The Map That Changed the Question: Why International Counts of Cybersecurity Education Depend on the Observation System

Abstract

Version 3.3 corrected English preprint and reproducibility materials for a measurement study of why international cybersecurity-education programme counts change under reasonable choices of source frame, retrieval language, temporal evidence, entity resolution and counting unit. The design separates a worldwide candidate-evidence layer from 14 purposively selected mechanism probes, a five-country verified pilot and two directed prospective rounds. The worldwide layer diagnoses observability and source concentration; it does not estimate global programme prevalence or support country rankings. Round 2 records contrasts for Luxembourg and Iceland but is not confirmatorily evaluable because 20 provider-level zero-result searches lack preserved query-level evidence. CYBERUS is coded conservatively as one integrated programme, with two-track and three-location alternatives reported as sensitivity scenarios. The record includes full and short PDFs plus a 52-file reproducibility package with protocols, version history, registry, evidence tables, deterministic audit scripts, human-coder provenance, MANIFEST.csv and SHA-256 checksums. This is an author preprint, not a peer-reviewed paper.