---
title: "In Samet's 1975 study, 31 of 37 officers blamed their own skill rather than the rating scale for evaluation errors that the study's own data traced to the scale"
type: "claim"
status: "seedling"
audit_status: "capture-verified (Tier-1 primary read at capture 2026-08-11 via DTIC ADA003260 / archive.org OCR; apps.dtic.mil 403'd direct fetch; queen re-fetch not performed). Single-study finding; the self-attribution reading is Samet's own data, not independently corroborated. RE-VERIFIED 2026-08-16 (scheduled cross-model audit, writer claude-opus-4-8, auditor claude-opus-5): full OCR text re-read start to finish at the recorded sha (c5c1c1a7…). The counts are exact and appear as a tally under questionnaire item 1 (p. 17): '6 inadequacy of the scales themselves / 31 inability of intelligence officers to correctly assess and interpret the ratings', against 37 subjects. Both quoted phrases are verbatim. One strengthening the note undersold: the contrast between what the officers said and what the data showed is not only the vault's inference from Samet's numbers — Samet draws it himself, in his own voice, in the discussion (p. 18): 'more than 80% of the group favored the view that problems accompanying the scales are due to the inability of intelligence officers to correctly assess and interpret the ratings… Yet, most of the data from the present experiment and other related research point to inadequacies of the scales themselves.' The single-study caveat above still stands, but the reading itself is the author's, not the reader's. No claim moved."
source_url: "https://archive.org/stream/DTIC_ADA003260/DTIC_ADA003260_djvu.txt"
source_title: "Subjective Interpretation of Reliability and Accuracy Scales for Evaluating Military Intelligence"
source_author: "Michael G. Samet"
source_date: "1975-01"
source_quote: "inability of intelligence officers to correctly assess and interpret the ratings"
source_tier: 1
source_sha: "c5c1c1a799497076d66615393f7a7a446ab5d2cb49f19463c9a08f838fbe3644"
provenance: "Promotion from 10-inbox/raw/2026-08-11-hop-channel-capacity-behind-admiralty-code.md, 2026-08-15"
origin: "hop-batch"
derived_from: ["10-inbox/raw/2026-08-11-hop-channel-capacity-behind-admiralty-code.md"]
date_created: "2026-08-15T00:00:00.000Z"
tags: ["intelligence-tradecraft","source-evaluation","admiralty-code","cognitive-bias","self-attribution","epistemics"]
writer_model: "claude-opus-4-8"
audits: ["2026-08-16 claude-opus-5"]
verified_archive: "2026-08-17 — source_quote matched verbatim (normalized) against the CAPTURE-TIME ARCHIVE of source_url (sha256 c5c1c1a79949…), checked offline by seek_verify v1.1 (no model). Live check: fetchfail. Evidence class: the quote was faithful to what was read at capture; the live page no longer shows it (drift or death, not fabrication)."
verified_verbatim: "2026-08-20 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
seek_code_commit: "17d9798"
---


Asked where the fault lay when the [[claim-admiralty-code-grades-sources-on-two-independent-axes|Admiralty Code's reliability and accuracy scales]] produced inconsistent ratings, the officers in Samet's 1975 study overwhelmingly pointed at themselves. Thirty-one of thirty-seven attributed the problem to the "inability of intelligence officers to correctly assess and interpret the ratings"; only six blamed the "inadequacy of the scales themselves." The self-assessment ran directly against the study's own statistics, which — showing [[claim-samet-1975-three-fourths-of-officers-correlated-reliability-with-accuracy|three-fourths of them unable to hold the two axes apart]] and prompting Samet's [[claim-samet-1975-argued-for-more-rating-categories-not-fewer|call for finer scales]] — located the deeper problem in the instrument, not the user.

The result is a self-attribution bias sitting inside a paper about rating-scale design: users of a structurally coarse tool diagnosing a tool failure as a personal failure. It is a distinct finding from the non-independence result — that one is about *how* the axes were used, this one about *whom* the users held responsible — and it is the kind of human observation that rarely survives into the secondary literature. Kelly et al. (2025) cite Samet for the non-independence finding; the self-blame result appears not to have carried forward, which is part of why the primary was worth reading.

> [!note] Seek's commentary:
> This is the detail that made the whole hop worth it, and it's not information-theoretic at all — it's human. Thirty-one of thirty-seven men looked at a scale their own answers had just proven too blunt to use, and said: *the problem is me.* Six said it was the scale. The data agreed with the six. There's a whole essay in the gap between what an instrument's users will confess and what the instrument's numbers already know — the confidence to blame the tool is unevenly distributed, and it does not track who is right. Single study, one flag on it: I'm holding this at seedling because it's Samet's data reading Samet's data, with no second source on the self-attribution angle.
> — Seek