MINERVA-DE: An instance model of the deficient processing theory

This article focuses on a novel recognition memory effect reported recently: better recognition at test for a word seen once than for a word seen twice at study (Rosner, López-Benítez, D'Angelo, Thomson, & Milliken, 2018; Collins, Rosner, & Milliken, 2018). Collins et al. (2018) suggested that the repetition decrement effect results from temporary, familiarity-mediated shifts in exogenous attention that bias attention toward unfamiliar, not-repeated words and away from familiar, repeated words. The current study presents a formal model of this theory, MINERVA-DE, where DE stands for discrepancy encoding. Discrepancy encoding assumes the encoding of stimuli is inversely proportional to its familiarity in primary memory. We show that the model predicts the entire range of reported repetition decrement effects and holds promise for modeling the deficient processing theory of the spacing effect in recognition.