Data, Materials, and Reproducibility Files for "Evaluating AI career advice after scaffold withdrawal: a cluster-randomized trial of TRACE in vocational education"

This repository contains de-identified analysis data, scoring documentation, study materials, analysis code, retained statistical outputs, and reproducibility files supporting the manuscript “Evaluating AI career advice after scaffold withdrawal: a cluster-randomized trial of TRACE in vocational education.”

The study used a teacher-blocked 2 × 2 cluster-randomized design involving 480 students in 24 higher-vocational classes. The repository supports analyses of unaided transfer and performance on a standardized AI-advice evaluation task, including class-level randomization inference, cluster-robust analyses, scoring sensitivities, completion analyses, and judgment-transition analyses.

The deposit also includes the internally dated analysis plan finalized on 12 February 2026, together with an English translation. The analysis plan preceded baseline data collection and intervention delivery but was publicly archived retrospectively; it should therefore not be interpreted as a prospective preregistration.

Public files exclude direct identifiers, participant linkage keys, verbatim participant responses, signed consent materials, and detailed operational records not required for reproducibility. The authors retain responsibility for the data, analysis, interpretation, and deposited materials.