The Impact of Disparities in Collaborators’ Encoding Levels on Collaborative Memory

This study aims to investigate the effects of different encoding-level combinations on collaborative memory retrieval and subsequent individual memory performance, with the goal of deepening the understanding of collaborative inhibition mechanisms and providing evidence-based strategies for optimizing collaborative learning. Using an experimental approach, 160 healthy university students were recruited as participants. Collaborative groups were organized into dyads, with three encoding combinations (40 participants each, i.e., 20 dyads per combination). Additionally, 20 participants each from semantic and perceptual encoding conditions were randomly paired to form nominal groups.The experimental materials comprised standardized two-character Chinese word lists presented via E-Prime. Sixty neutral low-frequency two-character words were selected from the Modern Chinese Frequency Dictionary of Common Words (Language Teaching Institute, Beijing Language and Culture University, 1986). These included 30 nouns and 30 verbs, with each word category further divided into two structural types: "fully left-right structured" (both characters having left-right configurations, e.g., "设计[design]") and "non-fully left-right structured" (at least one character lacking left-right configuration, e.g., "技术 [technology]", "家具[furniture]"), resulting in four word categories.During the encoding phase, all participants underwent either semantic (deep) or perceptual (shallow) single-encoding processing. In the retrieval phase, collaborative groups in each of the three encoding combinations (semantic+semantic, perceptual+perceptual, semantic+perceptual) separately conducted collaborative retrieval and individual retrieval tasks, while nominal groups performed independent retrieval throughout the entire process. Data collection encompassed the number of correct retrievals and errors during both collaborative and secondary individual retrieval phases, analyzed using mixed-design ANOVA.The complete dataset comprises two subsets: Data_1 (collaborative retrieval performance, including accuracy rates and false memory quantities during collaborative retrieval) and Data_2 (post-collaboration individual retrieval performance, containing accuracy rates and false memory metrics). This dataset systematically reveals the dynamic effects of encoding combinations on collaborative memory, particularly demonstrating how heterogeneous encoding pairs (semantic+perceptual) enhance memory performance through dual mechanisms of strategy differentiation (reducing interference via divergent retrieval approaches) and knowledge transfer (low-level encoders benefiting from high-level partners during collaboration). The granular behavioral metrics and structured trial-by-trial records support cross-disciplinary investigations into collaborative memory dynamics and educational strategy optimization.