Memory Capacity and Intelligence: Novel Techniques for Evaluating Rival Models of a Fundamental Information-Processing Mechanism

Abstract

Data from a fundamental cognitive task in which participants discriminate the relative frequency of visual or auditory binary stimuli were examined. Accuracy on this task correlates well with psychometric intelligence. The experimental paradigm is highly tractable, lending itself to rigorous analyses of precisely defined simulation models. Numerous models are evaluated, using multiple comparisons between response patterns of individual (and pooled) participants and predictive measures based on simulations for each trial sequence. Implications for theoretical accounts of short-term memory, discrimination, and absolute judgment, as well as the measurement of individual differences in cognitive ability, are discussed. The results suggest a reinterpretation of memory capacity and support a new kind of model (with a single estimable parameter) in which discrete, valued units of information are stochastically displaced by further input.