Replication Data and Code for “When Platform Ratings Are Not Commensurable: An Architecture-Aware Audit of Cross-Source Evaluation Data”
Description
This repository contains the derived analytical data, analysis code, model outputs, tables, and figures used to reproduce the empirical analyses reported in the manuscript. The study examines the information-quality implications of integrating cross-source platform ratings using a candidate frame of 10,590 Steam games released between 2015 and 2025, including 984 titles with Steam-embedded Metacritic scores. The analyses cover: (1) selectivity in Steam-embedded Metacritic availability; (2) product-level calibration and observed dispersion differences between Steam recommendation shares and embedded Metacritic scores; (3) ordinal correspondence and percentile-rank displacement; (4) pairwise ordering reversals with a Steam-side Wilson-interval reliability screen; and (5) robustness and linkage-sensitivity analyses. The package is intended to support transparency, reproducibility, and reuse of the study's architecture-aware commensuration audit. The underlying variables were collected from publicly accessible Steam endpoints and Steam appdetails metadata. No private user-level data or personally identifiable information are included. Users of the package should observe the applicable terms and conditions of the original data sources.