A classroom-based dataset on phase-level assessment of mathematical modelling processes in upper secondary education

Abstract

This classroom-based dataset documents phase-level mathematical modelling activity from the same 37 upper-secondary students across two study occasions, yielding 74 student-by-occasion records. The dataset includes explicit variables for participant identifier, study occasion, instructional condition, modelling task, measurement order, and five modelling-phase scores coded on an ordinal 0–1–2 scale (0 = not achieved, 1 = partially achieved, 2 = achieved). The repository includes the primary dataset, a complete codebook, a phase-specific scoring rubric, independent second-coder scores, inter-rater reliability statistics, technical-validation documentation, and the study instruments. Primary coding was conducted by ROUISSA Lamyae. Oudrhiri-Idrissi Mohamed independently re-coded the anonymised student productions using the same five-phase rubric for inter-rater reliability assessment. The reliability analysis, including calculation of Cohen’s kappa, was conducted by Lamyae Rouissa. The first occasion used the Konrad Adenauer head-statue modelling task without generative-AI access, whereas the second used the Uwe Seeler foot-statue task with ChatGPT accessed through the standard free web interface as a classroom resource. Because task, administration order, and AI availability changed across occasions, the dataset is intended for descriptive, exploratory, methodological, and paired within-student analyses rather than causal attribution of differences to ChatGPT. All publicly released records are anonymised and contain no direct participant identifiers. The dataset is released under the Creative Commons Attribution 4.0 International licence (CC BY 4.0).