ESTIMASI KESALAHAN PENGUKURAN DALAM BIDANG PENDIDIKAN BERDASARKAN TEORI TES KLASIK

Abstract

Measurement results play an important role in providing information about the achievement of student learning outcomes. Unfortunately, accurate and objective information in the measurement of psychological aspects is not easily obtained because of measurement errors. These measurement errors need to be known by the teacher so that the measurement results can be interpreted appropriately. To determine the magnitude of measurement errors, the indicator used by classical test theory is the reliability index. The higher the reliability index, the lower the measurement error; and the smaller the reliability index, the higher the measurement error. This paper aims to describe the approach in calculating reliability index in order to find out the measurement errors of the measuring instrument used. The method used in this study is the study of literature, which is to examine various literature related to measurement errors and reliability of measuring instruments. In estimating measurement errors, there are three approaches that can be applied to estimate the amount of reliability index of a measuring instrument. That approach is the test-retest, the parallel-forms approach, and the internal consistency approach. Each approach is developed in accordance with the nature and function of the measuring instrument concerned with practical aspects in mind.