Evaluating the ChatGPT-Assisted Learning Method (CALM) for Struggling Readers: The Emergence of the Theory of Distributed Reading Gains

This study aimed to determine the efficacy of ChatGPT-Assisted Learning Method (CALM) in enhancing reading proficiency among 9th grade students in a public secondary school situated in the Philippines. The study used a sequential explanatory mixed-methods, single-group pretest–posttest design. A total of N = 50 students were purposely selected who were identified as frustrated readers during the Phil-IRI assessment. The intervention period lasted eight weeks and was assessed using a pretest and posttest. To gain a deeper understanding of the participants' significant experiences, this was followed by a qualitative phase, in which five participants were selected as the saturated sub-sample to undergo one-on-one semi-structured interviews and focus groups. Thematic analysis was used to isolate the functional mechanisms of the AI-mediated scaffold. Results from the quantitative data revealed an increase in reading proficiency, with M = 5.66, and students reaching "Proficient" or "Advanced" levels rising from 54% to 92%. Cohen’s d = 1.45 indicates a large effect, while η2 = .34 suggests that 34% of the variance in reading proficiency is attributable to the CALM intervention. Hence, a reduction in the mean standard deviation from 4.13 to 2.82 indicates a significant improvement in reading performance. Furthermore, five themes emerged from the qualitative findings supporting the result from the quantitative phase which described as core functional mechanisms: (1) instructional responsiveness and mastery growth; (2) dialogic adaptability and strategic autonomy; (3) diagnostic feedback and self-regulated revision; (4) metacognitive prompting skill building; and (5) ownership and clarity through reconstructed expression, marking the internalization of the digital scaffold. This evidence establishes the Theory of Distributed Reading Gains (TDRG) through the CALM Feedback−Prompt−Scaffold (FPS) Instructional Framework cycle which operates how distributed gains arise in resource-constrained high-density environments.