Evidence sweep · September 2026
Each month the registry sweeps the published literature for new psychometric evidence on the instruments it tracks, and files what it finds through the public change process. This is the September 2026 sweep, the first filed through the automated pipeline. Headline: four relevant new publications confirmed, one licence question investigated and resolved, seventeen of the twenty-one watchlist instruments silent. No grades moved, which is the expected outcome of a normal month.
What was found
- WHO-5 Well-Being Index: an Arabic validation in university students in the UAE (Frontiers in Education, 5 August 2026) reporting internal consistency of 0.833, composite reliability 0.848, and gender invariance to the scalar level. A student rather than working population, so recorded as indirect evidence for workplace use.
- GHQ-12: a factor-structure evaluation in Chinese university students (Frontiers in Public Health, 28 July 2026) retaining the two-factor model on interpretability grounds while noting a better-fitting bifactor model, feeding the structural debate the record already marks as contested.
- Oldenburg Burnout Inventory: a Malayalam adaptation in secondary school teachers (IJCMPH, August 2026): content validity strong, total-scale consistency 0.84, but the disengagement subscale at 0.61, consistent with the record's note of uneven subscale reliability in translations.
- WEMWBS: a Chinese validation in cardiac patients (15 July 2026), caught from the gap between the registry's generation and the sweep window; clinical rather than workplace population, coefficients awaiting extraction.
The licence question
The sweep flagged a possible change to the WEMWBS licence terms at Warwick. Direct inspection of the live page resolved it: the page was revised on 18 August 2026, and the NHS charging bands the registry records are unchanged; the revision published fuller pricing ladders, including an extended commercial tier. Archived copies of the licence page and its new NHS sub-page were captured on 1 September, and the record's verification date will be refreshed at the next version.
Stated limits of this sweep
The sweep counted only studies it located and corroborated; near-misses and items whose publication dates could not be pinned are logged for the next pass rather than counted. Non-indexed and non-English journals, instrument manuals and conference literature are known blind spots. Silence on an instrument means nothing corroborated was found in the window, not that nothing exists.
Registry changes applied with this sweep are listed in the changelog. The full technical report, with per-instrument search coverage and every citation, ships with the public repository.