LAUSR

The paper explores one of the current issues of distance education - the quality of computer tests in terms of ensuring objective control of knowledge. This issue is especially important in today's pandemic and temporary quarantine. The main attention is paid to a statistical analysis of test quality based on test results using CTT and IRT methods. Using modern statistical methods, the authors analyzed the results of testing prepared and conducted during the quarantine period. As an object of study, a test on “Integration of functions of one variable” was chosen, which students completely mastered remotely. The tests were created on the basis of the MOODLE platform at Igor Sikorsky Kyiv Polytechnic Institute, by proffessors of the Department of Mathematical Analysis and Probability Theory. Data processing is carried out using a system for test analysis, created by the authors in the programming environment R. The system allows you to process tests in different areas: pedagogy, psychology, sociology, etc., different in structure; use both CTT and IRT apparatus; work with large data sets; to analyze not only test questions, but also respondents; more accurately differentiate respondents. Based on the study, the possibility of conducting electronic testing remotely was confirmed. The technology used in the study can be used to create and analyze the tasks of external evaluation, conducting session control during quarantine. The use of the methods studied in the work for the analysis of test tasks will increase the competence of high school teachers to conduct electronic remote testing.