Research suggests that people's memory; span for Arabic numerals is consist
ently greater than that for digit words written in an alphabet, even when t
hese stimuli have been matched for word length, lexicality, and semantics.
It is nor clear whether this "numeral advantage" holds true for other scrip
ts. The present study test ed this possibility in the case of Japanese kanj
i (the Chinese characters used in Japanese). 23 Japanese adults responded t
o stimuli comprised of two sets of random number sequences, one set present
ed in Arabic numerals and the other in digit word kanji. Each set began wit
h three two-item sequences followed by three three-item sequences and so on
, until three 12-item sequences were presented. Consecutive repetition of s
timuli was avoided. Memory span was measured as the length of the last corr
ectly recalled item sequence. Results indicated no numeral advantage effect
, as there was no significant difference between people's recall of numeral
s and digit words. Memory strategy in dealing with kanji appears to hear mo
re resemblance to numerals than to words in English.