How to Improve Test Performance by Moodle Statistics

Creation of properly balanced tests and their evaluation is an important part of the educational process, which is dealt with by the theory of testing. The most popular testing theories are Classical test theory and Item response theory. In this paper, we will focus on metrics used in the Item response theory, which deals with evaluating the performance of the test from the point of view of individual questions, through metrics such as the Discrimination index, Item difficulty, Random score estimate, and others. In the 2023/2024 academic term, 239 second-year students in the field of computer science participated in the Operating Systems course. After passing each test, we proceeded to evaluate the test performance based on the statistics available in the MoodIe system. We evaluated a total of 41 testing items, primarily focused on Discrimination index, Item difficulty, Intended weight and Efficiency weight. We identified 5 testing items, which do not perform well enough. In order to be able to evaluate the overall test performance, we decided to calculate the reliability of the test using the Kuder-Richardson-20 (KR20) metric. KR20 metric is used to measure the internal consistency reliability of a test. We calculated the KR20 value before adjusting all three tests, and also after adjusting the tests. In all three cases there was an improvement in the value of KR20 in the interval from 3.63% to 5.37%, which indicates higher internal consistency of all three tests. Conclusion: We evaluated test performance of 3 tests based on the MoodIe statistics. We identified five testing items that were shown not to work based on manual verification as well as on the MoodIe test statistics. We eliminated those testing items and observed an improvement in the test reliability calculated by KR20 metric, resulting in an improvement in test performance.