The development of serial short-term memory and the articulatory loop hypothesis

Abstract

The working memory model attributes the development of span in children to increasing proficiency in the use of an articulatory loop (AL). This hypothesis was investigated by testing samples of two different age levels with a fixed span task, and with a running memory task that required the recall of the 5 terminal items in each list. Some of the results supported the AL hypothesis: Fixed span and articulation rate correlated within each of the samples; the effect of age on fixed span disappeared when articulation proficiency served as a covariate in an analysis of covariance. Several findings challenged the validity of the hypothesis, however. For example, the reliable correlation coefficients found between fixed span and running memory were not generally reduced when the influence of articulation speed was partialled out. This indicates a common factor underlying individual differences in the span and running memory tasks, not related to articulation proficiency. Alternative approaches to the span—articulation proficiency relationships are discussed, including the possibility that span and articulation develop as separate facets of a common system.