The quality of assessment tools matters. Hence, this study is intended to assess the quality of test items and their overall properties. The study was conducted using311answer-sheets of licensing examination for new English teachers in selected Ethiopian universities. An Exam-sheet that contains 100 questions constructed in multiple choice formats. Statistical analysis was made to compute Item difficulty, item discrimination index and distractors efficiency. Point bi-serial correlation was used to check the reliability of the items. The results of this study indicated that the mean score of the examinees was 52.62 Â± 11.58, the mean p value and DI were 53 Â± 21% and 0.29 Â± 0.1 9, respectively. The DI was noted to be moderate at p value range above 60%. Having combined the two indices, 54(54 %) of the items could be considered to be good having a p-value from 31% to 60%. Overall 47% items had non-functional distractors, while 53% items had functional distractors. Mean DE was 66.63 Â± 33.35%. Satisfactory level of discrimination (DI = 0.413 and 0.305) was achieved with items having two and one nonfunctional distractors respectively while items with no nonfunctional distractors had lower DI (0.027) internal consistency reliability of the test as per KR20 is 0.9.Items having optimal level of difficulty and high discriminating power with functional distractors should be should be stored and used for subsequent assessments while the rest should be revised toimprove the quality of the assessment tools in general and the subsequent test items in particular.
Keywords: types of indices, power of distractors, item analysis, multiple choice questions, non-functional distractors