One February 5, 2025, Fair Trial Analysis LLC filed an Amicus Curiae Brief in Support of Warren King’s Petition for Writ of Habeas Corpus in the Superior Court of Butts County, Georgia (Case No. 2024-SU-HC-0010). The Brief shares the results of a double-blind randomized experiment that had 786 jury-qualified adults, representative of the county, vote for a life/death sentence based on the case facts with or without the dispute evidence.
Background on King’s Case
In 1994, Warren King, an 18-year-old Black man with no history of violent crime, and his older cousin Walter Smith were arrested for the murder of Karen Crosby, a white woman, during an attempted convenience store robbery in Surrency, Georgia. At trial, the prosecution’s key evidence was Smith’s testimony that King had initiated the crime and pulled the trigger. King was convicted of murder and sentenced to death in 1998.
In post-conviction proceedings, King argues that the prosecution violated his constitutional rights by failing to disclose that Smith had been promised a plea deal in exchange for his testimony. At trial, the prosecutor and Smith both denied any agreement. However, King later uncovered evidence that the prosecutor had, in fact, arranged for Smith to plead guilty in exchange for a life sentence rather than facing the death penalty. King contends that this suppressed impeachment evidence could have severely undermined Smith’s credibility and therefore prejudiced his trial.
What King Must Prove in Post-Conviction Proceedings
To obtain post-conviction relief for the Brady violation (suppression of impeachment evidence), King must show that there is a reasonable probability that, had the jury known about Smith’s plea deal, the outcome of his trial would have been different. This standard stems from United States v. Bagley and Strickland v. Washington, which require that suppressed evidence be material — meaning it must undermine confidence in the verdict. If the suppressed evidence would not have reasonably changed the jury’s decision, relief for the Brady violation must be denied.
Under Napue v. Illinois, the standard for showing harm is lower when the State knowingly uses or fails to correct false testimony. Instead of having to show a reasonable probability of a different outcome, the defendant only needs to show a reasonable likelihood that the false testimony could have affected the jury’s judgment. This means that under Napue, King could win relief even if the effect of the suppressed evidence was smaller than what Bagley or Strickland would otherwise require.
Analyzing the Effect of the Impeachment Evidence on the Trial Outcome
Fair Trial Analysis tested King’s claim by conducting a double-blind, randomized survey experiment with 786 jury-eligible adults from a population representative of Butts County, Georgia, the trial venue. Participants were randomly assigned to either (1) a summary of King’s trial without mention of Smith’s plea deal (mirroring the original trial) or (2) a summary disclosing the plea deal. The study measured how the disclosure affected sentencing preferences and found that about 31.5% of people who initially favored a death sentence changed to favoring life imprisonment when told about the plea deal. The research provides empirical evidence that the suppressed evidence significantly affected juror decision-making.
This is the first brief of its kind. We thought it was important to explain the methodology and make the analysis as transparent as possible. We included a link to the survey instrument in the brief, so others can see it for themselves, and posted replication materials in a public data repository.
We pray that this research will help Judge Robert Mack evaluate the effect of the disputed evidence on a jury trial. The prevailing standards for a case like this, while clear in the abstract, are notoriously difficult to apply as they require the judge to be something of a mind reader. The best way to understand how a trial error or omission affected a representative cross-section of the community is to study a representative cross-section of the community using controlled randomized experiments, the gold standard of research.
Georgia v. Warren King study files: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/O2SKO4