What we see depends on how we measure: Modeling-based instruction and the visibility of learning
International Journal of Educational Research, cilt.139, 2026 (SSCI, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 139
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.ijer.2026.103048
- Dergi Adı: International Journal of Educational Research
- Derginin Tarandığı İndeksler: Social Sciences Citation Index (SSCI), Scopus, Periodicals Index Online, EBSCO Education Source, Education Abstracts, Educational research abstracts (ERA), Index Islamicus, Academic Search Ultimate (EBSCO), Education Source Ultimate (EBSCO)
- Anahtar Kelimeler: Assessment format, Astronomy education, Epistemic explanation, Modeling-based instruction, Science education
- TED Üniversitesi Adresli: Evet
Özet
Assessment format plays a critical role not only in how learning is measured but also in which aspects of learning become visible. Grounded in a validity-based perspective and epistemic cognition, this study examined how Modeling-Based Instruction (MBI) in an elementary astronomy context influenced students’ conceptual understanding and epistemic explanation, and how these effects were represented across multiple-choice (MC) and open-ended (OE) assessments. The study adopts an assessment-centered approach to examine how assessment format shaped the evidence available for interpreting students’ learning. The findings revealed a clear format-sensitive pattern. While MBI did not yield a statistically significant improvement in MC-based performance, it was associated with a meaningful increase in OE-based explanation quality and epistemic reasoning. Gain comparisons indicated that learning outcomes were more visible in the OE tasks used in this study than in the conventional four-option MC items. Analysis of students’ modeling performances showed a differentiated epistemic profile, with relatively strong scientific accuracy but weaker performance in justification, evidence use, and epistemic use of models. Although process-level epistemic dimensions were positively related to OE outcomes, their unique predictive contribution diminished once baseline explanation ability was controlled. Across formats, results indicated limited convergence, clear divergence in performance patterns, and low classification agreement, suggesting that the specific MC and OE tasks used in this study provided complementary rather than interchangeable evidence about student learning. Taken together, findings demonstrate that conclusions about instructional effectiveness depended in part on the specific assessment formats used in this study, as these formats elicited different forms of learning evidence. The study contributes to science education research by highlighting how the design of assessment tasks and response formats can shape the interpretation of learning outcomes, particularly when instruction targets explanation, justification, and model-based reasoning.