West · Psychology, public policy, and law : an official law review of the University of Arizona College of Law and the University of Miami School of Law 2025 · pre-registered randomized experiments and meta-analysis · n=1,316

Four (and a Half) Preregistered Failures to Replicate the Weapon Focus Effect in Online Samples.

Cited 0 times in the scientific literature.

Level 2 - randomized trial

Level 2 by design analogy (series of pre-registered randomized experiments with internal meta-analysis).

PubMed 41234291 · doi:10.1037/law0000469 · record verified 2026-08-26

What was done

Researchers conducted five pre-registered randomized experiments evaluating the Weapon Focus Effect (WFE) in online samples with sample sizes between 100 and 800 per experiment (total n = 1,316). Materials were drawn from studies that previously reported significant WFEs. Experiments systematically varied presentation modality (videos versus slides), audio presence, weapon type (knife versus gun), perpetrator gender (male versus female), and participant population (student versus non-student). Outcomes assessed were target physical description and lineup identification accuracy, followed by random-effects meta-analyses across all experiments.

What was found

None of the individual experiments showed a statistically significant WFE for target description or lineup identification. Some experiments yielded significant anti-WFEs, where participants exhibited better memory for armed compared to unarmed perpetrators. Random-effects meta-analyses across all five experiments showed a nonsignificant effect of weapon presence on both memory outcomes. No specific effect sizes, test statistics, or p-values were reported in the abstract.

Why it matters

The weapon focus effect is widely accepted in legal and forensic settings to explain witness memory deficits. These findings show that the effect does not reliably replicate across online experimental conditions, suggesting it may be fragile or highly dependent on specific laboratory methodologies.

Limits

All experiments relied on online samples, which may differ from in-person laboratory environments in participant attention, screen size, and viewing conditions. The abstract omits exact point estimates, confidence intervals, and test statistics. Real-world high-threat eyewitness scenarios were not directly evaluated.

Cited by