Hanna de Jong-Markus

Methods 2 45 original language of the participants, on the other I also used ideas for categories that came from my earlier insights and from insights during the coding of earlier meaning units. The codes were continuously defined and redefined. Meaning units with the same codes formed categories (Elliot & Timulak, 2005; Miles et al., 2014). These categories were in turn also categorized (Elliot & Timulak, 2005) by looking for similarities and patterns between them, in this way forming increasingly abstract categories. Patton’s guidelines (2002) for dual criteria judging categories—internal homogeneity and external heterogeneity—were helpful in this phase. Next, the main findings were abstracted by finding an answer to the question, ‘What categories are required to communicate the essence of the phenomenon?’ (Elliot & Timulak, 2005). In some sub-studies (described in Chapters 4 and 5) this step was extended by imparting the main findings with related information or examples from other data fragments (distinctive meaning units or domains) or by checking whether other data fragments contained contradictions with the findings. Finally, the results were written down by presenting the findings and describing the relations between them. I tried to regularly use quotations, as I wanted to be clear about the origins of the results plus stress the voices of teachers. Data analysis of the focus groups study For the data analysis in the focus groups study a generally comparable procedure was followed, but with detailed reports of the focus groups. These reports were created by the assistant and checked by the researcher using the audio recordings. The meaningful units were distinguished in close consultation between researcher and assistant, and thematically coded partly by each (cf. Braun and Clarke, 2006; Elliott and Timulak, 2005) while they checked each other’s work. Because of the relatively small amount of data and prior experiences of the assistant, Microsoft Excel25 instead of Atlas.ti was used. The codes were then jointly categorized (cf. Elliott and Timulak, 2005). Based on these categories and codes the results were described, with some of the pre-formulated questions reformulated. One of the supervisors in the research team read along critically during the whole process, while the results were also discussed with the entire research team. To stay close to the experiences of participants, the data were open-coded and the results were only compared with literature afterwards instead of using literature for defining the categories. 25 https://www.microsoft.com/nl-nl/microsoft-365/excel