PENGEMBANGAN LMS-PBL TERINTEGRASI CHATBOT MATERI EKOLOGI UNTUK MENINGKATKAN KETERAMPILAN PEMECAHAN MASALAH DAN SELF EFFICACY MURID KELAS 7

DEVELOPMENT OF AN LMS-PBL INTEGRATED CHATBOT FOR ECOLOGY LEARNING TO ENHANCE SEVENTH GRADE STUDENTS' PROBLEM SOLVING SKILLS AND SELF EFFICACY

ABSTRAK

Yulandari, Silvi Novrian. 2026. Pengembangan LMS-PBL Terintegrasi Chatbot Materi Ekologi untuk Meningkatkan Keterampilan Pemecahan Masalah dan Self-Efficacy Murid Kelas 7. Tesis, Program Studi Magister Pendidikan Sains, Fakultas Matematika dan Ilmu Pengetahuan Alam, Universitas Negeri Surabaya. Pembimbing I Prof. Dr. Fida Rachmadiarti, M.Kes. dan Pembimbing II Mita Anggaryani, M.Pd., Ph.D.

Kata Kunci: LMS, Problem Based Learning, chatbot, keterampilan pemecahan masalah, self-efficacy, ekologi.

Keterampilan pemecahan masalah (KPM) dan self-efficacy murid pada pembelajaran ekologi di Gresik masih tergolong rendah, sementara persoalan banjir Daerah Aliran Sungai Kali Lamong yang dekat dengan keseharian murid belum banyak diangkat sebagai konteks pembelajaran IPA. Penelitian ini bertujuan menghasilkan Learning Management System berbasis Problem Based Learning yang terintegrasi chatbot (LMS-PBL) pada materi ekologi kelas 7 yang valid, praktis, dan efektif untuk meningkatkan KPM serta self-efficacy murid. Pengembangan produk mengacu pada model Hannafin dan Peck yang mencakup tahap penilaian kebutuhan, perancangan, serta pengembangan dan implementasi. Produk yang dihasilkan berupa LMS berbasis Moodle beralamat exploresains.com yang dilengkapi chatbot bernama BioBot atau Deka. Implementasi produk dilakukan dalam dua tahap dengan desain One Group Pretest-Posttest. Uji coba terbatas dilaksanakan pada 22 murid kelas 7A SMP NU Bahrul Ulum untuk memastikan kelayakan operasional produk sebelum diperluas. Uji coba luas dilaksanakan pada 80 murid, terdiri atas 42 murid SMP NU Bahrul Ulum (kelas 7C dan 7D) dan 38 murid SMP IT Utsman bin Affan (kelas VII-A dan VII-B), untuk menguji kekonsistenan produk pada konteks sekolah yang berbeda. Data dikumpulkan melalui lembar validasi, lembar keterlaksanaan pembelajaran, angket respons murid, tes diagnostik four-tier untuk KPM, dan kuesioner self-efficacy 17 butir. Analisis data mencakup persentase kevalidan dan kepraktisan, reliabilitas antarvalidator dengan Percentage of Agreement (Borich, 1994), uji normalitas Shapiro-Wilk, uji Paired Sample t-test atau Wilcoxon Signed-Rank Test, N-gain Hake, serta ukuran efek Cohen. Hasil penelitian menunjukkan LMS-PBL terintegrasi chatbot memenuhi kriteria valid, praktis, dan efektif. Pada aspek kevalidan, rerata validasi LMS mencapai 99,49% dengan kategori Sangat Layak dan PA antarvalidator sebesar 99,51% (reliabel), instrumen tes diagnostik four-tier memperoleh rerata 95,00% (Sangat Valid) dengan PA sebesar 96,73%, dan kuesioner self-efficacy memperoleh rerata 99,11% (Sangat Valid) dengan PA sebesar 98,47% serta reliabilitas Cronbach Alpha sebesar 0,994. Pada aspek kepraktisan, keterlaksanaan pembelajaran berada pada kategori Sangat Praktis dengan tren meningkat selama tiga pertemuan, sedangkan respons murid memiliki rerata 80,11% pada rentang 79,28% hingga 81,50% dengan kategori Baik. Pada aspek keefektifan terhadap KPM, hasil uji coba terbatas di kelas 7A menunjukkan peningkatan pretes-postes yang signifikan (p < 0,05) dengan rerata N-gain sebesar 0,719 (tinggi), sedangkan hasil uji coba luas juga menunjukkan peningkatan signifikan pada seluruh kelas (p < 0,05) dengan ukuran efek besar, yaitu N-gain sebesar 0,203 pada kelas 7C (rendah) dan 0,787 pada kelas 7D (tinggi) di SMP NU Bahrul Ulum serta 0,500 pada kelas VII-A (sedang) dan 0,458 pada kelas VII-B (sedang) di SMP IT Utsman bin Affan. Pada self-efficacy, skor total gabungan uji coba luas meningkat dari rerata 39,90 menjadi 48,17 (selisih 8,27), dengan uji Wilcoxon menghasilkan Z sebesar 7,170 dan ukuran efek r sebesar 0,710 (besar). Berdasarkan temuan tersebut, LMS-PBL terintegrasi chatbot pada materi ekologi berkonteks banjir terbukti valid, praktis, dan efektif untuk meningkatkan KPM serta self-efficacy murid kelas 7 baik pada uji coba terbatas maupun uji coba luas.

ABSTRACT

Yulandari, Silvi Novrian. 2026. Development of an LMS-PBL Integrated Chatbot for Ecology Learning to Enhance Seventh Grade Students' Problem Solving Skills and Self Efficacy. Thesis, Master Program of Science Education, Faculty of Mathematics and Natural Sciences, Universitas Negeri Surabaya. Advisors: Prof. Dr. Fida Rachmadiarti, M.Kes. and Mita Anggaryani, M.Pd., Ph.D.

Keywords: LMS, Problem Based Learning, chatbot, problem-solving skills, self-efficacy, ecology.

Students problem-solving skills (PSS) and self-efficacy in ecology learning in Gresik remain low, while the recurrent flooding of the Kali Lamong watershed, an issue close to students daily lives, has rarely been used as a learning context in science lessons. This study aims to develop a Problem Based Learning-based Learning Management System integrated with a chatbot (LMS-PBL) on ecology for seventh grade students that is valid, practical, and effective in improving students PSS and self-efficacy. The product was developed using the Hannafin and Peck model, which comprises three phases, namely needs assessment, design, as well as development and implementation. The resulting product is a Moodle-based LMS accessible at exploresains.com and equipped with a chatbot named BioBot or Deka. Product implementation was carried out in two stages using a One Group Pretest-Posttest design. A small-scale trial was conducted on 22 students of class 7A at SMP NU Bahrul Ulum to verify the operational feasibility of the product before it was expanded. A large-scale trial was then conducted on 80 students, consisting of 42 students from SMP NU Bahrul Ulum (classes 7C and 7D) and 38 students from SMP IT Utsman bin Affan (classes VII-A and VII-B), to examine the consistency of the product across different school contexts. Data were collected through validation sheets, learning implementation observation sheets, student response questionnaires, a four-tier diagnostic test for PSS, and a 17-item self-efficacy questionnaire. Data analysis included validity and practicality percentages, inter-rater reliability using Percentage of Agreement (Borich, 1994), Shapiro-Wilk normality test, Paired Sample t-test or Wilcoxon Signed-Rank Test, Hake N-gain, and Cohen effect size. The results show that the chatbot-integrated LMS-PBL met the criteria of validity, practicality, and effectiveness. In terms of validity, the LMS obtained an average validation score of 99.49% in the Highly Feasible category with an inter-rater PA of 99.51% (reliable), the four-tier diagnostic test reached an average of 95.00% (Highly Valid) with a PA of 96.73%, and the self-efficacy questionnaire reached an average of 99.11% (Highly Valid) with a PA of 98.47% and Cronbach Alpha reliability of 0.994. In terms of practicality, learning implementation fell into the Highly Practical category with an increasing trend across three meetings, while student responses averaged 80.11% within the range of 79.28% to 81.50%, categorized as Good. In terms of PSS effectiveness, the small-scale trial in class 7A showed a significant pretest-posttest improvement (p < 0.05) with an average N-gain of 0.719 (high), while the large-scale trial also demonstrated significant improvement across all classes (p < 0.05) with large effect sizes, namely N-gain values of 0.203 for class 7C (low) and 0.787 for class 7D (high) at SMP NU Bahrul Ulum, as well as 0.500 for class VII-A (medium) and 0.458 for class VII-B (medium) at SMP IT Utsman bin Affan. For self-efficacy, the combined total score from the large-scale trial increased from an average of 39.90 to 48.17 (a gain of 8.27), with a Wilcoxon test yielding Z = 7.170 and an effect size r of 0.710 (large). Based on these findings, the chatbot-integrated LMS-PBL on ecology flooding context is proven to be valid, practical, and effective in improving the PSS and self-efficacy of seventh grade students in both small-scale and large-scale trials.