Reproducibility Package for: Human-Study Dataset and Reproducibility Package for Staged Multi-Criteria Disclosure in ML Model Selection

This record documents the underlying human-study dataset and reproducibility materials for a staged multi-criteria disclosure study in ML model selection. The archived materials were prepared by Adil Joldić and Nina Bijedić and support multiple derived publications based on the same broader study, with version-specific processed files corresponding to different analytic snapshots and paper-level analyses. The record includes processed study data, frozen policy and stimulus specifications, and reporting artifacts needed to verify published analyses. It does not include the full live study application or the complete agent system codebase; instead, it focuses on the processed study data, frozen scenario specifications, and analysis artifacts required for paper-level reproducibility. Version 1 corresponds to the ISDA 2026 paper “From Unaided Choice to Multi-Criteria Disclosure: A Human Study of ML Model Selection in Higher-Education Quality Assurance”. This version is tied to export snapshot 20260312_124341Z and corresponds to the paper snapshot with: 37 participants / 222 confirmatory scenarios 89 participants / 533 sensitivity scenarios It includes: anonymized confirmatory and sensitivity scenario-level datasets, frozen criterion weights and frozen stimulus payload export, exported statistical summaries, a reproduction script and README. This version supports reproduction of the reported tables and key descriptive results for the ISDA 2026 paper from the frozen export snapshot.