Epistemic Agency in Self-Tracking: AI Qualitative Analysis

Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents. According to the Frontiers article (18 Aug 2026), an agency-based review of the "quantified self" literature shows that self-tracking spans from low-agency data collection to high-agency personal science, and the distinction hinges on "who owns the question". This post explains the Frontiers review's methods and numbers (search April–May 2025), extracts concrete statistics (721 records, 241 full papers, 46 self-tracking studies), and shows how AI-enabled qualitative research workflows can reproduce and extend that agency-based classification to inform researchers and product teams.

Key Takeaways

According to the Frontiers article (18 Aug 2026), a conceptual review of "quantified self" literature used a four-phase framework to measure "epistemic agency" and found 46 self-tracking studies that ranged from low to high agency; see Frontiers.

A literature search in Web of Science and PubMed in April–May 2025 returned 721 records, of which 241 full papers were reviewed, according to the Frontiers review (18 Aug 2026).
Of the 241 full papers, 46 articles used self-tracking as the data-collection method and were classified for agency: 17 low-agency, 10 mixed-agency, and 19 high-agency (Frontiers, 18 Aug 2026).
The Frontiers authors define epistemic agency across four phases: Questioning, Observing, Reasoning, Discovering, and use the phrase "who owns the question" to mark the shift from measurement to personal science (Sara Riggare et al., Frontiers, 18 Aug 2026).

What happened and how the Frontiers review measured agency

Answer: The Frontiers review applied a four-phase personal science framework to characterize epistemic agency in "quantified self" studies, then categorized eligible papers into three agency levels, as reported in the Frontiers article (18 Aug 2026).

According to the Frontiers methods section, the authors searched Web of Science and PubMed in April–May 2025 for the term "quantified self" and retrieved 721 unique records, of which 241 full papers were retained after screening (Frontiers, 18 Aug 2026).

According to the Frontiers analysis, only 46 of the 241 papers used self-tracking as the data-collection method and were therefore eligible for agency coding; the authors performed independent dual-author review for classification (Frontiers, 18 Aug 2026).

According to the Frontiers analytical framework, the four phases are Questioning, Observing, Reasoning, and Discovering, and articles were labeled low, mixed, or high agency by whether the individual performed each phase (Frontiers, 18 Aug 2026).

The Frontiers authors conclude that the decisive distinction between mere measurement and personal science is "who owns the question", a phrase emphasized in their discussion and applied to classify the 46 self-tracking studies (Frontiers, 18 Aug 2026).

Findings snapshot

Date	Metric	Value	Implication
April–May 2025	Search results (Web of Science + PubMed)	721 records	Corpus for screening (Frontiers, 18 Aug 2026)
April–May 2025	Full papers retained after screening	241 papers	Primary corpus for thematic categorization (Frontiers, 18 Aug 2026)
April–May 2025	Papers using self-tracking as data collection	46 papers	Eligible for agency-level coding using the four-phase framework (Frontiers, 18 Aug 2026)
Aug 18, 2026	Agency classification among 46 studies	17 low, 10 mixed, 19 high	Demonstrates full spectrum from sensor-platform studies to personal science (Frontiers, 18 Aug 2026)

Implications for qualitative researchers and UX teams

Answer: The Frontiers review implies that research teams must distinguish between data provision and knowledge co-creation when designing self-tracking studies (Frontiers, 18 Aug 2026).

According to the Frontiers discussion, low-agency designs typically treat participants as "sensor platforms" and prioritize data sharing, which risks erasing participant interpretation and lived context (Frontiers, 18 Aug 2026).

According to the Frontiers results, high-agency personal science studies often show the self-tracker as co-author and emphasize sharing of methods and learnings rather than raw data alone (Frontiers, 18 Aug 2026).

Practical implication for UX teams: design interfaces and export tools that enable participants to formulate questions, iterate measurement methods, and export their methods and visualizations, not only summary metrics (Frontiers, 18 Aug 2026).

Practical implication for qualitative researchers: annotate and preserve participant-generated reasoning (qualitative notes, context) alongside sensor streams to enable valid N-of-1 and cross-case synthesis (Frontiers, 18 Aug 2026).

How Evidano helps: map problems in agency to AI-enabled qualitative features

Problem: Participants treated as data sources, not co-researchers

Solution: Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.

How it helps: According to platform capabilities, Evidano ingests transcripts and logs so teams can tag which phases (Questioning, Observing, Reasoning, Discovering) participants performed and surface missing agency phases using automated thematic and frequency analysis.

Relevant feature: use Evidano features to combine transcripts, exported device logs, and survey responses into a single analysis workspace.

Problem: Low transparency about methods and no way to share method metadata

Solution: Evidano extracts methodological text and exports structured method summaries suitable for inclusion in appendices or community repositories.

How it helps: According to Evidano's document ingestion, tagging, and export features, researchers can automatically generate a reproducible method trace for each participant so "who owns the question" is documented.

Problem: Slow synthesis of mixed quantitative and qualitative N-of-1 data

Solution: Evidano performs thematic, content, frequency, and cross-segment analyses across transcripts and time-series annotations.

How it helps: Evidano's AI-assisted coding speeds Reasoning stage replication and supports cross-case synthesis that respects individual Discovering outputs while extracting generalizable themes, matching the four-phase needs identified by the Frontiers review (Frontiers, 18 Aug 2026).

FAQ: epistemic agency in self-tracking

What is epistemic agency in self-tracking?

Answer: Epistemic agency in self-tracking is the capacity of an individual to act as a producer of knowledge, posing questions, collecting and interpreting evidence, and articulating discoveries, as defined in the Frontiers review (18 Aug 2026).

Support: According to Catherine Elgin (quoted in the Frontiers article), "Epistemic agents should think of themselves as, and act as, legislating members of a realm of epistemic ends: they make the rules, devise the methods, and set the standards that bind them."

How did the Frontiers review operationalize agency?

Answer: The Frontiers review operationalized agency using a four-phase framework: Questioning, Observing, Reasoning, Discovering (Frontiers, 18 Aug 2026).

Support: According to the Frontiers methods, each phase was coded as performed by the individual or by researchers, and studies were conservatively assigned to low, mixed, or high agency levels based on explicit textual evidence (Frontiers, 18 Aug 2026).

Which concrete numbers should I cite from the Frontiers review?

Answer: Cite that the search in April–May 2025 returned 721 records, 241 full papers were reviewed, and 46 papers used self-tracking as data collection, split into 17 low, 10 mixed, and 19 high agency (Frontiers, 18 Aug 2026).

Support: Those counts are reported in the Frontiers Results and Methods sections and are suitable for methodological comparisons or replication plans (Frontiers, 18 Aug 2026).

Can AI reproduce the Frontiers agency classification?

Answer: Yes, AI-enabled qualitative workflows can reproduce and scale the Frontiers agency classification by tagging textual evidence of who posed questions, who analyzed data, and who disseminated findings, as suggested by the review's coding procedures (Frontiers, 18 Aug 2026).

Support: According to the Frontiers paper, agency coding relied on explicit textual cues such as co-authorship by the self-tracker and descriptions of self-directed Reasoning and Discovering; these cues are extractable from publications and participant logs with NLP and thematic analysis tools.

Conclusion & Next Steps

The Frontiers review (18 Aug 2026) shows that self-tracking in health sits on a spectrum from low-agency measurement to high-agency personal science, and the decisive criterion is "who owns the question" (Frontiers, 18 Aug 2026).

AI-enabled qualitative research can reproduce the review's four-phase coding at scale by extracting Questioning, Observing, Reasoning, and Discovering signals from transcripts, device logs, and papers, enabling design and ethical reflection targeted at increasing epistemic agency.

If you run self-tracking studies or design consumer health experiences, apply an agency audit to distinguish participants-as-sensors from participants-as-co-researchers, and preserve method metadata so discoveries are reproducible.

Get started: Try Evidano for free to import transcripts and device logs, run thematic and cross-segment analyses, and document who owned each phase of inquiry in your projects.

Topics

epistemic agency in self-tracking
personal science analysis
AI qualitative research
quantified self qualitative

Keep reading

Browse all articles