Abstract
The author discussed the methodological problems of the analyses of narrative texts (Saijo, 2002; Yamada, 2002) from
the standpoint of experimental psychology. Main critiques were as follows: (a) their descriptions of hypotheses were
abstract and ambiguous, (b) every illustration could not guarantee any logical justifications in hypothesis-testing, and
(c) their arguments had some important inferential jumps. It was suggested that the analysis of narrative texts was not
necessary to discuss their themes and, instead, the author offered several experimental designs to test their hypotheses
in terms of random sampling, control group, comparison group, control period, extraneous variable, and intervening
variable. Lastly, some promising narrative analyses using on-line chat, bulletin boards, and e-mails were proposed.
