YUNILDA, HENI (2025) PENGEMBANGAN INSTRUMEN TES DIAGNOSIS KOGNITIF MATEMATIKA SEBAGAI PERANGKAT IDENTIFIKASI AWAL SLOW LEARNERS. Doktoral thesis, UNIVERSITAS SULTAN AGENG TIRTAYASA.
![[img]](https://eprints.untirta.ac.id/style/images/fileicons/text.png) |
Text (DISERTASI)
HENI YUNILDA_7782220029_FULLTEXT.pdf Restricted to Registered users only Download (4MB) |
|
Text
HENI YUNILDA_7782220029_01.pdf Restricted to Registered users only Download (1MB) |
|
Text
HENI YUNILDA_7782220029_02.pdf Restricted to Registered users only Download (627kB) |
|
Text
HENI YUNILDA_7782220029_03.pdf Restricted to Registered users only Download (687kB) |
|
Text
HENI YUNILDA_7782220029_04.pdf Restricted to Registered users only Download (999kB) |
|
Text
HENI YUNILDA_7782220029_05.pdf Restricted to Registered users only Download (333kB) |
|
Text
HENI YUNILDA_7782220029_Reff.pdf Restricted to Registered users only Download (315kB) |
|
Text
HENI YUNILDA_7782220029_Lamp.pdf Restricted to Registered users only Download (2MB) |
Abstract
Low mathematics achievement among some senior high school students is often not identified at an early stage due to the lack of systematic and standardized diagnostic instruments. Therefore, this study aims to develop a Cognitive Diagnostic Mathematics Test Instrument as an early identification tool for slow learners. The development followed the Four-D Model, consisting of the following stages: (1) test specification formulation, (2) initial item construction, (3) limited trial, (4) item revision, (5) field testing, (6) technical analysis, (7) final instrument development, and (8) printing and distribution. Content validity was assessed using Aiken’s V index, which showed that all items demonstrated good validity (V ≥ 0.80). Readability testing and item analysis based on Classical Test Theory were conducted during the limited trial phase involving 40 students from SMAN 3 Kota Serang. The results were used to revise the items prior to field testing with 886 students from four senior high schools in Jakarta, Kota Tangerang, and Kabupaten Serang. The two-parameter logistic Item Response Theory (IRT 2-PL) analysis indicated that most items had high discrimination and appropriate difficulty levels for diagnostic purposes. To ensure construct validity, Kaiser-Meyer-Olkin (KMO) and Confirmatory Factor Analysis (CFA) were performed. The KMO results demonstrated that the data met the required sampling adequacy, with overall MSA values ranging from meritorious to marvelous (0.72–0.93). Meanwhile, CFA results revealed that most subdomains fit the recommended goodness-of-fit indices (CFI, TLI ≥ 0.90; RMSEA ≤ 0.08; SRMR ≤ 0.08), supporting the theoretically established factor structure. The final instrument consists of 105 test items (113 sub-items) distributed across five major domains: number, algebra, measurement, geometry, and data analysis and probability. The implementation trial involved 300 students from SMA Garuda Cendekia to assess the instrument’s accuracy. The analysis indicated that a cut-off score of ≤ 85 yielded a sensitivity of 80%, specificity of 65%, and overall accuracy of 72%. Students scoring ≤ 85 are classified as low-achieving and potentially slow learners, further divided into types 1, 2, and 3. The implementation results showed that type 3 achieved 100% accuracy, type 2 achieved 71%, and type 1 achieved 83%. Thus, the Cognitive Diagnostic Mathematics Test Instrument developed in this study has been proven to be valid, reliable, and effective for the early identification of slow learners.
| Item Type: | Thesis (Doktoral) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
|||||||||
| Additional Information: | Rendahnya capaian belajar matematika pada sebagian siswa jenjang SMA sering kali tidak teridentifikasi secara dini akibat keterbatasan instrumen diagnosis yang sistematis dan terstandar. Oleh karena itu, penelitian ini bertujuan untuk mengembangkan instrumen Tes Diagnosis Kognitif Matematika sebagai perangkat identifikasi awal slow learners. Model pengembangan yang digunakan dalam penelitian ini adalah Four-D Model, dengan tahapan: (1) penentuan spesifikasi tes, (2) penyusunan butir soal awal, (3) uji coba terbatas, (4) revisi butir soal awal, (5) uji coba lapangan, (6) analisis teknis, (7) penyusunan instrumen final, serta (8) pencetakan dan distribusi instrumen tes. Uji validitas isi dilakukan melalui indeks Aiken’s V, yang menunjukkan bahwa seluruh butir soal memiliki validitas yang baik (V≥ 0,80). Selanjutnya, uji keterbacaan dan analisis butir soal berbasis Classical Test Theory dilakukan pada tahap uji coba terbatas dengan melibatkan 40 siswa SMAN 3 Kota Serang. Hasil dari uji coba terbatas digunakan untuk merevisi butir soal sebelum dilakukan uji coba lapangan dengan 886 siswa dari empat SMA di Jakarta, Kota Tangerang, dan Kabupaten Serang. Hasil analisis Item Response Theory dua parameter logistik (IRT 2-PL) menunjukkan bahwa sebagian besar butir soal memiliki daya beda yang tinggi dan tingkat kesukaran yang sesuai untuk tujuan diagnosis. Untuk memastikan validitas konstruk instrumen, dilakukan analisis Kaiser-Meyer-Olkin (KMO) dan Confirmatory Factor Analysis (CFA). Hasil KMO menunjukkan bahwa data memenuhi syarat kecukupan sampel dengan nilai MSA keseluruhan berada pada kategori meratifikasi hingga sangat baik (0,72–0,93). Sementara itu, hasil CFA menunjukkan bahwa sebagian besar subdomain memiliki model yang sesuai dengan kriteria fit indeks yang direkomendasikan (CFI, TLI ≥ 0,90; RMSEA ≤ 0,08; SRMR ≤ 0,08), mendukung struktur faktor yang telah ditetapkan secara teoritis. Instrumen final terdiri dari 105 butir soal (113 item) yang tersebar dalam 5 domain utama, yaitu bilangan, aljabar, pengukuran, geometri, serta analisis data dan peluang. Uji implementasi dilakukan pada 300 siswa SMA Garuda Cendekia untuk menilai akurasi instrumen. Hasil analisis menunjukkan bahwa cut-off score ≤ 85 memiliki sensitivitas 80%, spesifisitas 65%, dan akurasi 72%. Siswa dengan skor ≤ 85 dikategorikan sebagai siswa berprestasi belajar matematika rendah dan dikelompokkan lebih lanjut menjadi slow learners tipe 1, 2, dan 3. Hasil implementasi menunjukkan bahwa kategori tipe 3 memiliki tingkat akurasi 100%, tipe 2 sebesar 71%, dan tipe 1 sebesar 83%. Dengan demikian, instrumen Tes Diagnosis Kognitif Matematika yang dikembangkan dalam penelitian ini telah terbukti valid, reliabel, dan efektif dalam mengidentifikasi awal slow learners. | |||||||||
| Uncontrolled Keywords: | cognitive diagnostic test, Item Response Theory, mathematical ability, slow learners in mathematics, test instrument development Item Response Theory, kemampuan matematika, lamban belajar matematika, pengembangan instrumen tes, tes diagnosis kognitif | |||||||||
| Subjects: | L Education > L Education (General) L Education > LB Theory and practice of education > LB1603 Secondary Education. High schools |
|||||||||
| Divisions: | 08-Pascasarjana 08-Pascasarjana > 7782-PENDIDIKAN (S3) |
|||||||||
| Depositing User: | Heni Yunilda Hasibuan | |||||||||
| Date Deposited: | 14 Jul 2025 09:15 | |||||||||
| Last Modified: | 14 Jul 2025 09:15 | |||||||||
| URI: | http://eprints.untirta.ac.id/id/eprint/50822 |
Actions (login required)
 |
View Item |