Kualitas Butir dan Estimasi Kemampuan Matematika Siswa SMP pada Soal Ujian Sekolah

• Main Authors: Novi Indriyani Kones, Raden Rosnawati
• Format: Article
• Language: Indonesian
• Published: Universitas Hamzanwadi 2021-07-01
• Series: Jurnal Elemen
• Subjects: characteristic items; online mathematic assessment; student ability
• Online Access: http://e-journal.hamzanwadi.ac.id/index.php/jel/article/view/3054

Mathematics school exams as summative assessments are expected to have good quality, so the description of students' abilities is under the actual conditions. However, due to COVID-19, math school exams are carried out online, which results in bad quality of instruments, so analysis of math exam questions is needed for information about the quality of questions, and an overview of students' abilities is obtained. This study aimed to analyze the quality of the questions, which included analyzing the characteristics items seen from the level of difficulty, differences, distractions, and estimation of students' abilities with the item response theory approach. This study employs a quantitative research method with an exploratory, descriptive approach. The subjects of this study were 758 grade 9 students of State Junior High School who took mathematics school exams in Gunung Jati Sub-district, Cirebon Regency. Data from student's responses or answers from mathematics exam questions in the form of multiple-choice with 40 items for the 1st school, 35 for the 2nd school, and 20 for the 3rd school. The research data fulfills the IRT's assumptions first. Then, the researchers found that from the online assessment analysis with SPSS software, the R and Ms. Excel program, it showed that the level of difficulty there were still some items that needed to be revised or removed if not needed. The difference between the items in the three schools could be said to be good even though the distribution was not evenly distributed. The effectiveness of the distractor is more dominant in having a bad distractor with a percentage of 100% and 65.7% and the ability of students to answer math exam questions by obtaining the highest ability information from different students from the three schools.