EN
The article examines the issue of the reliability and validity of the discourse completion task (DCT) in research projects on speech acts. It starts with a review of the main data collection methods in pragmatics. In the paper, the advantages and disadvantages of using a questionnaire for gathering data are discussed. Next, the studies comparing questionnaire data with natural spoken data are presented. The paper argues that relying on one data collection instrument, namely a DCT, does not yield reliable results; therefore the integration of authentic and controlled data is desirable. It is concluded that the study of a particular speech act cannot be carried out without some reliance on naturally occurring data.