Electronic Theses and Dissertation

Pandemi covid-19 yang melanda dunia selama dua tahun sudah mempengaruhi pendidikan di Indonesia. Dampak dari perubahan kebijakan dalam pendidikan selama pandemi covid-19 mengakibatkan adanya kemunduran siswa dalam keterampilan dan pembelajaran dasar yang dikenal dengan istilah learning loss. Salah satu cara yang dapat dilakukan untuk mengidentifikasi learning loss adalah dengan merancang soal Assessment Diagnostic Cognitive (CDA). Penelitian ini bertujuan untuk mendapatkan soal diagnostik materi pecahan yang berkualitas baik, serta mengidentifikasi learning loss materi pecahan siswa kelas VIII SMP di kota Banda Aceh. Penelitian ini menggunakan model penelitian pengembangan Tessmer dengan instrumen penelitian yang digunakan adalah lembar validasi, dan tes. Soal yang dikembangkan adalah soal dengan option bentuk politomus, yaitu berjenjang mulai dari D0 kesalahan paling berat, D1 kesalahan ringan, D2 kesalahan paling ringan, dan D3 kunci jawaban. Hasil validasi empiris menunjukkan 16 butir soal yang berkualitas baik, dengan reliabilitas untuk tiap atribut yaitu (A1) 0.9853, (A2) 0.9613, (A3) 0.9117, (A4) 0.9933, dan (A5) 0.9110. Analisis learning loss pada identifikasi perorangan mendapatkan hasil bahwa, umumnya siswa memilih option level D0 dan D1, sehingga sebanyak 51.4% siswa mengalami learning loss.

Kata kunci: Tes Diagnostik, Learning Loss, Validitas Soal, Bilangan Pecahan.

The Covid-19 pandemic that has hit the world for two years has affected education in Indonesia. The impact of policy changes in education during the Covid-19 pandemic resulted in a decline in students' skills and basic learning known as learning loss. One way that can be done to identify learning loss is to design Cognitive Diagnostic Assessment (CDA) questions. This study aims to obtain diagnostic questions about good quality fraction material, as well as identify learning loss in fraction material for Grade VIII students of junior high schools in Banda Aceh. This study uses the Tessmer development research model with the research instruments used are validation sheets and tests. The questions developed are questions with polytomous form options, namely tiered starting from D0 the most serious error, D1 the lightest error, D2 the lightest error, and D3 the answer key. The results of the empirical validation showed 16 items of good quality, with reliability for each attribute namely (A1) 0.9853, (A2) 0.9613, (A3) 0.9117, (A4) 0.9933, and (A5) 0.9110. Analysis of learning loss on individual identification shows that students generally choose option levels D0 and D1, so that as many as 51.4% of students experience learning loss. Keywords: Diagnostic Test, Learning Loss, Problem Validity, Fraction..