On April 25, 2025, Fair Trial Analysis filed an Amicus Curiae Brief in Partial Support of Brenda Andrew’s Petition for Writ of Habeas Corpus in the U.S. Court of Appeals for the Tenth Circuit (Case No. 15-6190). The Brief shares the results of a double-blind randomized experiment that had 1,128 jury-qualified adults, representative of Oklahoma County, Oklahoma, the trial venue, vote on trial outcomes with or without the dispute evidence.
Background on Brenda Andrew’s Case
Brenda Andrew was convicted and sentenced to death for the murder of her estranged husband, Rob Andrew, in Oklahoma. Evidence showed that Andrew conspired with her lover, James Pavatt, to lure Rob Andrew to her garage under false pretenses and fatally shoot him in 2001. The State presented circumstantial evidence of premeditation, including evidence of life insurance fraud and attempted earlier attacks. During trial, however, the prosecution also introduced extensive, highly prejudicial evidence concerning Andrew’s appearance, sexual conduct, and failings as a mother.
On direct appeal, the Oklahoma Court of Criminal Appeals (OCCA) affirmed her conviction and sentence but acknowledged that some of the State’s evidence — especially concerning Andrew’s appearance, sex life, and mothering — was irrelevant. Nevertheless, the OCCA found the errors harmless and denied relief.
In post-conviction proceedings, Andrew argues that the admission of irrelevant and inflammatory evidence deprived her of a fundamentally fair trial. The U.S. District Court denied relief, and a divided panel of the Tenth Circuit Court of Appeals affirmed.
The U.S. Supreme Court vacated the judgment of the Tenth Circuit and held that the lower court erred in refusing to recognize clearly established federal law governing Andrew’s due process claim. Because the Tenth Circuit incorrectly concluded that no such clearly established law existed, it failed to evaluate whether the improper admission of prejudicial evidence at Andrew’s trial violated due process.
Did the Prejudicial Evidence Make Andrew’s Trial Fundamentally Unfair?
On remand, the Supreme Court instructed the Tenth Circuit to decide whether a fairminded jurist could disagree with Andrew’s argument that the improper evidence — including testimony about her appearance, sex life, and personal character — so infected either the guilt or sentencing phases of her trial as to render them fundamentally unfair. The Court emphasized that the Tenth Circuit must address the guilt and sentencing phases separately, considering the relevance of the disputed evidence, the degree of prejudice it caused, and any mitigating instructions given at trial.
To prevail on her claim and obtain post-conviction relief, Andrew must demonstrate that the prejudicial evidence undermines confidence in the reliability of the verdict or death sentence
Analyzing the Effect of Improper Evidence
Fair Trial Analysis conducted a double-blind, randomized experiment with 1,128 jury-qualified adults representative of Oklahoma County. Participants were randomly assigned to receive either a trial summary with the improper sexual and character evidence included or an error-free summary omitting it. The study found that the improper evidence had little effect on the guilt verdict. At the same time, the improper evidence substantially increased the probability of a death sentence recommendation — doubling it from 11.2% to 22.4%. This empirical finding provides objective support for Andrew’s claim that the improper evidence made the sentencing phase of her trial fundamentally unfair.
This is one the first briefs of its kind. We explained the methodology as clearly as we could and made the analysis as transparent as possible. We posted replication materials in a public data repository. Our research offers scientific, neutral evidence about the likely effect of improper trial evidence, enabling fairminded judges to reconsider initial intuitions about harmlessness based on systematic data rather than conjecture.
We hope that this research will help the appellate panel evaluate the effect of the disputed evidence in the guilt and sentencing phases of 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.
Oklahoma v. Andrew study files: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/VFEZGY