Reproduction package for "One weight for many variables: why multipurpose cal-ibration of non-probability samples needs flexible working models, and what sharing it costs"

Résumé fourni par la source

Simulation and reproduction code for the paper. Contains the Monte Carlo study of the stacked multipurpose model-calibration weight, the rank-condition test and its cross-fitting counterpart, the double-robustness experiment, the frontier and lever experiments, the volunteer/reference overlap experiment, and the design-based study on the 2025 Uruguayan Continuous Household Survey microdata. Thirteen R scripts, twenty-five result files, a manifest mapping every table and figure in the paper to the script and named configuration that produces it, and a record of the R and package versions used. Seeds are fixed and the run configuration is passed as an argument rather than by editing the scripts, so the deposited text is fixed. The ECH microdata are not redistributed here; they are the public-use file published by the Instituto Nacional de Estadistica of Uruguay. Requires the weightflow R package (CRAN 1.2.0) for all results except the bounded calibration arms of Section 8, which require the development build pinned in the README.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.