Replication data and code for "What People Say, What Pools Need: Why Willingness to Pay Cannot Forecast the Survival of Voluntary Health Insurance

Replication materials for the paper "What People Say, What Pools Need: Why Willingness to Pay Cannot Forecast the Survival of Voluntary Health Insurance" (D. M. Dror). The repository lets a reviewer reproduce, offline and from deposited data, every number the paper's own analyses rest on: the income-direction classification across the WTP-for-health-insurance literature (§5.2: income is a positive correlate of declared WTP in 22 of 26 explicit-sign primary abstracts, 57 of 61 with determinant-by-convention abstracts added); the validation-gap systematic search (§6: 11,110 records across four query families in PubMed and Europe PMC, reduced to 4,855 unique, extended to OpenAlex, Crossref, and DOAJ; the search corpus and both search logs are deposited), in which no study validates declared willingness to pay against realised payment sustained across renewal; a WTP concept genealogy; and the two-scheme sustainability illustration (§7). A single self-locating master script, scripts/replicate_all.R, reproduces these numbers from the deposited files and prints a match report; it runs from any location without path edits and requires only R (>= 4.5) with base packages. The optional extension-search script (scripts/04_extension_search_harvest.R), which reproduces the OpenAlex/Crossref/DOAJ search, additionally requires the jsonlite package; it is not needed to reproduce the headline numbers. The six-district income-share correlations of §5.1 are derived from an access-restricted household survey (WOTRO/NWO Integrated Programme, grant W 01.65.309.00); the repository deposits the derived correlation and regression statistics, verified against the manuscript, together with two co-author source workbooks. The underlying raw microdata is not deposited and is available from the author on reasonable request under the original grant terms. Figures drawn from other authors' published papers are documented with their exact locus (table, page) for verification against the cited article and are not redistributed. The repository is organized into data/, analysis/, outputs/, and scripts/, with README.md, PROVENANCE_MAP.md (a claim-by-claim index), CODEBOOK.md, MANIFEST.txt (MD5 + row counts), and LICENSE.txt.