Re-Examining Memory Distortion Across Repeated Reproduction Using Human and Computational Approaches

Despite its renown and importance to the field of cognitive psychology, Sir Frederic Bartlett’s (1932) study on repeated reproduction of memory has seldom been directly replicated (Ahlberg & Sharps, 2002; Bergman & Roediger, 1999; Johnson, 1962). Keeping the same procedures, this proposed project aims to be a modern-day direct replication and extension of Bartlett’s study by adding new materials beyond just “The War of the Ghosts,” stronger controls for confounds, such as recall instructions, and novel ways of computationally measuring how memory is distorted over time using semantic similarity comparisons and representational similarity analyses. If successful, this research project would not only be the first direct replication of one of the most famous studies in cognitive psychology in the past 20 years, but would also allow Bartlett’s findings to be further generalized through the additional materials, controls, and measures. Additionally, the results of the novel computational analysis could provide a deeper understanding of how schema influences memory recall top-down. By plotting each successive recall in the same common meaning space, one can visualize the trajectory of memory distortion, and potentially point to the underlying schema that distorts memory. Ultimately, this project will not only provide insight into the nature of schemas but also contribute to a burgeoning literature on using novel computational techniques to analyze memory recall (Fenerci et al., 2025).