Abstract:

Choosing a degree course at tertiary level of education is not an easy task. It requires careful analysis of one’s desire for a particular field and the skillset that’s required for that field of study. Due to growing population rate, it is becoming more competitive to get admission in reputable university. Entrance test is the primary tool that is used by most of the universities in Pakistan to filter out the students with desired skill and knowledge set and selects the best pool of students, across the country. For that reason, universities should continuously improve the process of selection and the construction of Entrance test items. The purpose of this research is to check the reliability of Entrance tests format that were consistently taken over last two years by a private sector university in Pakistan. We analyzed, difficulty index, discrimination index and distractor strength for all the items in last two year’s data of about one thousand students. The Tests were paper based; Multiple Choice Questions (MCQs) type and the responses from answer sheets were extracted using Image processing techniques. Moreover, we analyzed the internal consistency of the test using Cronbach's alpha. The results of item analysis showed that the difficulty levels and discriminability are not homogeneous and a strong criticism emerged from the quality of distractors. The items with least reliability were removed and recommendations were given to improve the internal consistency, reliability and how to balance the distractors strength to make the test more convergent.

Keywords:

Assessment, Entrance Test, Reliability.