Automated Scoring of Young Learners’ Student-Generated Questioning in Elementary Reading: A Machine-Learning Approach

• Automated scoring can support assessment of SGQ in elementary reading. • The study compared feature-based and deep learning models for SGQ quality scoring. • Results show strong alignment between machine and human scoring of cognitive engagement. • Feature-based models offered interpretability but lower scoring performance. • Results support the use of ML for scalable and interpretable SGQ assessment.