A markov model for measuring storage loss and retrieval failure in retroactive inhibition

Abstract

A Markov model is proposed to measure storage loss and retrieval failure in retroactive inhibition. This model generalizes Batchelder and Riefer's (1986) storage-retrieval model to account for retroactive inhibition. Storage loss is assumed to result in a disintegration of clustered items, while retrieval failure is assumed to prevent access to clustered items. The model is tested against data from three free recall experiments in which categorized item lists were presented to subjects. The kind of item categorization (conceptual vs. phonemic-orthographic) and the interlist similarity (different vs. identical categories) were manipulated across experiments. The model fits the data of all three experiments well, thus supporting the central inhibition concepts of the model. Based on these inhibition concepts and the parameter estimates of the model conclusions are drawn on the issue of whether retroactive inhibition is due to storage loss or retrieval failure in this experimental situation. In addition, rates of storage loss and retrieval failure are compared across the experiments. This latter analysis is discussed as an example of how the model reveals information about retroactive inhibition in ways that the storage-retrieval model cannot.