An automatic inference system for the quality analysis of test items based on the Bloom's revised taxonomy

Good Test items can be used to effectively measure the ability of testers. In this paper, an inference system is proposed to automatically analyze the quality for test items based on the Bloom's revised taxonomy. The main concept is to find the characteristics of knowledge dimension and cognitive process dimension for test items by analyzing the “verbs” and “question words” retrieved from the test items. After the system is completed, a test with 35 test items is analyzed by the system and an expert, respectively. The results show that the inference system can automatically and effectively analyze the quality of test items and reduce the load for generating a good test.