Reproducibility package for: What the Balkan Barometer can and cannot rank

Code, derived tables and results behind a measurement audit of the Regional Cooperation Council's Balkan Barometer public-opinion survey, editions 2023 to 2025. The package reproduces every number in the working paper: the reconciliation of the published percentages against the released microdata (440 of 444 cells within one point); the pairwise rank resolution between the six Western Balkan economies at design effects from 1.0 to 3.0, with Holm correction and simulated rank intervals; the design-effect proxies estimated from the municipality and region structure; the population-weighted alternative to the equal-weight WB6 aggregate; the graded comparability of the headline series across editions and the trend contrast that follows from it; the measurement invariance analysis of the institutional and regional-cooperation batteries (configural to scalar under WLSMV, alignment with bootstrap standard errors, and ordinal logistic DIF); the response-quality indicators; and the sixteen-element transparency coding of the three editions. The package carries aggregate statistics, code and text only. The Balkan Barometer microdata, the saved database pages and the published reports are the RCC's and are not redistributed; the README lists each excluded file with its size and SHA-256 and gives the RCC page to download it from. Version 1.1 (29 September 2026) changes one thing: the manuscript PDF carried in docs/ is now version 2 of the working paper, which corrects the reference apparatus after a full citation and claim-support check. Two sentences were narrowed to what their sources show, nineteen missing DOIs were added, three citations were corrected to the APA form for three or more authors, a citation to the World Development Indicators was added where the text used them, and one issue number was completed. No code, no derived table and no result changed, so every number in results/ is identical to version 1.0.