A Bayesian procedure for separately estimating storage and retrieval components of forgetting

A Bayesian procedure was developed to estimate separately the probability of sufficient storage, θ_s, and the probability of successful retrieval, θ_τ, on an individual subject basis. The method involves a probabilistic analysis of the subject's performance on a task that randomly intermixes recall with recognition trials. Two experiments using the Brown-Peterson paradigm were performed to demonstrate and validate the estimation procedure. The experiments showed that both θ_s and θ_τ decrease as a function of increasing retention interval. The results of both experiments suggest the necessity of a two-factor theory for short-term memory.