Abstract
Baddeley, Thomson, & Buchanan (1975) found that phonological length of words determines memory span, i.e., the word-length effect. However, the longer words used in Baddeley et al. (1975) had more letters, and it was undeniable that number of letters also influenced memory span. Especially, native Japanese readers are found to rely more heavily on visual codes and less on phonological codes in letter processing than native English readers (Mizuno, Matsui, & Bellezza, 2007). We, therefore, hypothesized that memory span of native Japanese readers is rather influenced by number of letters than phonological length of stimulus words. An experiment was conducted to measure memory span and reading speed of native Japanese readers using three-mora words composed of one, two, and three Kanji characters. The results showed that the recall rate and reading speed of three-character words was less than those of two-character and one-character words, indicating the validity of our hypothesis.
