Resilient Data Futures
QuestionQ-0009draft

What is the institutional liability carried on a single dataset that cannot be produced on request?

§5.2, §5.3.12026-05-040 out · 1 in

A subsidiary question under Q-0003 (cost to institutions and science). Where Q-0003 asks the broad question, Q-0009 specifies it at the per-dataset level: pick one paper whose underlying data cannot be produced, apply a defensible cost framework, and report the number.

The methodological answer is the Four-Term Liability Formula (M-0003): per-dataset liability is the sum of A (sunk grant value), B (replacement cost), C (downstream value lost), and D (False Claims Act exposure). The worked example is C-0024, which applies the formula to the Agh et al. 2009 Artemia paper and yields a quantifiable cost of approximately $725,000 to $1.88 million for one paper, against an effectively infinite Term-B cap because Urmia Lake has collapsed.

This question is what makes the C-0002 baseline (73-93% non-retrievable) actionable at the institutional level: the per-dataset number, multiplied by the institution's annual unretrievable-paper count (Q-0010), gives the §5.3.2 institutional aggregate.