Ruochen Ning, Thomas Canhao Xu

DOI: http://doi.org/10.4208/itl.20260106

Innovative Teaching and Learning, Vol. 8 (2026), Iss. 1 : pp. 89–115

Published online:    2026-5

Abstract

This design based study introduces a mechanism informed checklist and a ready fading hints, contrastive examples, and formative probes. We re analyze a representative classroom assessment (Quiz: quartiles/IQR, normalization, confusion labeling to illustrate errors; <5% feedback failures) and provide reproducible templates. We argue that when randomized controlled trials (RCTs) are infeasible, mechanism alignment and falsifiable thresholds enable pragmatic, trustworthy improvement.

Keywords

AI in education, data science instruction, evidence informed checklist, cognitive load, fading scaffolds, formative assessment

View Original Text: Click Here