Claim Missing Document
Check
Articles

Application of Thermodynamic Wave KIT-based Deep Learning to Reduce Misconceptions in Grade XI High School Students Regarding Heat Muzdalifa A. Madusila; Asri Arbie; Meilan Demulawa
Kasuari: Physics Education Journal (KPEJ) Vol. 9 No. 1 (2026): June 2026
Publisher : Universitas Papua, Jurusan Pendidikan Fisika FKIP Unipa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30862/kpej.v9i1.1108

Abstract

This study aims to reduce the level of misconceptions experienced by eleventh-grade high school students on Heat material through the application of a deep learning approach combined with the use of the Wave and Thermodynamics KIT. The research method used was a quasi-experimental one-group pretest–posttest design, involving one class as the experimental class and two classes as replication classes. The research instrument was a three-tier multiple-choice diagnostic test used to map students' misconceptions before and after the learning treatment. Data were analysed using descriptive statistics, two-proportion tests, and normalised gain (n-gain) calculations. The research findings showed a very significant decrease in the frequency of misconceptions, namely from 6.10 to 1.00 in the experimental class; from 5.53 to 1.40 in the first replication class; and from 6.33 to 1.33 in the second replication class. This decrease was also accompanied by an increase in the conception score in each class, as reflected in average n-gain values of 0.69 in the experimental class, 0.72 in the first replication class, and 0.76 in the second replication class. Thus, it can be concluded that the application of deep learning supported by the Wave and Thermodynamics KIT is effective in reducing high school students' misconceptions about Heat material while simultaneously increasing their conceptual understanding to a medium to high level.
Pengaruh Pembelajaran Berbasis Simulasi Phet Dengan Pendekatan Deep Learning Terhadap Pengurangan Miskonsepsi Siswa Pada Materi Gelombang Menggunakan Analisis Rasch Olivia Talalu; Mohamad Jahja; Meilan Demulawa
Science and Education Journal (SICEDU) Vol 5 No 3 (2026): Science and Education Journal 2026
Publisher : LPPM Universitas Pahlawan Tuanku Tambusai

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31004/sicedu.v5i3.1264

Abstract

Miskonsepsi pada materi gelombang menjadi salah satu permasalahan dalam pembelajaran fisika karena karakteristik konsepnya yang abstrak dan sulit diamati secara langsung. Penelitian ini bertujuan menganalisis pengaruh pembelajaran berbasis simulasi PhET yang diintegrasikan dengan pendekatan Deep Learning terhadap pengurangan miskonsepsi peserta didik pada materi gelombang serta mengidentifikasi perubahan kemampuan konseptual menggunakan Model Rasch. Penelitian menggunakan pendekatan kuantitatif dengan metode quasi-experiment dan desain Pretest-Posttest Replicated Experimental Design. Subjek penelitian terdiri atas 59 peserta didik dari dua kelas XI. Data dikumpulkan menggunakan Three-Tier Diagnostic Test dan dianalisis secara deskriptif serta menggunakan Model Rasch melalui Winsteps 5.11.2. Hasil penelitian menunjukkan peningkatan kategori Paham Konsep dari 12,0% menjadi 88,0% pada kelas XI-1 dan dari 19,7% menjadi 86,9% pada kelas XI-2. Sebaliknya, total miskonsepsi menurun dari 59,1% menjadi 8,3% pada kelas XI-1 dan dari 54,8% menjadi 7,2% pada kelas XI-2. Analisis Rasch juga menunjukkan peningkatan Person Measure serta reliabilitas item yang tinggi pada post-test. Temuan ini menunjukkan bahwa integrasi simulasi PhET dan pendekatan Deep Learning efektif mendukung perubahan konseptual dan mengurangi miskonsepsi pada materi gelombang.