Unpublished cell disposition log
StatMedAnalysis | 7 November 2026 | Use aggregate or hypothetical figures only.

Make one row for each expected but unusable cell. Save publisher symbols as text; do not convert them to numeric zero.
1. State the target grid (product × geography × variable × period) and the expected count.
2. Compare returned keys with the expected grid. Distinguish an absent row, an annotated value and an observed numeric zero.
3. Apply a prespecified disposition. Do not infer suppression if the reason is unknown.

Expected grid definition: ____________________  Count: ____  Retrieved unique keys: ____
Key | Release/table | Source text or absent row | Reason and evidence | Primary disposition | Alternative
___ | _____________ | _________________________ | ___________________ | ___________________ | ___________
___ | _____________ | _________________________ | ___________________ | ___________________ | ___________
Expected = usable + unusable + absent: ______ = ______ + ______ + ______
Fictional example: three expected cells; values 6, 12, absent. Usable = 2, unusable present = 0, absent = 1. Report coverage 2/3 and preserve the blank; imputing zero would change the observed-cell mean from 9 to 6.
Completion check: counts reconcile and readers can distinguish a missing geography from a published zero.