Data collection and analysisData we

Data collection and analysis

Data were collected through interviews (see Appendix 1 for interview guideline) and participant observation as they were seen as useful methods for exploring learners’ perceptions and gaining rich data of participants’ behaviour and actions within a context (Mackey & Gass, 2005). Data for this study were generated from participant observation of eight lessons in each aforementioned class and interviews right after the completion of the course in the semester. Given that learners may not be initially familiar with the TBLT method, participant observation was carried out from the third lesson onwards in each class. Unlike the unfamiliar presence of an outside researcher, the teacher-researcher was advantageously able to observe the class with minimal influence. Observations focused on learners’ performance and their reactions during task completion. Observation notes were made by the teacher-researcher immediately after each lesson and were summarised for analysing. To have an in-depth understanding of TBLT in CFL, semi-structured interviews were conducted focusing on learners’ feedback of the CFL class, learners’ experiences with TBLT, and learners’ suggestions for future learning. Interviews were carried
out in English immediately following the completion of the course. Due to one participant’s personal schedule, 17 participants were interviewed individually, and each interview lasted for around 30 minutes. All interviews were transcribed by the first author.

Data were analysed by an inductive approach and coding was grounded in the data. Firstly, we repeatedly read through the interview transcripts and field notes from the observation through open coding, and assigned the codes to pieces of data pertinent to the research questions and other emerging issues (Mackey & Gass, 2005). These codes provided an index to store and retrieve the data by recursive examination. Secondly, we merged the related codes into categories and developed them into a number of themes. Once these themes were established, we pooled all data touching on these themes. Finally, we moved iteratively from the data to the established themes and conceptualised final themes for this study through axial coding. In this study, the interview transcripts formed the primary data and the observation notes as the secondary. In the ‘Findings’ section, the italicised words in the brackets in the displayed quotes are the researchers’ interpretations.