Cross-modal feature and conjunction errors in recognition memory

Abstract Four experiments were conducted to investigate whether a modality-specific familiarity contributes to feature and conjunction errors and, hence, to recognition memory. In each experiment, the presentation modality of compound words was manipulated at study (auditory or visual), and in Experiment 2 the presentation modality for the test also was manipulated. In Experiment 3, participants were pushed to respond quickly in order to create a reliance on familiarity rather than recollection. In Experiment 4, a direct manipulation of response deadline was employed. Across experiments, auditory and visual tests did not produce different hit rates or feature and conjunction error rates, and shifts in study-to-test modality did not affect hit rates or feature and conjunction error rates. The response deadline manipulation of Experiment 4 affected old/new discrimination but not feature and conjunction effects (feature/new and conjunction/new discrimination), producing a dissociation. Unlike implicit perceptual memory, modality information does not appear to contribute significantly to the familiarity underlying feature and conjunction errors. The familiarity underlying feature and conjunction errors, and thus in recognition memory, is different from the familiarity underlying perceptual implicit memory.