Item analysis of National Geography Olympiad multiple-choice questions (MCQs) in Indonesia

Abstract

Analysis of questions is crucial in improving reused questions for subsequent tests; it also serves to eliminate items that do not fit the question framework. This research focuses on questions and test quality while exploring the relationship between the difficulty index of questions (p-value), the discrimination index of questions (DI), and the effectiveness of distractors in the questions (DE). The study was conducted on participants who took part in a geography Olympiad with a total of 259 students answering multiple-choice geography Olympiad questions from the year 2019. One hundred multiple-choice questions were provided as a competition in the field of geographical science. The internal consistency reliability of the test was 0.82, using the Kuder Richardson 20 coefficient (KR-20). The average score was 33.85 with a standard deviation of 9.85. The average difficulty index (p) and DI of the questions were 0.33 and 33%, respectively, indicating that a considerable number of items had moderate difficulty and high discrimination power. Effective distractors should be integrated into future tests to enhance the quality of questions and assessments. Regarding the discrimination index (DI), it can be observed that 10%–30% of test items fall within a range that is fairly good or acceptable with potential for improvement.

Keywords

• Geography Olympiad
• level of difficulty of questions
• differentiability of questions
• effectiveness of distracting questions
• Michael Hardman, University of Salford, UK
• Assessment