Analisis Kualitas Tes dan Butir Soal Integral pada Evaluasi Formatif Matematika Teknik

Abstract

<p><strong>Abstrak: </strong>Analisis kualitas butir soal pilihan ganda merupakan alat penting untuk mengidentifikasi <em>item</em> yang dapat dipertahankan, direvisi, atau dikeluarkan. Tujuan penelitian adalah menguji kualitas <em>item</em> berdasarkan validitas, reliabilitas, tingkat kesukaran, daya pembeda, dan efek pengecoh. Metode penelitian berfokus pada <em>item analysis </em>dari 35 soal pilihan ganda yang dilakukan untuk 83 mahasiswa Pendidikan Teknik Bangunan. Sampel penelitian adalah <em>total sampling</em>. Analisis statistik menggunakan Microsoft Excel dan IBM SPSS. Teknik pengumpulan data melalui <em>LMS </em>spada.uns.ac.id. Analisis data menggunakan analisis deskriptif kuantitatif. Hasil analisis menghasilkan validitas isi 0,89, konstruk 97,14%, valid. Reliabilitas <em>Kuder-Richardson</em> 0,876 dan <em>Intraclass Correlation Coefficient</em> 0,880, reliabilitas tinggi. <em>Difficulty index</em> (DIF I) 28 (80%) soal memiliki tingkat kesukaran baik, tiga (8,57%) terlalu sulit, empat (11,43%) terlalu mudah, dan Mean ± SD 61,83% ± 16,61%. <em>Discrimination index </em>(DI) 35 (100%) soal memiliki daya pembeda yang dapat diterima hingga sangat baik dengan Mean±SD 53,90% ± 17,26%. <em>Distractor effectiveness</em> (DE) Mean ± SD 92,86% ± 17,75% dengan 92,86% distraktor fungsional secara keseluruhan. Dari keseluruhan hasil analisis disimpulkan butir soal tes integral pada evaluasi formatif Matematika Teknik memiliki kualitas soal yang baik untuk penilaian kognitif mahasiswa. Penelitian selanjutnya dapat dilakukan investigasi tentang korelasi DIF I, DI, dan DE untuk meningkatkan kualitas <em>item </em>pada bank soal.</p><p><strong><em>Abstract: </em></strong><em>The item analysis of multiple choice questions (MCQ) is essential for identifying items that can be retained, revised, or removed. The research analyzes items' quality based on validity, reliability, difficulty index, discrimination index, and distractor effectiveness. The study focuses on item analysis of 35 MCQ administered to 83 Building Engineering Education students. The research sample is total sampling—statistical analysis using Microsoft Excel and IBM SPSS—data analysis quantitative descriptive analysis. The study results had a content validity of 0.89 and a construct of 97.14%, which is valid—Kuder-Richardson reliability of 0.876 and Intreclass Correlation Coefficient of 0.880, high reliability. Difficulty index (DIF I) 28 (80%) items had a good difficulty, three (8.57%) were too difficult, four (11.43%) were too easy, Mean ± SD 61.83% ± 16.61%. Discrimination index (DI) 35 (100%) items are acceptable to excellent with Mean ± SD 53.90% ± 17.26%. Distractor effectiveness (DE) Mean ± SD 92.86% ± 17.75% with 92.86% functional distractors. The study concluded that the integral test in the formative evaluation of Engineering Mathematics had good-quality questions for students' cognitive assessment. Further research investigates the correlation of DIF I, DI, and DE to enhance the quality of items in the question bank</em>.</p>