Date of Publication
4-2025
Document Type
Dissertation
Degree Name
Doctor of Philosophy in Science Education Major in Physics
Subject Categories
Science and Mathematics Education
College
Br. Andrew Gonzalez FSC College of Education
Department/Unit
Science Education
Thesis Advisor
Lydia S. Roleda
Defense Panel Chair
Maricar S. Prudente
Defense Panel Member
Voltaire Mallari Mistades
Joel T. Maquiling
Minie Rose C. Lapinid
Abstract (English)
This study investigated ChatGPT’s potential to enhance flipped Physics instruction in two ways: 1) as an AI-driven chatbot guide for students’ pre-class learning and 2) as a tool for generating formative and summative Physics assessments. Quantitative data were collected from students’ pre-test scores across different flipped learning models (videos only, videos plus ChatGPT, and ChatGPT alone), while qualitative feedback captured their preferences and experiences. Statistical analyses revealed no significant differences in pre-test performance among the three groups, indicating that ChatGPT is as effective as the traditional usage of videos in flipped Physics instruction. Qualitative findings highlight ChatGPT’s added value in providing personalized feedback, time-efficient study support, and flexible problem-solving inquiries during flipped instruction.
Simultaneously, multiple-choice assessments generated by ChatGPT-4 were evaluated using Classical Test Theory. Results showed a balanced range of item difficulties and acceptable discrimination indices, suggesting that AI chatbots, specifically ChatGPT, can create valid physics assessments, potentially reducing teachers' workload in creating assessments. The study also discussed how to efficiently generate large-scale assessments and contextualized assessments, guiding educators in specifying content parameters to ensure alignment with the learning objectives. Nonetheless, expert validation remains critical to address occasional inaccuracies, emphasizing the importance of human-AI collaboration.
Keywords: Flipped Instruction, Artificial Intelligence, Chatbots, ChatGPT
Abstract Format
html
Abstract (Filipino)
Sinuri ng pag-aaral na ito ang potensiyal ng ChatGPT na mapahusay ang flipped Physics instruction sa dalawang pangunahing paraan: (1) bilang isang AI-driven chatbot guide para sa pre-class na pag-aaral ng mga estudyante, at (2) bilang kasangkapan sa pagbuo ng mga formative at summative na pagtataya sa Physics. Gumamit ang pag-aaral ng quasi-experimental na disenyo upang ihambing ang mga pre-test score ng mga estudyante mula sa tatlong flipped learning models, videos lamang, videos na may ChatGPT, at ChatGPT lamang,habang ang mga datos na kwalitatibo ay nagtipon ng kanilang mga pananaw at karanasan.
Ipinakita ng statistical analyses na walang makabuluhang pagkakaiba sa pre-test performance ng tatlong grupo, na nagpapahiwatig na ang ChatGPT ay kasing-epektibo ng tradisyonal na paggamit ng mga video sa flipped Physics instruction. Samantala, ibinunyag ng kwalitatibong resulta na nagdadagdag ng halaga ang ChatGPT sa pamamagitan ng pagbibigay ng personalized na feedback, mas episyenteng oras sa pag-aaral, at mas malayang pagtatanong sa problem-solving habang nagaganap ang flipped instruction.
Kasabay nito, sinuri rin ang mga multiple-choice test items na binuo ng ChatGPT-4 gamit ang Classical Test Theory. Nagpakita ang mga item ng balanseng antas ng kahirapan at katanggap-tanggap na discrimination indices, na nagpapahiwatig na ang AI chatbots, partikular ang ChatGPT, ay kayang lumikha ng valid na physics assessments na maaaring magpababa ng workload ng mga guro sa paggawa ng pagsusulit. Tinalakay rin ng pag-aaral ang mga pamamaraan sa mas episyenteng pagbuo ng malakihang pagtataya at contextualized assessments, kabilang ang paggabay sa mga guro sa pagtukoy ng content parameters upang matiyak ang alignment sa learning objectives. Gayunpaman, nananatiling mahalaga ang ekspertong beripikasyon upang tugunan ang mga posibleng pagkakamali, na binibigyang-diin ang kahalagahan ng human-AI collaboration.
Mga Keyword: Flipped Instruction, Artificial Intelligence, Chatbots, ChatGPT
Abstract Format
html
Language
English
Format
Electronic
Keywords
Physics—Study and teaching (Secondary)—Technological innovations; Flipped classrooms; Artificial intelligence—Educational applications; Chatbots
Recommended Citation
Camungol, I. F. (2025). Leveraging AI in flipped physics instruction: As students’ chatbot guide and physics assessment generator. Retrieved from https://animorepository.dlsu.edu.ph/etdd_scied/60
Upload Full Text
wf_yes
Embargo Period
4-2028
