- BadBanks & CAT Overlap. Simulation frameworks for adaptive testing under calibration error, item reuse, and overlap constraints; evaluating rescoring, selection, and stopping rules.
- Balancing Bias / APE (AI‑assisted scoring). Ensemble LLM raters calibrated against human biases in MFRM; improving reliability and fairness.
- GMM Diagnostics & IMV. Identification checks, posterior instability diagnostics, and predictive comparisons between GMM and simpler growth models.
- LEVANTE. Cross‑cultural growth modeling and fairness evaluation across multilingual contexts (U.S., Canada, Colombia, Germany).
- Automated scoring at operational scale (ACER). Measurement criteria for when machine-generated scores can carry part of an operational marking load, and the human oversight required around them.
- Scaling and linking for large-scale assessment (ACER). Scale comparability across forms, cycles, and instruments not originally designed to be linked.