Draft Preliminary note on identifiability and PII review · click to expand

This appendix is being circulated to the committee as a draft. It contains material that has not yet completed a final review for personally identifiable information. Before the appendix is finalized for submission, every mention of a named individual will be reviewed against the taxonomy in §B.6.5: students named in instructor-produced materials will be anonymized unless explicit written consent for educational use is documented; named guest speakers will be retained as public professional identities with their professional context attached; Luma-platform workshop feedback will be reviewed for anonymization; and external journalism is retained as already published and consented.

Index
§B.6

Trustworthiness

Trustworthiness in analytic autoethnography is established through Anderson's analytic rigor criteria, through triangulation across artifact types and contexts, and through honest acknowledgment of the limits of the data and of my position as the researcher.

B.6.1 Anderson's analytic rigor as the trustworthiness frame

Section B.3 lays out Anderson's five criteria and evidences each against my corpus. The five criteria are themselves the trustworthiness frame for analytic autoethnography. A work is trustworthy as analytic autoethnography to the extent that it satisfies the criteria; a work that drifts toward evocative narrative without the criteria's analytic discipline is a different kind of work and is evaluated differently.

My corpus satisfies the criteria as evidenced in §B.3. Complete member researcher status is documented across every site of teaching practice (CV-1, CV-2, DK-3, DK-4, the HC series, the KN series, the KP series, WB-2026-03-03). Analytic reflexivity is most fully evidenced in the Iteration 1 Weekly Updates Prelim Document (WU-1.W01 through WU-1.W15) and is supplemented by my public-facing reflective writing across the cross-iteration channels. Narrative visibility of my self is present throughout the corpus, with first-person voice in the Research Impact Essay (RE), the newsletter (KN-EP series), the podcast (KP-EP series), and the webinar (WB-2026-03-03). Dialogue with informants beyond self is evidenced by twenty-three named student outputs in DK-1.W01, twenty-nine Luma responses (LF-3.R01 through LF-3.R29), the Aspen Public Radio coverage (AP-2024-05-16), the UW KidsTeam research data (KT-THEMES and the broader KT corpus), six student teach-out presentations (SP-2.*), and the documented guest-speaker rosters across both iterations. Commitment to theoretical analysis is satisfied by the three anchor concepts I develop in Chapter D (hallucination-as-pedagogy, compression-as-curriculum-maturation, multi-channel teaching practice), each tested against multiple independent sources.

B.6.2 Triangulation across artifact types and contexts

The strongest single trustworthiness move I make is triangulation: anchoring each substantive claim in multiple independent artifacts of different types from different points in the timeline.

My hallucination-as-pedagogy claim is anchored in four independent sources: my contemporaneous reflective journal (WU-1.W01-Q1, January 2024), an independent K-12 research data set (KT-THEMES-C5, July 2024), my public-facing reflective writing (KN-EP1-Q1, April 2025), and my recorded teaching delivery (TR-4.D1, September 2025). These four sources span roughly twenty months and three distinct types of artifact: instructor reflection, independent learner observation, public-facing writing, and recorded delivery. The hallucination phenomenon as pedagogical opportunity surfaces in all four.

My compression-as-curriculum-maturation claim is anchored in the four iteration artifact sets. Iteration 1 (CV-1, DK-1.FG, DK-1.W01, FP-1, WU-1 series) shows the fifteen-week structure with the four-theme architecture explicit. Iteration 2 (CV-2, SY-2, DK-2 series) shows the same architecture in a second fifteen-week iteration, with documented tool turnover and module reshuffles. Iteration 3 (DK-3, LF-3, LR-3) shows the architecture surviving compression to five days. Iteration 4 (DK-4, TR-4.D1 through TR-4.D5) shows the same five-day compression carrying through to a second cohort.

My multi-channel teaching practice claim is anchored across the eight documented channels described in §B.3.5 and developed in §D.4. Each channel is a separate corpus with its own artifact provenance.

Triangulation is the practical instrument by which analytic autoethnography moves from instructor's stray observation to defensible theoretical claim.

B.6.3 Honest acknowledgment of the limits of the data

I name two limits of my data that the methodology chapter must surface rather than paper over.

Asymmetry in contemporaneous reflective material. Iteration 1 is uniquely well-served by the fifteen-week Weekly Updates Prelim Document. Iterations 2, 3, and 4 do not have a structured contemporaneous reflective counterpart. The cross-iteration reflective channels (KN-EP series across the 2025 iterations, KP-EP series in May 2025, WB-2026-03-03 in March 2026) carry reflective writing across the later iterations within an acknowledged retrospective-public frame, and analytic autoethnography permits this configuration when the frame is named.

Aspen Public Radio quoted at source-level. The Aspen Public Radio article (AP-2024-05-16) is held as a source-level reference in the evidence-table. The article's named quotes of me and of my Iteration 1 student Ashley Stafford are part of the corpus by reference; the dissertation cites the article at source-level rather than at sentence-level.

These limits are not invalidating. They are common in analytic autoethnographic work that surveys a multi-year practitioner-pioneer practice across multiple delivery channels. Naming them is what trustworthy autoethnography requires.

B.6.4 Selection of the corpus itself

The corpus is a curated artifact. I chose what to include, in what level of detail, and with what framing. Selection is itself a methodological act that requires acknowledgment.

What I included: every artifact I produced as the instructor of the four iterations that I could locate and ingest (Canvas LMS exports, slide decks, syllabi, the Weekly Updates Prelim Document, the Final Project assignment); the publicly-available cross-iteration material I authored (Research Impact Essay, Keep Up Newsletter, Keep Up Podcast, CU RMACC webinar); learner-facing material that captures student or learner voice (named student outputs in DK-1.W01, the Iteration 2 student teach-out presentations, the Iteration 3 Luma feedback corpus); collaborative research material relevant to the K-12 strand (the UW KidsTeam corpus); and institutional context (the AI-IRT Seed Grant proposal, the ENED Preliminary Exam Part 2).

What I included only partially: the Aspen Public Radio article cited at source-level rather than sentence-level (§B.6.3), and feedback channels for Iterations 1 and 2 that would parallel the Luma corpus for Iteration 3 (these were not collected through a structured instrument because the semester iterations did not run through Luma).

A different researcher with access to the same setting might have included different material. This is the honest position selection bias requires: not neutrality, which is unavailable, but transparency about what was included and what was not.

B.6.5 Identifiability and PII review scope

The autoethnographic appendix differs from the main document in the kinds of identifiable mention it contains. The main document analyzes the instructor-authored curriculum and treats course evaluation feedback as anonymized triggers for revision (main §2.6.5). The appendix, by contrast, is autoethnographic: it draws on the instructor's reflexive position, names participants in the four iterations where they appear in the artifact corpus, and quotes from public-facing materials (the Keep Up Newsletter, the Keep Up Podcast, the CU RMACC webinar, the Aspen Public Radio article). This wider corpus carries identifiable mentions that need to be accounted for separately.

Four categories of identifiable mention appear in this appendix, each handled differently:

Handling of these categories has been and will remain consistent with the Family Educational Rights and Privacy Act (FERPA, 20 U.S.C. § 1232g; 34 CFR Part 99) and the University of Colorado Boulder's policies on student records. The appendix's autoethnographic framing follows the standards of analytic autoethnography (Anderson 2006), in which the instructor's reflexive position and named relationships in the practice are part of the analytic subject. As with the main document's human-subjects scope (main §2.6.5), the analytic claims of this appendix do not require IRB-reviewed measurement of student outcomes.

B.6.6 Honest acknowledgment of my position as the researcher

I am the subject of the work and the researcher of the work. This is the definitional condition of autoethnography, and it carries a definitional limitation: my interpretive choices are not neutral. I have a stake in my work succeeding, in my findings landing as contributions, and in the dissertation being awarded. Any reader of an analytic autoethnography must take this stake into account.

I respond to this limitation through three moves. First, I source every substantive claim to artifacts produced at the time of the events described, rather than to retrospective reconstructions, wherever the artifacts allow. The Weekly Updates Prelim Document is the strongest such anchor for Iteration 1. Second, I subject every anchor concept to multi-source testing as described in §B.5.3 and §B.6.2, rather than committing to a concept on the strength of a single observation. Third, I expose my analytic process and my data to my dissertation chair, my second committee member (both of whom are co-PIs on the seed grant that funded the work and both of whom have been present in the setting as guest lecturers), and my full committee, so that interpretive choices that depend on my position can be tested against readers who share knowledge of the setting but do not share my investment.

This is the trustworthiness posture analytic autoethnography asks for: insider position acknowledged, multi-source evidence assembled, chair and committee engaged as interlocutors. I make no claim to neutrality, and analytic autoethnography does not require it.

Index