To Implement Computerized Adaptive Testing by Automatically Adjusting Item Difficulty Index on Adaptive English Learning Platform

Abstract

In previous language teaching studies, there were still some issues that have rarely been discussed or been of concern. One of the issues is the limitation to provide the appropriate test items according to the degree of the students’ capability in computerized adaptive testing. However, the difficulty level of a test item relies on expert judgment, which it is too time-consuming and expert-dependent. In order to effectively achieve automatic adjustment of the difficulty levels of test items to implement computerized adaptive testing, an attempt is made to address this issue by developing an adaptive English learning platform. This platform was used to analyze the difficulty levels of the items along with 162 individual ability values from the massive data of the 53,191 item answer records. The research results showed that analyzing the records using big data analysis and the item response theory enables the platform to automatically determine and adjust the difficulty levels of the items and to provide the fitting test items that match learner capabilities in computerized adaptive testing. Therefore, teachers do not have to manually adjust the difficulty levels of the test items. In addition, students can use computerized adaptive tests to improve their learning performance.