Test Quality Through Item Analysis in Inferential Statistics

This study was carried out to assess the quality and effectiveness of fifty multiple-choice question (MCQ) test items with four options in a midterm examination in statistics among eighty-two, first-year Bachelor of Science in Psychology students at Bohol Island State University–Main Campus in Tagbilaran City during the second semester of academic year 2024–2025. The 27% of each upper and lower group were used to determine the performance and quality of each test item through item difficulty, discrimination index (DI), distractor effectiveness, and the relationships among these indices. The ratio of the test items was 82%, 14%, and 4% for easy, medium, and difficult items, respectively. For DI, 66% of the items were deemed acceptable (good and satisfactory), while 34% required revision or rejection (poor and bad). Also, 77% of distractors were functioning, while non-functioning distractors (NFDs) were more prevalent in less challenging items. There were twenty-two items rated as excellent, seventeen as good, ten as moderate, and only one as poor, reflecting generally strong item quality. Items with lower DI tend to have a higher number of ineffective distractors. The DI and distractor efficiency (DE) suggest that items with better distractors differentiated between high and low performers. Conversely, difficulty and discrimination indicated that easier items were less effective in distinguishing performance levels. Overall, the test demonstrated sound distractor quality and fair discrimination with an excellent reliability level. The construction and analysis of tests are recommended to improve assessment quality and ensure the accurate measurement of learning outcomes.