Comparison of test items within and between faculties at the Indonesian Open University
Date
1993
Authors
Pakpahan, Sondang Purnamasari
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The purpose of this study was to investigate the differences in the quality of UT's (Universitas Terbuka's) test items in terms of item difficulty, discrimination, and test reliability, using the criteria proposed by Ebel and UT. A comparison was made between test items used previously and a sample of revised items drawn from the same sources. This investigation was conducted in two faculties, the Faculty of Education (FKIP) and the Faculty of Mathematics and Natural Science (FMIPA).
This study involved test results of three Physics and three English courses from FKIP, and two Mathematics and three Statistics courses from FMIPA drawn from three test administrations. The new tests consisted of revised and non-revised items drawn from the two old test administrations.
The major finding was that most of the test items in FKIP and FMIPA met UT's standards. FKIP and FMIPA had the same approximate qualities in item difficulty. However, the qualities of item discrimination and test reliability, were slightly better in FKIP. Most of the items in both faculties are still classified as 'hard' with marginal discrimination. Most of the tests in both faculties were also more difficult, less discriminating, and less reliable than testing experts posit as ideal, especially for science courses such as Physics, Mathematics, and Statistics.
There were three other major findings. First, item revisions in three of the six FKIP courses (Introduction to Quantum Mechanics, English for Arts and Science, and English for Education) resulted in desired changes in both the item difficulty and discrimination. Item revisions in the other courses, Alternating Current, only met the desired criterion for i tem discrimination. Second, item revisions in two of the five FMIPA courses (Calculus I and Survey Sampling Method) resulted in desired changes in both the item difficulty and item discrimination. Item revisions in the other course, Applied Experimental Design, only resulted in desired changes in item discrimination. Third, overall, a large percentages of revised and non- revised items in FKIP and FMIPA still had low discriminations (less than 0.30), implying that further item revision should be conducted in both FKIP and FMIPA.
Description
Keywords
UN SDG 10: Reduced Inequalities