Estimasi kesalahan pengukuran standard setting dalam penilaian kompetensi matematika tingkat SMP di Kabupaten Sumbawa

• Main Authors: Weni Wendari, Samsul Hadi
• Format: Article
• Language: English
• Published: Universitas Negeri Yogyakarta 2018-06-01
• Series: Jurnal Penelitian dan Evaluasi Pendidikan
• Subjects: cut score; standard setting; generalizability theory
• Online Access: https://journal.uny.ac.id/index.php/jpep/article/view/16492
• ISSN: 1410-4725; 2338-6061

Penelitian ini bertujuan untuk mengetahui metode yang lebih akurat dalam mengestimasi kesalahan pengukuran standard setting pada metode Ebel, Bookmark, dan Contrasting group. Data penelitian ini merupakan dokumen Dinas Pendidikan dan Kebudayaan Kabupaten Sumbawa berupa respon peserta Ujian Nasional Matematika Paket P0C5520 tahun ajaran 2015/2016 yang berjumlah 352 siswa. Guru juga dilibatkan dalam penelitian sebagai panelis dalam Focus Group Discussion  (FGD). Data yang terkumpul kemudian dianalisis melaui tiga tahap.. Tahap pertama yaitu persiapan, kegiatan pada tahap ini mencakup penyiapan data, penggolongan SMP, dan penentuan karakteristik butir. Tahap kedua yaitu FGD dilakukan dalam dua putaran. Tahap ketiga yaitu mengestimasi kesalahan pengukuran dengan menggunakan pendekatan Generalizability Theory dengan bantuan program eduG. Hasil penelitian menunjukkan bahwa metode Contrasting group memiliki estimasi kesalahan pengukuran paling kecil dibandingkan metode Ebel dan Bookmark, oleh karena itu, metode Contrasting grouplebih akurat dibandingkan dengan dua metode lainnya.Kata kunci: cut score, standard setting, generalizability theory   MEASUREMENT ERROR ESTIMATION OF STANDARD SETTING IN MATHEMATICS COMPETENCY ASSESSMENT FOR JUNIOR HIGH SCHOOL IN SUMBAWA REGENCY Abstract This research aims to find the most accurate methods in estimating measurement error of standard setting among Ebel, Bookmark, and Contrasting group methods. The data used in this study were 352 students’ responses on Mathematics National Exam Package P0C5520 in the academic year of 2015/2016. The document was collected from the Department of Education and Culture in Sumbawa Regency. Teachers were also involved in this research as panelists in the Focus Group Discussion (FGD). The data collected were then analyzed through three stages. The first stage was preparation stage, including the activities of data preparation, school classification, and item characteristics analysis. The second stage was two-round FGD. The third stage was estimating the measurement error using Generalizability Theory approach assisted by eduG program. The research result shows that Contrasting Group method produces the smallest measurement error estimation compared to Ebel and Bookmark methods, therefore, Contrasting group method is considered as the most accurate method. Keywords: cut score, standard setting, generalizability theory