3.4 Data Analysis
Phase 1 semi-structured interviews were analysed using an inductive thematic analysis. Phase 2 survey Likert scale questions and text response questions were analysed using basic statistical analysis and thematic coding.
3.4.1 Phase 1
According to Braun & Clarke’s Using Thematic Analysis in Psychology (2006), thematic analysis is a foundational method for qualitative research. Thematic analysis offers researchers a consistent, repeatable method for identifying, analysing, and reporting patterns in data. Using themes allows complicated data from qualitative research to be made available to a wider audience (Braun & Clarke, 2012). The use of themes that “resonate with the data, with the participants, with the ongoing too muchness of research” through thematic analysis creates a much stronger and more effective report (Cannon & Edber, 2025). Resonance developed in this way promotes “empathy, identification, and reverberation of the research by readers who have no direct experience with the topic discussed … to transform the emotional dispositions of people and promote regard” (Tracy, 2010, p. 844). Presenting student experiences with digital learning material was designed to engage the empathy of post-secondary employees, making thematic analysis employed as Tracy described an ideal method for this research.
The analysis of the student co-design sessions followed Braun & Clarke’s inductive thematic coding (2006). This bottom-up approach ensures themes arise from the data themselves as opposed to using the data to validate preconceived ideas. This approach is particularly important for research without a significant amount of existing literature to test against. Inductive thematic coding ensures the themes presented to readers are driven by what is in the data (Braun & Clarke, 2012). Identifying patterns in the students’ experience with digital learning material in a systematic and deliberate manner ensures a clear and digestible report for readers while honouring the students’ lived experience.
To analyze the data, Braun & Clarke’s 15-point checklist for good thematic analysis was employed (2006, p. 96). The first step was transcription. In-person sessions were digitally recorded and transcribed using Microsoft Word’s Transcribe tool within OCADU’s O365 cloud environment. Remote sessions were hosted on Zoom. The Zoom meetings were recorded with the audio separated from the video (which was securely deleted) and transcribed in the same manner as the in-person recordings. Transcripts were read and verified against the digital recordings. Through validating the transcription, and in review of field notes and artifacts, the Graduate Student Researcher became familiar with the general ideas and key points in conversation revealing patterns and repeated concepts.
During the verification, reading and rereading initial codes were developed. Data was coded within 72 hours of each interview. Initial themes were noted, but no themes were formalized until all interviews were completed. Codes were developed inductively and iteratively, tweaking and refining codes for uniform language and reducing synonyms through merging of similar codes. Within each transcript each code was tagged with a comment. Then, each code was copied to a heading of the same name. This allowed a grouping of similarly coded sections of each transcript. Codes were checked against session transcripts then collated with excerpts, artifacts, and notes to ensure a strong link between the data and codes. Once all session transcripts were coded, themes were developed.
Codes were grouped into broad topics to organize coded data with similar evidence. As with coding the data, this process began to reveal patterns which formed the basis for themes. This process was done on an Excel spreadsheet (Appendix D). Potential themes were then checked against one another, the individual codes first identified, and the entire set of codes for consistency, relevancy, and coherence. Themes were refined, revised, and merged to ensure relevance and discretion. For example, the code ‘time’ became an obvious theme as all students mentioned the significant amount of extra time, they needed to complete tasks. However, some content initially coded as ‘barriers to learning’ or ‘negative experience’ were explicitly about time costs and were re-coded to that theme. The first session occurred on August 20, 2025, and the last session occurring on September 4, 2025. Data was coded and themes were developed by September 30, 2025.
Inductive, data-driven, and iterative coding ensured themes were reflective of the students’ responses and did not allow researcher biases to influence the data analysis. Of particular importance was identifying themes that respected the co-designers’ intent and reflected their specific input while ensuring the themes would be meaningful to readers as well as relevant to the topic and research question being explored (Braun & Clarke, 2012). Given the researcher’s dual role and pre-existing relationship with co-designers, bracketing was employed to “suspend … researcher biases … [and] reflect on the social, cultural, and historical forces that shape their interpretation” (Creswell & Miller, 2000, p. 127). Bracketing is a process in which researchers consider their values and assumptions and attempt to put them aside, in brackets, while conducting research (Draper, 2004). In the data analysis, bracketing helps to exclude personal bias to ensure co-designers’ experiences were considered impartially with maximum accuracy of each co-designer’s intent, not the researcher’s preconceptions.
Once the resource for phase 2 was developed, co-designers were contacted for feedback. Co-designers were contacted via email between six and seven weeks after their session. All co-designers responded, some only affirmed their participation while two provided specific notes for minor changes to language.
3.4.2 Phase 2
As phase 2 data shifted from qualitative to mix-methods during this project, additional data analysis methods were included. Based on the National Academies of Sciences analysis techniques for small population research and Statistics Canada’s guidance on survey theory for small samples, percent calculations were applied to the ordinal Likert responses (Rao & Fuller, 2017; National Academies of Sciences, 2018). Thematic coding was applied to the open-ended text responses per Braun & Clarke (2006) as discussed above.
