1. Familiarisation with the dataset | Padlet responses were organised in an excel spreadsheet. The authors familiarised themselves with responses, noting any initial impressions and ideas. |
2. Coding | Responses were systematically coded by BS & HMW, based on sematic similarities. Where participants had made multiple points within one posted comment these were separated so that each point could be coded individually (520 posts responding to 8 prompts were coded as 967 separate points). The researchers discussed these codes and differences of opinion with the wider team. It was noted that while a substantial portion of codes corresponded to the ‘coping reservoir’ conceptual framework (deductive coding), additional codes were also required (inductive coding). |
3. Generating initial themes | Codes were grouped together to form provisional candidate themes by RMD, LF, OTE and HMW, adding notes on the impressions of data items. Candidate themes were discussed as a research team, particularly checking if they resonated with the experiences of researchers who were medical student and recent graduates. |
4. Developing and reviewing themes | The researcher team reviewed and refined the provisional themes, revisiting the original dataset to check that the candidate themes presented an accurate summary of concepts that was grounded in the data. Themes were organised and reorganised until the team was satisfied that data were represented in a meaningful way. |
5. Refining, defining and naming themes | OTE & HMW produced a synopsis of each theme, selecting data extracts that reflected the content. This was circulated to the other authors, who further refined the descriptions so that the scope and content of each theme was clear. Team consensus was sought for themes and names. |
6. Writing up | The researcher team revised the written analysis, in line with the SRQR reporting standards [22] (Appendix 2), documenting the reasoning behind theoretical, methodological and analytical choices. Contextual descriptions of the themes were constructed, with a logical flow, embedding participants’ responses to illustrate the validity of the analysis. |