Evidence of a linear relationship between speech rate and memory span
in children has been obtained in several studies (e.g. Hulme, Thomson,
Muir, & Lawrence, 1984). This evidence is used to support an explanat
ion of the development of memory span based on the working memory mode
l (Baddeley, 1990). The model argues that speech rate is related to th
e amount recalled and that developmental increases in speech rate allo
w faster rehearsal with age and, hence, greater recall. However, the l
inear relationship between speech rate and memory span has generally b
een reported in terms of group means for speech rate and memory span r
ather than individual level correlations between the two variables. Th
e present studies replicate the group relationship, but find that corr
elations between individual subject's speech rates and memory spans, w
hen the effects of age are partialled out, are no longer significant.
Nor was the size of the word length effect related to the difference i
n speech rate between short and long words. It is argued that the grou
p mean relationship between speech rate and memory span is clear and r
eplicable, but that the speech rates of individual children are not go
od predictors of those children's memory spans. The implications of th
ese results for the working memory explanation of span development are
discussed.