Internal design document, rendered for transparency. Operational sections (deployment, job recipes, credentials) have been removed. Dated 2026-09-11; may be amended by later work recorded in the changes log.

report-1-round1-independent

This report measures INDEPENDENT round-1 inter-coder reliability: Krippendorff's alpha between coder-1 and coder-2 labelling the same documents independently, with no adjudication folded in, on the eligible-1922 frame (the 2,000 round-1 gold documents minus the 78 platform-artifact rows ruled analysis-ineligible on 2026-08-02).

Construction. one unit per eligible document per axis, values = (coder-1 label, coder-2 label) exactly as submitted; no adjudication, no escape-label collapsing

Threshold. 0.70 (human-human) — point estimate against the recorded threshold; the 95% bootstrap CI is reported for context and never gates (docs/technical_blueprint.md:238)

Not this report. the post-adjudication residual (report-2-post-adjudication-residual) and machine-vs-gold (report-3-machine-vs-gold) are different numbers on different constructions — do not merge them

Frame. eligible-1922 — 1922 documents.

axis	n	alpha	95% CI	passed	reason
actor_alignment	1922	0.5051	0.4785 to 0.5314	False	below threshold: 0.5051 < 0.7
function	1922	0.5373	0.5082 to 0.5642	False	below threshold: 0.5373 < 0.7
theme	1922	0.6914	0.6690 to 0.7145	False	below threshold: 0.6914 < 0.7

axis	published	computed	abs diff	within 1e-3
actor_alignment	0.5051	0.505056	0.000044	True
function	0.5373	0.537316	0.000016	True
theme	0.6914	0.691422	0.000022	True

input	sha256
`FOR-PROFESSOR-2026-07-25/results/labels_coder1.csv`	`ac182d198bf122c7...`
`FOR-PROFESSOR-2026-07-25/results/labels_coder2.csv`	`ac18cbd7be1cb670...`
`01-08-requests/03-PI-ruling-3-artifact-rows/artifact-rows-78.csv`	`ca068865472bef84...`

seed reliability-frames-20260911, 2000 bootstrap resamples, generated 2026-09-11T04:58:12.268855+00:00.