What the study analysed
The convergent mixed-methods dissertation, published open access on eScholarship, examined how undergraduates describe and interpret holistic wellbeing. The quantitative strand validated a new eight-dimension wellbeing survey with a final analytic sample of 251 full-time undergraduates. The qualitative strand followed 18 first-year students enrolled in a yoga-based general education seminar in Spring 2025.
Each participant produced nine weekly journal reflections, compiled into a document of roughly 10 to 25 single-spaced pages, alongside a 30 to 60 minute semi-structured interview. Interviews were transcribed with Otter.ai, manually corrected, and de-identified with pseudonyms before any analysis began. The study was approved by the university’s Institutional Review Board.
A deductive frame, kept open to new codes
The researcher entered a set of predetermined codes drawn from the eight dimensions of holistic wellbeing and from contemplative learning constructs such as self-awareness, emotional regulation, and perspective shifts, then instructed the platform to stay open to codes emerging directly from participants’ language.
Three analyses were run and compared: one over journals, one over interview transcripts, and one combining both sources for overarching themes. In the journal analysis the deductive structure performed as intended, organising responses under all eight wellbeing dimensions. A second pass over the interview data with revised prompts produced an alternative five-theme structure, which the researcher weighed against the first four-theme structure before the combined analysis settled on three overarching themes.
How Evidano fit the workflow
AI organised the volume; interpretation stayed with the researcher.
- 1Data compiled and de-identified
One document per participant, combining welcome survey responses and every journal submission, with interviews transcribed and manually corrected.
- 2Structured prompts and predetermined codes entered
Prompts were aligned to the research questions and to the eight wellbeing dimensions, while leaving room for emergent codes.
- 3Three datasets analysed independently
Journals, interviews, and a combined set, each uploaded separately to preserve the context of the source.
- 4Every code reviewed and interpreted by the researcher
Initial codes and suggested themes were refined and read back against the research questions, with the author’s familiarity as course instructor used to check nuance, divergence, and counterexamples.
“I used AILYZE, an AI-assisted qualitative analysis platform, to support the coding process by organizing large volumes of text data, identifying preliminary patterns, and generating initial thematic structures.”
“Based on these inputs, AILYZE generated initial codes and suggested themes for each dataset, which I then reviewed, refined, and interpreted in relation to my research questions.”
“In the present study, using AI enabled multiple rounds of thematic exploration across journal reflections and interview transcripts, allowing for the comparison of alternative coding structures and thematic interpretations that would have been substantially more time-intensive using traditional approaches alone.”
“While tools like AILYZE cannot replace the interpretive role of the researcher, they may offer valuable support in organizing large datasets, identifying preliminary patterns, and facilitating iterative cycles of coding and thematic development.”
Why AI-assisted over manual coding alone
Comparing alternative coding structures across hundreds of pages of journals and transcripts is slow enough that most researchers settle for a single pass.
Multiple rounds of thematic exploration across both data sources, so alternative structures can be compared rather than assumed.
The platform is named and cited in the data analysis chapter and the reference list.
Predetermined codes were entered up front, with emergent codes still welcome.
Transcripts were de-identified and pseudonymised before upload.
Every code was reviewed, refined, and interpreted by the author, who describes the platform as a thought partner rather than an independent analyst.
About Evidano (previously AILYZE)
Evidano is AI-assisted qualitative data analysis software for interviews, focus groups, open-ended survey responses, and documents. It supports AI-assisted thematic analysis with your own or an AI-suggested codebook, transcription and translation in 100+ languages, clickable quotes and citations, and exportable visual reports. End-to-end encryption, no third-party data sharing, and your data is not used to train AI. From USD 50/month for unlimited analysis.
Run the first pass in days, not weeks
Keep the judgment with you. Let Evidano do the heavy lifting on your interviews, focus groups, open-ended responses, and documents.
More case studies
AcademiaVolda and Melbourne researchers used AI interview analysis to triangulate manual coding
Two researchers coded 11 interviews by hand with over 90% agreement, then ran the anonymised transcripts through Evidano and compared the results for triangulation and validation.
AcademiaStanford and UC Davis researchers used AI to cross-check their qualitative coding
A published mixed-methods study uploaded deidentified transcripts into Evidano to cross-check its themes. Every AI code was reviewed by the researcher.
AcademiaCornell’s CLASP used AI thematic analysis on its program film. Adult educators concurred.
A published Cornell study used Evidano (previously AILYZE) for thematic analysis of its program film, then compared the AI’s reading with expert educators’ reactions at AAACE 2023.
Related published evidence
Researcher interviews and published citations covering comparable studies.
- Published research2026
Sharjah Education Academy: AI thematic analysis of 9 teacher interviews, human-verified
Published in Frontiers in Education (2026): encrypted data, human-verified themes, and a transparent AI coding workflow across English and Arabic interviews.
9 Teachers interviewed
Frontiers in Education, 11:1690181 (Frontiers Media, 2026), open access
- Published research2026
Five Ukrainian universities used AI-assisted grounded theory on 51 literature sources
A peer-reviewed, Scopus-indexed grounded theory study processed 51 literature sources through Evidano to formulate the categories that fed its constructed theory.
51 literature sources processed
Revista de Cercetare si Interventie Sociala, Vol. 93 (2026), pp. 43–66, peer-reviewed and Scopus-indexed
- Published research2026
University of San Carlos study used AI interview analysis on 15 government officials
In-depth interviews with 15 local government officials were analysed thematically in Evidano, using valence coding to classify responses across three Philippine municipalities.
15 local government officials interviewed, 58 minutes on average
Preprints.org, posted 16 June 2026 (preprint, not peer-reviewed; CC BY 4.0)
Explore more evidence
Researcher interviews, the wording authors published when citing Evidano, and accuracy comparisons against manual coding.
- Resource
Research spotlights: researchers in their own words
Interviews with researchers at leading universities and institutes about how they actually used Evidano in work they went on to publish.
- Resource
How Evidano was used, in published methods sections
Verbatim excerpts from 20 published studies describing what Evidano did in their analysis — as a thought partner, a second coder, for initial coding, or for synthesis.
- Resource
Methodology guides and research writing
Practical guides to qualitative methods — thematic analysis, grounded theory, evidence synthesis and more — alongside the wider Evidano article library.
- Resource
Human vs AI: validated accuracy benchmarks
Three head-to-head comparisons against expert human analysis — 371 interview transcripts with Arizona State and Penn State, 298 evaluation reports for a UN evaluation group, and UNICEF’s manual-coding review.