# REPORT — pinillossimpsonndexperimentalevidencesupporting

## What I processed

- Target PDF: `inbox/Pinillos&Simpson2014..pdf`
- Paper key: `pinillossimpsonndexperimentalevidencesupporting`

## Pipeline outputs used (in priority order)

- `papers/pinillossimpsonndexperimentalevidencesupporting/out/bundle.json` (status/index)
- `papers/pinillossimpsonndexperimentalevidencesupporting/out/fulltext.md` and `papers/pinillossimpsonndexperimentalevidencesupporting/out/text.txt` (context + most stats)
- `papers/pinillossimpsonndexperimentalevidencesupporting/out/tei.xml` (recovery of truncated Figure 1.1 condition stats)

## Manual interventions / issues

### Used the raw PDF for page numbers and to recover truncated numeric stats

- `out/fulltext.md`/`out/text.txt` truncate some Experiment 2 (Airplane) condition statistics mid-sentence.
- I used `pdfgrep -n` and `pdftotext -f/-l` on `papers/pinillossimpsonndexperimentalevidencesupporting/pdf/paper.pdf` to:
  - recover the full condition means/SDs/Ns for Experiment 2 (PDF page 13), and
  - attach PDF page numbers to analysis-critical provenance quotes.

## Effect size computations (Quarto; reproducible)

- Computed all `effect_size.d` and `effect_size.v` from group means/SDs/Ns (between-groups SMD) in:
  - `papers/pinillossimpsonndexperimentalevidencesupporting/analysis/effect_sizes.qmd`
  - rendered to `papers/pinillossimpsonndexperimentalevidencesupporting/analysis/effect_sizes.html`

## Uncertainties / questions for human review (fill these in)

### Q1 — Full bibliographic citation (year/editors/pages/DOI)

- The PDF text includes the chapter title and authors, and page headers suggest the volume title “Advances in Experimental Epistemology”, but I did not find:
  - publication year for this chapter,
  - editors/publisher details,
  - chapter page range in the volume, or
  - a DOI.
- I set `paper.year: null`, `paper.published: null`, and used “(n.d.)” in `paper.citation`.
- **Please provide** the complete reference (and confirm whether it should be coded as published/peer-reviewed).

### RESOLVED — Study 2 (Air/Bridge) vignette wording

- The appendix was located in `out/external/pinillos IRI paper with shawn october 2014.pdf`.
- It contains the Air appendix material and the full Bridge low/high vignettes plus the level-of-agreement prompts.
- Study 2 Air and Bridge effects have therefore been changed to `needs_review: false`; the numerical effect sizes were unchanged.

### RESOLVED BY CODING DECISION — Modal-reading Typo experiment

- Study 6 reports condition Ns and descriptive statistics for the Typo “Knows vs Hopes” modal-reading test, but does not explicitly report participant recruitment or compensation.
- Coding decision: recruitment/compensation were coded as `mTurk`/`money` by continuity with the surrounding Pinillos & Simpson 2014 studies in the same paper. The compensation amount remains unreported.
- The located Pinillos & Simpson appendix does not contain the full Typo vignette wording. Study 6 says it used the low-stakes and high-stakes Typo vignettes, so the full vignette wording was copied from the local Pinillos 2012 extraction (`papers/pinillos2012knowledgeexperimentspracitcal`).
- The stale `needs_review` flag for Study 6 was cleared in the source YAML. No final-analysis CSV update was made here.

## Notes

- I extracted stakes effects from:
  - Study 1 Experiment 1 (Water purifier),
  - Study 1 Experiment 2 (Airplane; split by probability for matched low/high stakes contrasts),
  - Study 2 probes (Coin/Air/Bridge; one study entry per probe due to different scales and separate Ns),
  - and the Typo modal-reading test (stakes effect on the Knows evidence-seeking question).
- I did not extract Study 3 (RKP/ACTION) because it does not add a new high-vs-low stakes contrast effect in the stakes-meta schema format (it analyzes normative questions/correlations rather than a direct stakes manipulation effect per scenario).