Supplementary Materials for Peer Review
Manuscript: What Really Drives Teachers to Learn? Insights from an Explainable Machine Learning Approach
Prepared for anonymous technical-check revision. No author names or identifying author details are included.

Contents
1. data/raw_survey_export_deidentified_all_records.csv and .xlsx
   - De-identified raw survey export with 531 records.
   - Direct identifiers are not included. School identifiers are recoded as anonymous school-cluster codes.

2. data/raw_valid_item_level_deidentified.csv and .xlsx
   - De-identified valid-case item-level dataset with 472 records.
   - This file retains item-level variables for the valid analysis sample.

3. data/processed_analysis_dataset_deidentified.csv and .xlsx
   - Processed analysis dataset with 472 records.
   - Includes the continuous outcome TAML, the binary classification target MOTIVATION_HIGH, demographic variables, leadership, workplace, learning-conception, and self-efficacy composites.

4. data/variable_codebook_and_processing_log.xlsx
   - Variable definitions, value labels, processing log, and overview.
   - CSV versions are also provided for machine readability.

5. instrument_and_protocol/Survey_Instrument_and_Protocol_HSSC.docx
   - Survey domains, item-code structure, response-format notes, and data-collection protocol.

6. analysis_scripts/reproduce_analysis.py
   - Reproducible analysis template for preprocessing, cross-validated model comparison, XGBoost fitting, and SHAP interpretation.
   - The script requires the packages listed in requirements.txt.

7. Supplementary_SHAP_Tables_HSSC.docx
   - Previously prepared supplementary SHAP tables, included for reviewer convenience.

Privacy and confidentiality
The original identifiable SPSS files are not included because they contain direct identifiers or administrative identifiers. The shared de-identified files contain the variables necessary to inspect the methodology and reproduce the reported analysis workflow.
