligible citizens, 45% of whom had prior jury service.  Subjects whose attitudes toward the death penalty prevented them from being impartial in the guilt/innocence phase were excluded.  Subjects viewed a videotape of a simulated trial and were asked to return a verdict.  Death-qualified jurors were significantly more likely to find a defendant guilty than jurors who were not death-qualified.  However, the study had one important weakness: the experiment did not employ random sampling, which makes generalizing the study to a larger population impossible.

In addition, the court in Grigsby pointed to a study by Dr. Reid Hastie (although the court did not appear to indicate which study) indicating the important impact that the initial attitudes of each juror (in this case, jurors’ initial support for or opposition to the death penalty) have on the jury’s final verdict.  Hastie’s research demonstrates that the verdict initially selected by the largest portion of the jury, prior to any jury deliberation, will most likely become the final verdict.  Jurors’ initial attitudes are critical in explaining this difference: a juror’s initial view of a case plays a strong role in the juror’s initial determination of the verdict.  The court in Grigsby found these studies compelling, opining that the studies simply proved what courts had already known about death-qualified juries and setting aside McCree’s conviction.

On appeal to the Supreme Court in Lockhart, McCree offered several studies in addition to Ellsworth’s work.  One of these studies, published by Haney in Law and Human Behavior, investigated whether the experience of obtaining a death-qualified jury during voir dire creates expectations or preconceptions in prospective jurors, predisposing them to interpret and receive evidence in certain ways and influencing their eventual verdict and sentencing decisions.  The subjects of this study were 67 adult men and women who responded to a newspaper advertisement asking for subjects to participate in a study of juror decision making.  People who indicated that they could not impose the death penalty in any case or could not be fair and impartial in deciding guilt if they knew that the death penalty was a possibility were excluded from the study.  Subjects were shown a videotape of a simulated voir dire and answered a questionnaire designed to assess their attitudes about the trial and its participants.  Half of the subjects watched a videotape that included a death qualification segment. 

The study found that exposure to death qualification during voir dire increased subjects' belief in the guilt of the defendant and their estimate that he would be convicted. It also increased their estimate of the prosecutor, defense attorney, and judge's belief in the guilt of the defendant. The death-qualification process led subjects to perceive both prosecutor and judge as more strongly in favor of the death penalty, and to believe that the law disapproves of people who oppose the death penalty. In addition, jurors exposed to death qualification chose the death penalty as an appropriate punishment much more frequently than control jurors. Thus, persons who had been exposed to death qualification not only differed from non-death-qualified subjects, but they differed in ways that were consistently prejudicial to the defendant.  However, this study is limited by its incredibly small sample size (67) and the likely self-selection bias that comes of obtaining subjects through newspaper advertising.

Despite the non-random sampling weakness of the Ellsworth study and the small sample size of the Haney experiment, these studies represent an improvement on the studies cited in Witherspoon because they used a population of adults, many of whom had prior jury experience, rather than college students.  However, the court in Lockhart seemed to take these limitations into consideration in ruling, as did the Court in Witherspoon, that the studies introduced by McCree were still too “tentative and fragmentary” to support McCree’s claim and were not useful in predicting the behavior of actual jurors.  In addition, the Court asserted that the studies could not serve as the basis for a constitutional rule against death-qualified juries because only one of McCree’s studies accounted for the presence of individuals who could not impartially decide a capital defendant’s guilt or innocence. 

Post-Lockhart Social Science

Ironically, the Supreme Court’s treatment of the complied data in Lockhart left the door ajar for the application of social science research in capital punishment litigation (Rozelle, 2006, 783-784). The Lockhart majority identified three key flaws that barred the inclusion of the relevant and new data before the court (Rozelle, 2006). The studies were criticized for their (1) lack of analysis of actual juror behavior; (2) lack of engagement of the process of jury deliberation; and (3) their failure to account for the presence of nullifies (Rozelle, 2006). These concerns both provoked and informed the research community’s response embodied in the Capital Jury Project (CJP), a vast empirical study that directly addressed the flaws focused on by the court (Rozelle, 2006).
The CJP is a multi-state, cross-sectional empirical survey projected conducted by a legal researchers and social scientists (Brewer, 2004, 534). The participating states were selected to be representative of the broad variations in capital discretion statutes from across the country (Brewer, 2004). Four jurors were randomly selected rom each state’s sample of capital punishment cases from 1988 onwards (Brewer, 2004). Each respondent then took part in a three to four hour in-person interview that contained a mixture of closed and open questions aimed at covering the extent of their particular decision-making process as well as accounting for demographic data (Brewer, 2004).
The 50-page instrument used in every interview aimed to document the process by which each particular juror came to his or her particular sentencing decision (Brewer, 2004). The study seeks to identify any outside influences that affected each interviewed juror and to clarify at which point in the sentencing trial they made determinations on sentencing and to what degree of certainty (Brewer, 2004).
Accordingly, one pathway to assessing the likelihood of a death-qualified jury to impose the death penalty is offered through measuring the point in the trial at which respondents decided that sentence was appropriate (Bowers, 1995, 1087-1089). Respondents were asked what sentence they thought the defendant should be given after they were found guilty but before the penalty stage of the trial and its requisite consideration of aggravating and mitigating circumstances (Bowers, 1995). Nearly half of the respondents noted that they had decided what the punishment would be once the defendant had been found guilty whether that meant life without parole or the death penalty. 28.6% decided on the death penalty (Bowers, 1995). This finding is problematic in light of the post-Furman stipulation that the sentencing-phase of the trial should produce a decision based upon a consideration of presented evidence.
To control for the impact of that evidence respondents were asked to measure how strong their believe about appropriate sentence was from “absolutely convinced” to “not too sure.” 7 out of ten jurors who thought a death penalty was appropriate were “absolutely convinced” follow a determination of guilt (Bowers, 1995, 1090). Nearly all were at least “pretty sure.” Across the full course of the sentencing trial the data showed that almost four out of five jurors who decided on the death penalty stuck with that decision up to voting on the sentencing.
Additionally, the data provides significant results as to the impact of race upon the imposition of the death penalty. Where a black defendant was charged with murdering a white victim(s), all white juries imposed a death sentence in 71.9% of cases (Bowers, Steiner & Sandys, 2001, 193). Alternatively, when one black male was on the jury the death penalty was imposed in 42.9% of cases (Bowers, Steiner & Sandys, 2001, 193).  The makers of the study concluded that the significant variables were white male jurors and black jurors (Bowers, Steiner & Sandys, 2001). The death penalty became significantly more likely where it included five or more white males whereas the inclusion of a black male juror substantially reduced the likelihood of the death sentence being imposed particularly in cases involving a black defendant and a white victim (Bowers, Steiner & Sandys, 2001).
The study also produced findings on the race of individual jurors and the extent to which it impacted decision-making. Where black and white jurors from the same case were found, respondents decision as to which punishment was appropriate were measured at three distinct points in the sentencing process (Bowers, Steiner & Sandys, 2001). The study found that whites were three times more likely than black jurors to decide on the death penalty at the pre-sentencing guilt phase of the trial; after sentencing they were four times more likely and by the vote on punishment they were seven ties more likely (Bowers, Steiner & Sandys, 2001, 200).
The study extensively demonstrates systematic issues with the sentencing process carried out by death-qualified juries in relation to established constitutional doctrine. The findings clearly indicate the propensity of these juries to contain participants who have determined the appropriateness of the death penalty pre-sentencing. These findings are further exacerbated by the extent to which this propensity intersects with the racial composition of the jury and the race of the individual juror.
The limitations of the CJP do not diminish its usefulness in the context of capital punishment litigation. The study is particularly susceptible to hindsight bias because of its reliance on the memory of its respondents (Hans, 1995, 1235-1236). This should not detract from the value that the study adds to capital punishment litigation. Most importantly,the study serves as a rigorous empirical validation of the conclusions advocated by the data formerly rejected by the Supreme Court.

The Influence of Pretrial Publicity

Death-qualified jurors tend to develop biases against the defendant due to pretrial publicity. Pretrial publicity is information about the defendant and case published through medias prior to the beginning of the trial. Such information in the media tends to be sensationalized, creating a dramatized—and often inaccurate—version of the story. This is particularly true of death penalty cases, which involve violent and disturbing crimes. (Butler, 2007). The issue of pretrial publicity is further exacerbated by modern news reporting, where information is disseminated online immediately, even before official police reports come in.
A prominent social studies study regarding the correlation between death-qualified jurors and pretrial publicity is by Brooke Butler. (Butler, 2007). In her study, Butler used a sample of two-hundred residents belonging in Florida’s 12th Judicial Circuit in order to gauge residents’ support of the death-penalty in relation to their exposure to pretrial publicity. She utilized a then current trial, Florida v. Joseph Peter Smith, where the defendant was charged with kidnapping, sexual assault, sexual battery, and first-degree murder. The case was well-known due to the victim’s abduction being caught on tape.It was also a case where pretrial publicity was either inaccurate or misleading. The media reports also described the defendant in degrading ways.
Butler’s study results found that death-qualified residents were more likely to identify the defendant correctly. They were more familiar with the facts of the case, more likely to think the defendant was guilty, and more likely to recommend the death-sentence. They also thought that pretrial publicity would have a minimal impact on the defendant’s right to due process.
In her article summarizing the study’s results, Butler acknowledges the deficiencies in her study; namely, that her sample’s demographic skewered towards conservative, wealthy white men.The study also took place in Florida’s 12th Judicial Circuit. Butler admits that the results of her study may not be representative of other states. However, in all six of Flowers’ trials, the jury makeup similarly consisted of a white men majority. As mentioned, this deficiency is the current issue in his upcoming Supreme Court hearing, and studied mentioned in this article have shown that death-qualified jurors often skewer towards white men regardless of state or location.
Another study by Christine Ruva, Cathy McEvoy, and Judith Bicker Bryant examines the effects of pre-trial publicity and jury deliberation in the context of a murder trial. (Ruva, Mcevoy & Bryant, 2007). Although their study did not specifically examine death-qualified juries, it produced similar results to Butler’s study. Here, the jurors exposed to pretrial publicity were also more likely to vote guilty. They gave harsher sentences to defendants they found guilty and were more likely to know the basic facts of the case before the trial began. Furthermore, they were more likely to be biased against the defendant and find them less credible.
Defendants have often challenged their verdicts on the issue of jury bias and pretrial media prejudice. But due to the availability of information, the courts have ignored or denied remedies on grounds that juries need not be entirely ignorant of the case. In Mu’min v. Virginia, the Supreme Court struck down the petitioner David Mu’min’s argument that the jury had been biased towards his guilt due to widespread negative pretrial reporting on his personal history and criminal background. Mu'min v. Virginia, (1991). By affirming Mu’min’s sentence, the Supreme Court affirmed the precedent where a possible juror’s exposure to pretrial publicity, no matter how extensive, does not disqualify them during voir dire.
Scholars have espoused change-of-venue motions in order to mitigate the negative effects of pretrial publicity bias. The issue of pretrial jury bias was the basis for the change in venue for Flowers’ second trial. The venue location was changed from Montgomery County, where Winona is located, to Harrison County. However, the issue of jury information bias and venue change is tenuous in court. In Murphy v. Florida and Patton v. Yount, the Supreme Court decided the standard for granting a venue change to be a “substantial likelihood” that juror prejudice will prevent the defendant from receiving a fair trial. Murphy v. Florida, (1975). Patton v. Yount, (1984). (Mariniello, 1994). This is a high standard to hold, particularly for death penalty cases, which almost always receive wide publicity that reach local and non-local jury candidates.
The Tardy murders received widespread attention before Flowers was even arrested. Neither Winona nor its neighbors had seen a crime of such scale in their state. After the murders, there was a period of months where authorities did not identify nor arrest a suspect. The unprecedented nature of the murders in conjunction with a silent period meant that the murders attracted statewide attention. The delay caused frustration to mount among both Winona residents and the families of the victims. The district attorney was under pressure to name a suspect. By the time Curtis Flowers was arrested, the Tardy murders had become a political and emotional issue. Yet Flowers’ sixth trial, despite the opinion of his guilt or innocence being starkly divided along racial lines, was again held in Winona.

The Flowers Trials

Flowers I
In the first trial, the jury selection process yielded 97 qualified individuals: 77% were white while 23% were black. Brief for Petitioner at 4, Flowers v. Miss., appeal docketed, (2018) (No. 17-9572). Though the methods of selecting these jurors was not purely based on death qualifications, the jury selection process shows a racially skewed pool for the eventual jury. This supports the idea that death qualified juries tend to be predominantly white. 

Jury selection in the second and third trials mirrored the jury selection process of the first trial. Brief of Appellant at 109, Flowers v. State,240 So. 3d 1082 (Miss. 2017)(No. 2010-DP-01348-SCT).
Flowers IV and V
The prosecution did not seek the death penalty in the fourth and fifth trials. Thus, the issue of death-qualification did not come up during voir dire. These were the only two trials in which the jury hung and also the only juries on which there was more than one black juror. In Flowers IV, the juror qualification process resulted in a potential jurors pool that was 44% black and 56% white, and the jury was formed by seven whites and five blacks. Brief for Petitioner at 12, Flowers v. Miss., appeal docketed, (2018) (No. 17-9572). In Flowers V, the jury was formed by nine whites and three blacks. Id. at 13. The greater presence of blacks in the two juries where there was no death qualification process supports the idea that death-qualification yields juries that are less racially diverse. Further, these two hung juries support the social science findings that death qualified juries are less engaged in deliberation, more prone to convict and more likely to believe the prosecution’s arguments.
Flowers VI
In Flowers VI, the prosecution once again sought the death penalty. During voir dire, 10 of the 156 potential jurors, 8 black and 2 white, were struck for saying they can't impose the death penalty. (Craft, 2018). This may correlate with research on racial bias in the death qualification of a jury because if fewer white people are unopposed to the death penalty, more are eligible to serve as jurors. At the peremptory challenges phase, the prosecution struck down potential juror Tashia Cunningham alleging: “On her questionnaire, she put she would not consider death or life. She was back and forth in questioning on what her opinion was on the death penalty.” Id. Flowers challenged Cunningham’s strike, alleging that her views on the death penalty were similar to the views of white jurors. However, the trial court held that the State’s reasons were race neutral, despite the fact that in Flowers III the court had held that striking a black juror who had “virtually indistinguishable” views on the death penalty as white jurors who were not struck was sufficient to raise an inference of race discrimination but did not amount to a Batson violation. J.A. at 398, Flowers v. Miss., appeal docketed, (2018) (No. 17-9572). Similarly, the prosecution struck Flancie Young Jones, saying, among other reasons: “She was back and forth all over the place on her opinion about the death penalty... her only explanation for that change was that she said she had lied on her questionnaire.” (Craft, 2018). Both Cunningham and Jones are black women. These methods of striking black jurors demonstrate that the death qualification may be used to racially define the jury and death qualification may serve as a justification, disguising more racially-motivated reasons for removing a juror.  


The death sentences in Flowers I, II, III and VI, where the juries were overwhelmingly white after the death qualification process, are in line with the social science evidence that shows that death qualified juries are biased against the defendant, more likely to believe the prosecution’s story and convict, and more likely to be racially imbalanced.