Evidence Talk and Transferable Problem Solving in Secondary Science Learning

Abstract

This study investigated how evidence talk contributed to transferable problem solving among public secondary science learners in San Pablo, Isabela. Using a multimethod explanatory-predictive design, data were obtained from 156 learners through the Evidence Talk Scale and a scenario-based Transferable Science Problem-Solving Task. Partial least squares structural equation modeling, predictive assessment, and importance-performance map analysis were employed to determine the strength and predictive relevance of evidence-centered classroom discourse. Results showed that learners demonstrated a generally high level of evidence talk but weaker performance in critically responding to alternative explanations and revising ideas after discussion. Transferable problem solving was generally proficient, although difficulties were evident in using evidence to construct solutions and justifying final answers. Evidence talk significantly predicted transferable problem solving, β = 0.57, p < .001, and explained 35% of its variance. Predictive assessment further indicated moderate out-of-sample predictive power. Among the dimensions of evidence talk, justification of reasoning showed the greatest predictive importance, while critical response and revision of ideas emerged as areas requiring greater instructional attention. The findings indicate that classroom discourse becomes more instructionally valuable when learners are required not only to state evidence but also to defend, question, compare, and revise scientific explanations. Science instruction should therefore provide sustained opportunities for evidence-based dialogue and unfamiliar problem-solving tasks that require flexible application of scientific understanding.