Determinants of metacognition in physics learning among high school students: a mixed-methods and machine learning approach
Metacognition in physics fundamentally shapes how students map out, oversee, and assess their cognitive steps during problem-solving. To untangle the complex factors behind this ability, we adopted an explanatory sequential mixed methods framework. First, we deployed a machine learning model via the Unsupervised Possibility Fuzzy C-Means (UPFC) algorithm in R, allowing us to group students based on structural similarities in their metacognitive profiles. We then followed up with qualitative, in-depth interviews analyzed through thematic coding in Atlas.ti to contextualize these groups. Our analytics uncovered two distinct student groups: those with high metacognitive engagement and those with low engagement. The driving forces behind these profiles stem from both the inside and outside. Internally, students’ capacity for self-evaluation, active task monitoring, and cognitive resilience when facing tough physics problems proved decisive. Externally, the classroom climate, active teacher scaffolding, resource accessibility, and emotional backing from family and peers heavily steered their development. Ultimately, these dynamics offer actionable blueprints for educators aiming to cultivate sharper metacognitive habits in the physics classroom.
Authors
- Edi Istiyono (ORCID: https://orcid.org/0000-0001-6034-142X)
- Okta Alpindo (ORCID: https://orcid.org/0000-0002-6315-5689)
- Widihastuti Widihastuti (ORCID: https://orcid.org/0000-0001-8242-658X)
Institutions
- Yogyakarta State University (ID)
- Universitas Maritim Raja Ali Haji (ID)
Publication Details
- Journal
- Investigações em Ensino de Ciências
- Published
- 2026-11-20
- DOI
- https://doi.org/10.22600/1518-8795.ienci/2026v31n3p769
- Primary Topic
- Innovative Teaching and Learning Methods
- Type
- article
- Field-Weighted Citation Impact
- 0.00