AUTOMATION OF TEST ITEMS QUALITY ANALYSIS AND IT STUDENTS TEST RESULTS ASSESSMENT

Abstract

The subject of this article is statistical models and methods for test items quality assessment, as well as processes and information support technologies for test results and test quality evaluating. The goal is to improve the quality of test items and the accuracy of the evaluation of test results in the context of constant changes in the content of study programs and the content of test tasks, and to reduce the time for test results processing through automation, and developed algorithms and an information system usage. The following tasks was solved: analysis of the test preparation and testing process; an analytical review of statistical methods for the test items quality assessment and the test items quality indicators; generalized process of the test preparation and conduct of testing development; and automation of IT students test results assessment. To solve these problems, the following methods was used: statistical analysis methods, the Rasch IRT (Item Response Theory) model, and software application development technologies were used. The following results are obtained. The process of the test preparation and testing was analyzed; its modification is proposed taking into account the peculiarities of IT students testing. The test task quality indicators, models and methods for it calculation were selected, in particular, the Rasch IRT model. The interpretation of the test item quality indicators values, and the calculation of the coefficients of the anomalies in the test item quality, what are the basis for the generation of an advisory solution to improve the test quality, are given. A prototype of an information system for support of processes of test items quality indicators evaluating, test results assessment and decision making about test modification has been developed. The system was tested during the module control of third-year students results evaluating. Conclusions: automation and information support of the test items quality indicators calculating and analyzing, and test results assessment will improve the accuracy of the assessment and reduce the time for test results evaluating.