Analysis of Rasch Model for the Validation of Chemistry National Exam Instruments

Abstract

Information about score obtained from a test is often interpreted as an indicator of the student's ability level. This is one of the weaknesses of classical analysis that are unable to provide meaningful and fair information. The acquisition of the same score if it comes from a test item with a different level of difficulty, must show different abilities. Analysis of the Rasch model will overcome this weakness. The purpose of this study was to analyze the quality of the items by validating the national chemistry exam instrument using the Rasch model. The research sample was 212 new students of the Department of Chemistry at the State University of Medan. The data collected was in the form of respondent's answer data to the 2013 chemistry UN questions, which amounted to 40 items multiple choice and uses the documentation method. Data analysis technique used the Rasch Model with Ministep software. The results of the analysis show the quality of the Chemistry National Exam (UN) questions is categorized as very good based on the following aspects: unidimension, item fit test, person map item, difficulty test level, person and item reliability. There is one item found to be gender bias, in which men benefit more than women. The average chemistry ability of respondents is above the average level of difficulty of the test items

Validation and Reliability of Lifestyle Instruments for Indonesian Adolescents with Hypertension Family History: A Rasch Model

“…The Rasch model analysis uses principal component analysis the residuals, which measures the extent to which the diversity of the instrument measures what it is supposed to measure. The unidimensionality requirement is that the raw variance explained by the measure is at least 20%, and the unexplained variance value does not exceed 15% [11], [12].…”

“…The Rasch model also has several advantages, because it fulfills the five principles of the measurement model, namely, (1) being able to provide equal intervals on a linear scale; (2) have the ability to predict the missing data; (3) can provide a more precise estimate because it does not only depend on the number of correct answers; (4) identification of error responses and guesses so as to detect model inaccuracies; and (5) produce replicable measurements [10], [11]. Therefore, analysis using the Rasch model can evaluate the strengths and weaknesses of the instrument properly [12].…”

Instrumen Literasi Digital Guru Menggunakan Model Rasch

“…Kata Kunci: guru, instrumen, literasi digital, model rasch Berdasarkan perannya yang signifikan dalam penelitian maka intrumen haruslah memiliki kualifikasi tertentu yang memenuhi persyaratan ilmiah (Imania & Bariah, 2019). Suatu instrumen yang baik akan mampu memberikan kepastian bahwa temuan yang dihasilkannya valid dan bisa diandalkan untuk memberikan penilaian yang akurat dan tepat terhadap kemampuan responden (Darmana et al, 2021). Sehingga sebelum digunakan dalam penilaian perlu dilakukan pengukuran ketepatan dan kelayakan instrumen.…”

Analisis Item Soal Kimia SMA Menggunakan Rasch Model

“…Menurut Arikunto (2013), suatu tes dianggap sah sebagai alat untuk mengevaluasi sesuatu apabila memenuhi kriteria validitas, reliabilitas, objektivitas, kegunaan, dan keterjangkauan. Suatu instrumen yang kompeten akan memastikan bahwa temuan yang dihasilkannya valid dan dapat diandalkan untuk memberikan penilaian yang akurat terhadap kemampuan siswa (Darmana, Sutiani, Nasution, Ismanisa, et al, 2021). Pendekatan analisis teori klasik merupakan salah satu yang sering digunakan oleh pendidik dalam proses penilaian kualitas instrumen penilaian.…”

“…Analisis yang dilakukan dengan menggunakan Rasch model tidak hanya sangat sederhana untuk dilakukan tetapi juga menghasilkan temuan analitis yang akurat. Selain itu, analisis item soal dengan menggunakan Rasch model memberikan sejumlah manfaat (Darmana, Sutiani, Nasution, Ismanisa, et al, 2021). Rasch model dapat mendeteksi jawaban yang salah dan skor data yang hilang, serta fakta bahwa abilitas tidak sepenuhnya bergantung pada jawaban yang benar (Eliza & Yusmaita, 2021).…”