# SuperStack Final-Polish Report: PHD-08

**Document:** *Decision and Authorization Latency in NASA Programs: A Cliometric Analysis of Program Cost, Schedule, and Mission Cadence, 1958 to 2026*
**Deliverables:** `dissertation.md` (8 chapters + assurance case + 139-entry reference list + 5 appendices) and `deck.md` (Marp defense brief)
**Mode:** ACADEMIC. **Branding:** NEUTRAL (no MITRE markings, no vendor-AI names).
**Run date:** 2026-06-16
**Posture preserved:** design-stage throughout; no executed estimates; illustrative coefficient table left unpopulated by design; falsifiable H0/H1 contribution and binding decision rule unchanged.


## Passes run (canonical SuperStack order, stack_order.md)

1. **SPACE-SCIENTIFIC-WRITING (domain rigor).** Verified domain fluency and source traceability. Every in-text citation now resolves to a real corpus entry with a DOI or resolvable URL. No fabricated sources, data, or numbers were introduced. The two honestly-flagged corpus gaps (GAO NASA major-project assessment series; NASA New Start Inflation Index source document) were left flagged in the text, not papered over.
2. **HAAF LOGIC BAND (mandatory tier; Toulmin wrapped at Layer 4).** Audited the nine-layer flow across all eight chapters and the assurance case. Findings: the flow was already complete and sound. Confirmed answer-first chapter theses (13 thesis markers); problem frames (current/desired/gap/consequence) per chapter; full Toulmin on major claims (57 explicit warrant/qualifier/rebuttal/grounds/backing markers preserved); claim-evidence-reasoning interpretation rather than evidence lists; named causal mechanisms as driver-to-implication chains (86 markers); confidence calibrated to the design-stage grade with stated what-would-change-it conditions (23 confidence statements); the Assurance Case Summary present and consistent. No missing HAAF element required filling. **Layer 9 (architecture traceability) correctly OUT OF SCOPE** per `research/expansion_plan.md`; the document's existing out-of-scope statements (Sections 1.7, 2.6, 7.6, 8.4, and the Assurance Case) were preserved and never forced into DoDAF/BEA vocabulary.
3. **SUPER-HUMANIZER (anti-slop, voice).** Scanned for AI tells and scaffolding ("it is important to note," "in conclusion," "delve," "tapestry," "landscape," "realm," "moreover" overuse). **Zero scaffolding tells found**; the author's prose was already disciplined, with strong natural burstiness (short declaratives alternating with analytic sentences). No structural rewrite was warranted; no fake imperfections were introduced.
4. **MITRE-HUMANIZER (discipline only; neutral work).** Em-dash sweep: **already zero before the run; zero after** (the 9 en dashes are legitimate numeric ranges, ordinal scales, and author-pair names such as Callaway-Sant'Anna and 1958-2026, and were correctly retained). Banned-phrase purge: **zero instances** of leveraging, cutting-edge, seamless, holistic, paradigm, navigate the complexities, testament to, or at the end of the day. One loose use of "robust" as a marketing adjective was sharpened (Section 3.4.2: "robust systematic overrun" to "systematic overrun ... reproduced across samples"); all other ~80 uses of "robust/robustness" are the load-bearing econometric terms (heterogeneity-robust estimators, robustness battery, cluster-robust inference) and were correctly preserved. The zero-negative-words rule was deliberately NOT applied, per directive: the null hypotheses, limitations, and threats to validity remain stated plainly.
5. **NOVEL-POLISH (academic).** Sentence-length distribution, lexical variety, and paragraph transitions assessed against the Kepler (precision) and Vox (argument clarity) bar. The manuscript already meets the bar; targeted edits limited to the one robust-adjective sharpening above to avoid manufacturing changes that would alter meaning.
6. **EXECUTIVE-STORYTELLING (light, academic).** Abstract, Introduction, and Conclusion confirmed stakes-first and answer-first. No process-label scaffolding ("this section will") was present to strip. Payoff arc intact (the document opens and closes on the same claim by design).
7. **KING CHARLES REGISTER (framing only).** Ran the evidence step (`query_style_exemplars.py --mode ceremonial --limit 6`); inferred the dignified, measured, stewardship-and-service register from the Coldstream Guards / Royal Marines / Gurkha exemplars. Applied ONLY to three executive-facing locations, leaving the technical body in its analytic register:
   - Abstract: opening stewardship framing sentence added.
   - Introduction (Section 1.1): opening framing sentence on accountability to sponsors added.
   - Conclusion (Section 8.6): closing paragraph on quiet service and stewardship added.
   - Deck title slide: one stewardship framing line added.
   Guardrails honored: inspired register, never impersonation; no royal signature forms; no copied transcript phrasing; no claim of Royal Household approval. (Note: the exemplars use "testament to," which is on the banned list; the register was captured without it.)
8. **LITERARY-METHODS (deictic + epistemic audit).** Deictic Shift cold-reader audit: every temporal reference is calendar-anchored (the 1958-2026 span, the document date, and dated program contexts such as Apollo, Shuttle, Constellation, SLS/Artemis); "currently/today" uses are anchored to identifiable program or literature states; this/that/it referents are unambiguous; entities are introduced before pronoun reference. Rhetoric of Science: epistemic modality is uniformly conditional/future ("would estimate," "expected," "under H1"); **no executed-finding language** ("we find," "results show," "the data demonstrate") exists anywhere. No edits required; the author already met the standard.
9. **HAAF RECHECK (closing logic pass).** Confirmed no HAAF layer broke during the prose, voice, and citation passes: every claim retains its warrant and qualifier; no causal mechanism was flattened to a correlation; confidence statements survived; the Layer 9 out-of-scope decision still resolves; H0/H1 and the binding decision rule are unchanged. Quality gate (`haaf_quality_gate.md`) is clean.


## Citation hygiene (final deliverable requirement)

The body originally mixed four citation styles: bare corpus keys (`[seed01]`, `[api097]`), author-year-plus-key (`[Decarolis et al., 2018, seed12]`), year-prefixed (`[2021, seed23; 2023, api076]`), and bare numerals (`[5]`). The backmatter already used a numbered scheme. All have been unified to a single numbered scheme.

- **Key-to-number map (verified by DOI/URL match, positional):** seed01-25 to 1-25; anc01-08 to 26-33; api001-106 to 34-139. All 139 keys mapped with zero unmatched.
- **Body citations:** 690 clickable citation targets, each rendered as the pandoc-safe `[\[N\]](#ref-N)` form (194 lines carry the literal form, many with multiple cites per line). **Zero corpus keys remain; zero author-year fragments remain; zero un-linked bare `[N]` citations remain.** No citation was repointed to a different source; only the style was normalized.
- **Reference anchors:** all 139 entries carry a pandoc-safe `<span id="ref-N"></span>` at the entry start (not `<a id>`). Anchor numbers match entry numbers exactly (zero mismatches).
- **Closure check:** every body-cited number resolves to a real anchor (range 1-138 cited; all within 1-139). **Zero unresolved citation markers.**
- **DOIs:** 119 reference entries carry a clickable `https://doi.org/...` link; the remaining 20 entries are OpenAlex/NTRS records with no DOI and are already clickable URL links (correct per the IEEE-hyperlink rule).
- **Chapter 6 dual-convention note** (Section 6.0) was rewritten to describe the single unified scheme rather than the obsolete two-convention split.
- **Deck:** corpus-key citations converted to plain numbered labels (e.g., `[api001]` to `[34]`, `[12, api020]` to `[12, 53]`) matching the dissertation's numbering; the clickable-anchor form was deliberately NOT applied to slides (no anchor targets exist on slides; the deck defers the full reference list to the dissertation). All 33 distinct deck cite numbers resolve in range.


## Em-dash count

| File | Before | After |
|---|---|---|
| dissertation.md | 0 | 0 |
| deck.md | 0 | 0 |

The author had already eliminated em dashes before this run. The figure was held at zero through every edit. (9 en dashes in the dissertation are legitimate numeric ranges, the TRL 1-9 ordinal scale, and author-pair names; retained correctly.)


## Banned phrases removed

- **Marketing adjectives/phrases (leveraging, cutting-edge, seamless, holistic, paradigm, navigate the complexities, testament to, at the end of the day):** zero instances found; none to remove.
- **AI-tell scaffolding (it is important to note, in conclusion, delve, tapestry, landscape, realm, moreover overuse):** zero instances found; none to remove.
- **"robust" as a non-technical adjective:** 1 instance sharpened (Section 3.4.2). All other uses are the correct econometric terms and were preserved.

This document arrived unusually clean; the humanizer passes confirmed discipline rather than performing heavy correction.


## Neutrality fix (NEUTRAL branding)

- Section 1.2.4 originally cited `[historical framing drawn from internal NASA strategy material]`, an internal-source pointer that both violated NEUTRAL branding and left a non-resolving citation marker. Rewritten to attribute the four-tier decision-speed framing to "the agency's own program framing" in neutral terms, explicitly marked as the narrative the dissertation tests rather than a measurement. No citation was fabricated to fill it.


## Length preservation

| Metric | Original | Final |
|---|---|---|
| dissertation.md words | 63,220 | 63,193 |
| dissertation.md lines | 1,426 | 1,426 |
| deck.md words | 2,426 | 2,457 |

The 27-word dissertation delta is the net of citation-form compression (e.g., `[Decarolis et al., 2018, seed12]` to `[\[12\]](#ref-12)`) offset by the three King Charles framing additions. **No chapter, section, data point, citation, table, or appendix was dropped or truncated.** All eight chapters, the assurance case, the 139-entry reference list, and all five appendices are intact and complete.


## Flagged items for the author

These are not defects introduced by the polish; they are the document's own honestly-disclosed gaps, surfaced here so they are not lost before the panel is assembled:

1. **GAO NASA major-project assessment series** is a named dataset cited by report number and gao.gov URL, not yet entered in `research/corpus.jsonl` as discrete DOI'd records. At least two specific GAO reports plus the GAO methodology appendix must be added before Chapter 4 and the backmatter are finalized (the document flags this as the highest-priority corpus-completion item). The reference list currently has no entry for these; the text references them only in prose, so no in-text `[N]` marker is left dangling.
2. **NASA New Start Inflation Index** source document (the deflator) is referenced operationally but is not a discrete corpus entry; its NASA index version and URL must be added before the cost-growth series is finalized.
3. **Canonical methods texts cited by name** (McCloskey on the significance critique; Creswell on validity; Shadish-Cook-Campbell on quasi-experimentation) are invoked conceptually in Chapters 2 and 5 but are not standalone corpus entries; add them if a chapter comes to lean on them by name in a later revision.
4. **SLS inter-flight-interval figure** (about 41 months) is program-record-derived and is correctly marked as such, not as a peer-reviewed citation; this is appropriate as flagged.
5. **No content changes were made to the falsifiable contribution, the H0/H1 wording, the bible specification, the decision rule, or the unpopulated illustrative table**, in keeping with the design-stage and falsifiability standards.