Analysis of Computational Thinking Instrument for High School Student Using Rasch Model

Abstract

Listen

Translate article

This study is aimed-to construct and analyze the Computational Thinking Instrument at Physics subject using Rasch Measurement Model. The instrument was developed by the researcher to assess students’ computational thinking in Grade Eleventh in high school on heat and transfer topics under STEM Quartet Integrated Learning. The type of the instrument is a multiple choice with multiple-choice reasoning. This test consists of 13 questions to measure five concepts of Computational Thinking which are abstraction, decomposition, algorithm, evaluation, and generalization. The test was tried out on 120 students (87 female and 33 male aged 16-18 years old) in West Java and Banten. The item of the questions on the test was analyzed using Winstep 5.3.2.0. The reliability of the Instrument can be shown at the Cronbach’s alpha, Item Reliability, Person Reliability, Item Separation, and Person Separation. The result shows that Cronbach’s alpha (KR-20) is 0.75 which means the instrument has high reliability. The value of the item reliability is 0.95 and person reliability is 0.72. This means that although the consistency of student in answering questions is sufficient, the instrument has high reliability. The values of the item and person separation are 1.61 and 4.43 which means that the instrument has great separation. The validity of the instrument can be seen in unidimensionality and item fit order. The unidimensionality shows that the value of Raw Variance Explain by Measure is 32.8%, the Unexplained variance 1-5 contrast is below 15% and the eigenvalues are also below 3%. Overall, 8 of 13 items of the test meet 3 criteria of the item fit order, then 5 items have 2 of 3 criteria of the item fit order. We can conclude that this Instrument is reliable and valid. Then, the instrument can be used to measure students’ computational thinking on heat and transfer topics