Students’ Evaluative Thinking during AI-Supported Observation as a Cognitive Comparison in Acid Strength Learning

Abstract

This study examines how students enact evaluative thinking during the observation phase of Predict–Observe–Explain (POE) learning when supported by generative AI on the topic of acid strength. A qualitative study was conducted with eight eleventh-grade students. Generative AI was integrated exclusively during the Observe phase to provide explanations related to acid strength (HF and HI), including pKa values and conjugate base stability, for comparison with students’ initial predictions. Data from students’ written responses, reflective journals, and AI interaction logs were analyzed thematically. The findings indicate three categories of evaluative thinking: descriptive–comparative evaluation, emerging evaluative awareness, and reflective evaluative reasoning. Most students demonstrated evaluative thinking at a descriptive–comparative level, while reflective evaluative reasoning involving explicit conceptual revision appeared in a limited number of cases, suggesting the need for instructional scaffolding to support deeper evaluative engagement when generative AI is used as a cognitive comparison resource.