Last updated: July 6, 2026
\n\nReflexivity is the practice of examining how your own assumptions, position, and choices shape a qualitative analysis. A 2026 CHI study by Ye and colleagues found that AI can support reflexivity — not by coding for you, but by prompting reflection in the moment, making code evolution visible, and turning coder disagreements into productive dialogue. The key shift: design AI tools for interpretive depth, not just speed.
\n\nReflexivity is the researcher's ongoing, critical examination of how their background, assumptions, and analytic decisions influence what they see in the data. In reflexive thematic analysis — the approach popularized by Braun and Clarke — subjectivity is not a bias to be eliminated but a resource to be made explicit. Themes are actively constructed by the analyst, so the credibility of the work depends on the analyst accounting for their own positionality.
\nThis is precisely the part of qualitative work that most AI tooling ignores. A model can suggest codes quickly, but it cannot notice that you are pattern-matching toward a hypothesis you already hold. That gap is what the 2026 study set out to address.
\n\nAccording to Ye et al. (CHI 2026), existing qualitative software \"prioritizes efficiency over interpretive depth.\" The authors built Reflexis, a collaborative analysis workspace designed around deliberation-supporting features, and evaluated it with 12 paired analysts working through reflexive thematic analysis together.
\nRather than measuring how fast coding went, the study looked at whether the tool deepened reflection and improved the quality of collaborative interpretation. The reported findings were qualitative: Reflexis \"encouraged more granular reflection,\" made code evolution transparent, and turned coder disagreements into \"positionality-aware dialogue\" instead of friction to be smoothed over.
\n\nThe study points to a design pattern that generalizes beyond one tool. Here is how \"efficiency-first\" AI differs from \"reflexivity-supporting\" AI in qualitative analysis:
\n| Design goal | Efficiency-first AI | Reflexivity-supporting AI |
|---|---|---|
| Coding | Auto-assigns codes to save time | Suggests codes but prompts the analyst to justify or revise them |
| Disagreement | Resolves to a single \"correct\" label | Surfaces the disagreement as material for interpretation |
| Code history | Hidden; only the final codebook is shown | Visible; how and why codes evolved is tangible |
| Reflection | Not prompted | In-situ prompts at meaningful decision points |
| Positionality | Treated as noise | Treated as a resource made explicit in memos |
The lesson is not \"add more automation.\" It is to point the model at the moments where a human is about to make an interpretive leap, and slow that moment down.
\n\nWhen a tool hands you a finished set of codes, it is easy to accept them — a form of automation bias. The reflexive tradition treats that as a threat: if the analyst never wrestles with the data, the \"themes\" belong to the model's priors, not the researcher's grounded interpretation. The 2026 study's contribution is a design framework for tools that \"prioritize rigor and transparency to support deep, collaborative interpretation,\" rather than treating analysis as a task to be finished as fast as possible.
\nThe study also flags limits. Analysts wanted control over when prompts fired, and wanted memos that connect across the project rather than isolated notes — a reminder that reflection cannot itself be fully automated without becoming a checkbox.
\n\nQualitati is an AI qualitative research platform built to keep the researcher in the interpretive loop. Its ThemeLens thematic-analysis pipeline maps codes to research questions and anchors every theme to participant quotes, so you can audit what the model based a theme on instead of trusting a black box. The QDA Workspace supports human-led inductive and deductive coding with AI assistance, keeping the codebook — and the reasoning behind it — in your hands. Reflexive memoing and positionality still belong to you; the tools are there to make your interpretation more transparent, not to replace it.
\n\nIs reflexivity relevant if I use AI for coding? Yes — arguably more so. AI can introduce its own patterns, so you need to account for both your positionality and the model's influence on the analysis.
\nDoes the study say AI improves reflexivity on its own? No. It found that a tool designed around reflection and deliberation supported reflexivity; generic, efficiency-focused AI was the problem it was reacting to.
\nWhat is reflexive thematic analysis? An approach (associated with Braun and Clarke) in which themes are actively constructed by the analyst, and the analyst's subjectivity is treated as a resource that must be made explicit rather than removed.
\nHow big was the study? The evaluation involved 12 paired analysts using the Reflexis system for collaborative reflexive thematic analysis (Ye et al., CHI 2026).
\n\nLast updated: July 6, 2026. This article is an independent editorial summary of third-party research (Ye et al., \"Reflexis,\" CHI 2026); it is not affiliated with or endorsed by the authors. Claims and figures are drawn from the cited source.
","related":[{"slug":"synthetic-survey-panels-marginal-fidelity-2026","title":"Do Synthetic Survey Panels Simulate People? A 2026 Test","description":"Do synthetic survey panels simulate real respondents? A 2026 test finds matching published averages is not proof: a no-persona prompt beat panels 9 of 9 times.","category":"Methodology","date":"2026-10-22","readTime":"7 min read","author":"Qualitati Research Team"},{"slug":"llm-as-judge-qualitative-interpretation-2026","title":"Can LLM-as-Judge Rate Qualitative Interpretations? A 2026 Test","description":"Can LLM-as-judge replace human review of qualitative interpretations? A 2026 study of 3,560 AI interview readings finds it screens models but misses nuance.","category":"Methodology","date":"2026-10-21","readTime":"7 min read","author":"Qualitati Research Team"},{"slug":"rehearse-interview-protocol-simulated-participants-2026","title":"Can You Rehearse Interviews with AI Participants? A 2026 Study","description":"Can you rehearse an interview guide with AI-simulated participants? A 2026 study of 21 researchers found real protocol revisions and three hidden risks.","category":"Methodology","date":"2026-10-21","readTime":"7 min read","author":"Qualitati Research Team"}]}