A Reproducible Survey-Scoring And C5.0 Decision-Tree Workflow For Classifying Self-Reported Higher-Order Thinking In...

Generative artificial intelligence is increasingly used in higher education, but researchers still need transparent procedures for collecting, scoring, and organizing learner-level survey data in this rapidly changing context. This article presents a reproducible survey-scoring and C5.0 decision-tree workflow for classifying self-reported higher-order thinking among college students who have used generative artificial intelligence tools for academic learning. The protocol covers participant recruitment, questionnaire administration, response-quality screening, composite-score calculation, binary coding, training-testing partitioning, C5.0 tree construction, pruning, model-output export, and interpretation of node-based classification paths. The workflow is demonstrated using a single-university survey dataset of 776 undergraduate students collected in China from March 7 to March 15, 2023. Higher-order thinking is operationalized as a self-reported questionnaire score rather than as directly observed cognitive performance. In the demonstration dataset, the pruned tree retained eight learner-related variables: generative artificial intelligence anxiety, trust in generative artificial intelligence, problematic smartphone use, academic procrastination, academic performance, parental upbringing, negative emotions, and attitudes toward generative artificial intelligence. The retained model achieved 89.52% accuracy in the training subset and 86.21% accuracy in the testing subset. However, class-specific evaluation showed uneven performance, with substantially weaker recall for the Low-HOT class than for the High-HOT class. Therefore, the tree should be interpreted as an auxiliary, interpretable workflow demonstration rather than as a validated screening tool. This protocol may support researchers who require a documented and repeatable procedure for survey-based learner profiling in generative artificial intelligence-supported learning environments.