The recordings of the tasks were first transcribed, and then the transcripts were analysed in various ways later on. The first step of the analysis involved finding the six stages of the fully-formed oral narrative as identified by Labov (1972) and described in the literature review section. The detailed evaluation criteria (see Appendix H) were compiled on the basis of articles using the Labovian system in the field of applied linguistics (Hatch, 1992; Liskin-Gasparro, 1996; McCarthy & Carter, 1994). The next stage was an in-depth analysis of the narratives uttered; an important issue here was the identification of appropriate chunks within the spoken data, which would reflect the psychological planning process (Foster, Tonkyn, & Wiggleworth, 2000), that is, the thoughts of the speaker. From the possible ways of segmentation, the method of dividing the texts into analysis of speech units (AS-units; see Foster et al., 2000) was adopted, as this system was specifically designed to handle oral data. The analysis followed the guidelines suggested by Foster and colleagues (see their article for a thorough description of the process). I expected that by counting the AS-units, the number of thoughts formulated by the respondent could be captured. The next step of the analysis involved identifying the narrative clauses (Labov, 1972) in the texts as defined in the literature review section above. The total number of narrative clauses as well as their ratio to the total number of AS-units were thought to reveal qualitative differences between the narratives produced. As regards the post-task interviews, notes were taken on the basis of the recordings, and these were analysed later on; full transcripts of the post-task interviews were not prepared.