Psychometric Properties of Chemistry Multiple-Choice Paper for Science Students: An Extensive Analysis By Rasch Model

Abstract

An achievement test is an essential element in the teaching and learning process as its primary purpose is to measure student performance. However, high-quality test items require extensive time and effort to be produced, particularly multiple-choice questions. Teachers are known as content experts and test developers however some of them may lack knowledge in test development. Therefore, some of the constructed items may be flawed, biased, or unreliable to measure the students’ performance. The present study aims to determine the psychometric properties of newly developed multiple-choice questions for the Chemistry test paper using the Rasch Model. A study was conducted among 435 respondents from four randomly selected secondary schools in Klang Valley, Malaysia. Data were analyzed using a software called Winsteps. From the point of unidimensionality, Principal Component Analysis explained the test dimension of the instrument was moderate and acceptable, with 27.8% of raw variance measured. The reliability estimates for items were 0.99 while for person reliability is 0.87. multiple-choice paper. The aforementioned findings provide a dimension of validity of the test. In conclusion, the Rasch model has proven that the Chemistry test paper is a valid and reliable unidimensional instrument in measuring students’ ability and item difficulty.