Rethinking visual working memory capacity: Proactive interference cannot account for real-world object advantage
Undergraduate Just-In-Time Abstract
Poster Presentation 23.350: Saturday, May 16, 2026, 8:30 am – 12:30 pm, Banyan Breezeway
Session: Undergraduate Just-In-Time 1
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
Anantya Engineer1 (), Yong Hoon Chung2, Jamal Wiliams1; 1University of California Santa Cruz, 2Dartmouth College
Traditional theories of visual working memory (VWM) suggest that we have a capacity limit of around 3-4 items. However, people have been found to remember more real-world (RW) objects compared to simple stimuli like colors (Brady et al., 2016). Recent findings challenge the cause of this advantage by showing it can be eliminated when RW stimuli are repeated over multiple trials, suggesting that the RW benefit may instead be due to an imbalance of proactive interference (PI). This suggests that simple stimuli such as colors are subject to more interference from previous trials since many studies showing the RW benefit use unique objects on every trial. Under this account, once PI is matched across colors and RW objects, the RW object advantage should disappear. In the present study, we explore this hypothesis through direct and conceptual replications of recent work. Surprisingly, we find no evidence of a RW object benefit and instead find a robust “meaningless” benefit, where colors are remembered better than RW stimuli (exp 1). This effect appears to be driven by the distinctiveness of each color since replacing the colors used in previous studies with CEILAB colors this “meaningless” benefit disappears (exp 2). Across several studies we systematically manipulate (1) spatial cues, (2) encoding time, (3) stimulus sets, and (4) probe types (e.g., recognition memory vs change detection) to eliminate or exaggerate the effects that PI might have on VWM. Together, these findings draw attention to the fragility of any PI that is claimed to be the driving force of performance differences across conditions, and demonstrate the effect that stimulus distinctiveness can have in a design. Importantly, they also highlight the importance of properly curated stimulus sets when evaluating theoretical claims regarding WM capacity.
Acknowledgements: “This work was supported by the laboratory of Jamal Williams, University of California, Santa Cruz”