Abstract
Reading literacy remains a persistent challenge in many English as a Foreign Language (EFL) contexts. However, primary EFL learners frequently encounter difficulties in integrating linguistic, visual, spatial, and gestural modalities. This study investigates how primary learners exercise epistemic agency in multimodal reading and construct knowledge across different modes. Utilizing multimodality theory and epistemic agency, the study implemented an explanatory sequential mixed-methods design involving thirty fifth-grade students from Madrasah Ibtidaiyah Negeri (MIN) 1 Tulang Bawang Barat, Lampung. In Phase 1, a one-group pretest–posttest design measured changes in multimodal comprehension using a validated twenty-item instrument. Results showed significant improvement, with mean scores rising from 61.40 (SD = 8.25) to 74.30 (SD = 7.80), t(29) = 8.52, p < 0.001, and Cohen’s d = 1.56, indicating a large effect. In Phase 2, six students representing varied improvement levels were interviewed and engaged in task-based reflections. Thematic analysis identified five dimensions of epistemic agency: taking initiative, decision-making across modes, justification of understanding, reflection and self-regulation, and collaboration and dialogue. The integration of both phases demonstrated that learners' agentive engagement with multimodal resources underpinned comprehension gains, transforming affordances into strategies for meaning-making. These findings affirm that multimodal pedagogy fosters not only measurable comprehension outcomes but also autonomy, critical thinking, and collaborative engagement. The study recommends embedding multimodal literacy into EFL curricula and providing scaffolds that balance cognitive support with opportunities for agency.