The validity of speech act data taken from different kinds of elicitation instruments has been widely discussed in interlangauge and cross cultural pragmatics literature. In this study an attempt is made to evaluate and investigate data taken from two of the most popular speech act instruments namely, written DCT and closed role play. The responses of forty Iranian university students in their native language (Persian) were investigated across the speech act of request. The distinguishing feature of the study is using the same participants for the two methods; in this way we can have an account of intra-participant variations. The social status and distance of the participants were both equal. Differences were found in the length and content of the responses. Respondents to role play tended to have longer responses and it was mainly because of the longer and higher number of alerters and supportive moves used in the role play. In the written data we witnessed more direct strategies used in the request Head Acts. Likewise, modification devices used in the oral data had a softer tone and in terms of request perspective the oral data provided more impersonal responses while the requests in the written data were more hearer-oriented. Overall, it seems that the data gathered through role play is more natural than DCT. Finally it is discussed that the choice of data gathering instrument in pragmatics research is to a large extent dependent on the goals of the researcher and research questions.