Compensating for identification errors in memory recall tasks

Abstract

A mathematical model for approximating the contributions of identification and recall errors on the ability to repeat a list of N items is developed. In memory recall tasks where the items are independent and unrelated to each other, the probability of correctly repeating a list of items can be approximated as the product of the probability of correctly recalling all the items and the probability of correctly identifying an isolated item raised to the power of N. This relationship suggests that unaccounted for reductions in identification performance can severely affect estimates of memory.