Speech act theory has greatly influenced both the study and teaching o
f language use; however, given its philosophical roots, it has been sh
ort on empirical support. In recent years numerous speech act studies
have been conducted. In these studies, though, there has been little d
iscussion of research methodology, specifically data collection techni
ques. This paper reports the results of a study (based on Rose, 1994),
which addresses methodological validation in speech act research. We
administered discourse-completion tests (DCTs) and multiple-choice que
stionnaires (MCQs), designed to elicit requests, to two groups of 36 J
apanese female undergraduates. As with Rose (1994), there were signifi
cant differences in most situations (11 of 12), with those completing
the MCQ choosing to opt out or hint more frequently than those complet
ing the DCT. These results seem to indicate some serious problems with
DCTs, which need to be addressed if they are to be used in speech act
studies.