<![CDATA[Tujuan penelitian ini adalah untuk mendapatkan sebuah instrumen kemampuan berpikir reflektif matematis siswa Madrasah Aliyah yang valid dan reliabel. Penelitian ini merupakan sebuah penelitian pengembangan (Research & Development) dengan model pengembangan 4D Thiagarajan dengan tahapan define, design, development, dan dissemination. Data yang diperoleh pada penelitian ini merupakan data kualitatif berupa komentar atau saran perbaikan dari expert , dan data kuantitatif berupa skor yang diperoleh siswa pada saat uji coba instrumen tes ini. Uji coba penelitianya dilakukan kepada 60 siswa kelas 10 MAN 12 Jakarta untuk mengetahui validitas, reliabilitas dan tingkat kesukaran instrumen. Analisis hasil uji coba dilakukan dengan pendekatan model Rasch dibantu oleh software Winsteps. Penelitian ini menghasilkan 5 butir soal uraian valid, reliabel, dan tingkat kesukaran yang baik. Uji validitas menggunakan tabel item dimensionality dan tabel item fit order dimana Undimensionality berdasarkan nilai raw variance by measure dan kesesuaian butir soal sesuai salah satu kriteria, yakni 0,50 < MNSQ < 1,5; -2,0 < ZSTD < +2,0 dan 0,4 < Pt Mean Car < 0,85. Uji reliabilitasnya berdasarkan tabel summary statistic pada nilai person dan item reliability sebesar 0,73 dan 0,76 yang berkriteria cukup dan nilai alpha Cronbach sebesar 0,92 yang berkriteria bagus sekali. Lalu tingkat kesukaran instrumen terbagi tiga tingkatan, yaitu sulit, sedang dan mudah dengan menggunakan tabel item measure. Berdasarkan hal tersebut, maka instrumen tes ini layak digunakan untuk mengukur kemampuan berpikir reflektif matematis siswa Madrasah Aliyah. The purpose of this study was to obtain a valid and reliable instrument of mathematical reflective thinking ability of Madrasah Aliyah students. This research is a development research (Research & Development) with a 4D Thiagarajan development model with the stages of define, design, development, and dissemination. The data obtained in this study are qualitative data in the form of comments or suggestions for improvement from experts, and quantitative data in the form of scores obtained by students when testing this test instrument. The research trial was conducted on 60 10th grade students of MAN 12 Jakarta to determine the validity, reliability and level of difficulty of the instrument. The analysis of the test results was carried out using the Rasch model approach assisted by the Winsteps software. This study resulted in 5 items of valid, reliable, and good level of difficulty. The validity test uses the item dimensionality table and the item fit order table where Undimensionality is based on the raw variance by measure value and the suitability of the items according to one of the criteria, namely 0.50 < MNSQ < 1.5; -2.0 < ZSTD < +2.0 and 0.4 < Pt Mean Car < 0.85. The reliability test is based on the statistical summary table on the person and item reliability values of 0.73 and 0.76 which have sufficient criteria and Cronbach's alpha value of 0.92 which has excellent criteria. Then the instrument difficulty level is divided into three levels, namely difficult, medium and easy using the item measure table. Based on this, this test instrument is suitable to be used to measure the mathematical reflective thinking ability of Madrasah Aliyah students.]]>
