Human Memory and the Reliability of Eyewitness Testimony: An Empirical Analysis of Misidentification, Memory Distortion, and Evidentiary Reliability

Abstract

Eyewitness testimony remains one of the most influential forms of evidence in criminal adjudication, yet empirical research demonstrates that human memory is vulnerable to systematic distortion. This study examines the reliability of eyewitness testimony through an empirical analysis of publicly available exoneration data, experimental research, systematic reviews, and documented eyewitness-identification cases. Rather than relying upon fabricated primary-survey responses, the study employs secondary empirical data from the Innocence Project, the National Registry of Exonerations, the National Research Council, and peer-reviewed psychological research available up to December 2024. The analysis demonstrates that eyewitness misidentification is a recurring contributor to wrongful convictions and that its significance cannot be explained solely by individual memory failure. The Innocence Project's dataset of 375 DNA exonerations in the United States records eyewitness misidentification in 69% of cases, with 42% involving cross-racial misidentification, 32% involving multiple misidentifications of the same person, and 84% involving misidentification by a surviving victim. The National Registry of Exonerations further recorded mistaken witness identification as a contributing factor in 43 of the 158 exonerations reported for 2024, representing approximately 27% of those cases.