Predictive Difficulty Beyond Hidden-State Cardinality: Data and Code Artifact

This data and code artifact supports the manuscript Predictive Difficulty Beyond Hidden-State Cardinality: A Shared Process Component and Learner-Specific Responses in Controlled Hidden Markov Models. The artifact contains selected hidden Markov model parameters for the development, confirmatory, predictor-specific, strict-replication, external-generator, and dataset-realization robustness analyses; run-level and aggregate experimental results; selection and protocol records; cleaned Colab notebooks; validation and reconstruction scripts; and outputs used to reproduce the reported tables and figures. The release is intended to support reproducibility and auditing of the experimental analyses reported in the manuscript. The regenerated sequence arrays used in the dataset-realization robustness experiment are not included, but they can be reproduced from the included HMM parameters, recorded split seeds, and public runner. Software under code/ is released under the MIT License. Data, derived results, figures, and documentation under data/, outputs/, and docs/ are released under Creative Commons Attribution 4.0 International (CC BY 4.0).