Replication package: Elicitation Design and the Comparability of Emigration-Intention Estimates among Health Professionals: A Systematic Review and Meta-Analysis

Abstract

Constructed data, estimation code, estimated artefacts, figures and checks for the article "Prevalence of Emigration Intention among Health Professionals: A Systematic Review and Meta-Analysis". The package contains: the analysis panel of 80 prevalence estimates drawn from 76 studies in 32 countries and covering 54,808 respondents, each coded for source setting, region, cadre, career stage and the wording of the intention question; the source-verification log recording every extracted value against the sentence in the primary report that supports it; the screening decisions with written reasons and confidence grades; the determinant evidence map of 26 quote-verified effects; the full estimation run; every estimated artefact as a CSV or a log, one per reported table or diagnostic; every figure at both the reference and the submission scale; and the checks that verify all of it. Every number reported in the article is produced by a script in 01_code and written to an artefact in 03_artefacts. No number is entered by hand. Every check carries a planted-failure self-test and is run in both directions: it must fire on the defect it exists to catch and pass on the repair. Raw inputs are published journal articles and are not redistributed here. The 76 included studies are cited in full in the article's reference list, and 02_constructed_data carries what those articles become after extraction, coding and verification. README.md gives the run order, the software versions and the package requirements. MANIFEST.csv lists every file with its size and SHA-256 checksum. Software: Stata 16 or later for the estimation, built-in meta suite only. R 4.3.3 for the figures and derived quantities; the png package is required by the crop step, nothing else beyond base R. Python 3 for the checks.