Experimental Topics Seminar
September 26, 2003, 3:30 pm
Byrne 403
Lists of monosyllabic words are better recalled than lists of otherwise equivalent multisyllabic words on tests of immediate serial recall. This segment-based word length effect is traditionally observed in experiments that use pure lists: in each list, only short or only long items occur. With mixed lists, in which both short and long items occur, the empirical evidence is mixed: Cowan et al. (2003) found a word length effect in mixed lists whereas Hulme et al. (in press) did not. A series of experiments is reported that resolves this issue: Whereas word length effects were invariably found in pure lists, no such effects were observed in lists in which both short and long items appear, contrary to predictions of all current theories of short-term/working memory. A new model that accounts for the data is briefly outlined.