Cornell Study Finds Gendered Language in Death Penalty Trials

Researchers analyzed 137 capital murder trials to uncover how specific biases appear in legal court records.

Updated on Oct. 29, 2026 in Women’s Issues

Isometric editorial illustration of stacked geometric blocks, representing the linguistic categorization of legal trial data. — Researchers at Cornell University have identified persistent gendered language patterns in capital murder trials, suggesting significant bias in court records. AI Illustration. Upload story photo >

Live Poll

Do you believe the U.S. judicial system provides equal treatment regardless of a defendant's gender?

Cornell University researchers have published a study identifying distinct gendered language patterns in U.S. capital murder trials. The analysis suggests that legal proceedings frequently incorporate biases related to gender roles, sexuality, and identity.

Why it matters

The study aims to help legal professionals identify potentially prejudicial language that could impact the fairness of death penalty proceedings. Identifying these linguistic disparities is a step toward mitigating systemic bias in the justice system.

Researchers analyzed 137 U.S. murder trials, with records averaging 450,000 words each. The study utilized 500-word blocks of text to identify specific linguistic patterns across thousands of pages of court documentation.

The players

Cornell University

This private Ivy League research university is located in Ithaca, New York.

Emma Harvey

She is a researcher developing a tool to identify potentially biased language at the sentence level in legal documents.

The details

By applying natural language processing, the researchers compared trial records matched by crime characteristics, defendant demographics, and victim relationships. Findings indicate that women's trials more often featured language regarding gender roles and sexuality, while men's trials focused on violence, drugs, and racial tropes.

The study includes death sentences issued between 1990 and 2022.
The Supreme Court vacated a federal appeals judgment in the Andrew v. White case in January 2025.
Researchers identified the death row status of study defendants in July 2026.
Findings were presented at an ACM Conference from October 12-14, 2026.

Culture Shift

This research follows the Supreme Court decision in Andrew v. White, which vacated a federal appeals judgment in January 2025. It marks a shift toward using computational linguistics to evaluate and challenge long-standing biases within the American legal system.

The findings highlight how language used in courtrooms across the United States may influence the perception of defendants based on their gender. Awareness of these linguistic trends can help advocates and the public monitor for bias in local legal proceedings.

The takeaway

Legal professionals are increasingly turning to natural language processing to uncover patterns of bias in court records. Understanding these disparities can help ensure that judicial outcomes are based on facts rather than gendered social stereotypes.