Evaluating Students’ Scientific Inquiry Strategies in Large-Scale Digital Assessments

Abstract

As education shifts toward digital learning environments, new assessments are needed to evaluate not only what students know but also how they learn using available tools and resources. The OECD Programme for International Student Assessment (PISA) 2025 Learning in the Digital World (LDW) assessment addresses this challenge by engaging students in open-ended, interactive tasks that require them to apply learning content to solve inquiry-based problems. This paper focuses on two key scientific inquiry strategies assessed through LDW tasks: Control of Variable (CoV), which reflects how systematically students conduct inquiry exper-iments, and Deriving Relationship from Data (DRD), which involves interpreting re-lationships or patterns from the experimental results. We present a scoring rubric designed for fine-grained analysis of student performance, allowing for partial credit to be linked to degrees of applying and reasoning with the strategies to derive solutions to assigned problems. Using a learning-by-modeling task in the LDW framework, we apply this rubric to pilot data collected from 6,800 students across 63 countries. Our findings show that students’ success with the CoV and DRD strate-gies is influenced not only by prior knowledge but also by how they engage with instructional phases and utilize digital and scaffolding tools during the task.