Calculation of discriminativity and validity for evaluation of the quality of the pedagogical test from the course of mathematics on the topic "Ranks "

The last twenty years the Department of Higher Mathematics and Applied Informatics of Samara State Technical University has been using tests on all chapters of the Mathematics course to control students' knowledge. To understand how objectively a particular test allows a given assessment, it is necessary to examine its quality. Many theories are devoted to this problem. This article will analyze test quality and test tasks based on classical test theory. The study consists of the following stages: compiling a control test; conducting it on a sample of students that meets all the requirements of statistical analysis; processing of this sample and calculation of basic characteristics; analysis of the resulting coefficients, which allows us to conclude whether the test under study can be used as a control of the level of knowledge of students and what changes need to be made in order to improve it. This article analyses some characteristics when examining the quality of the test on the topic "Series" of the course of mathematics, namely, discrimination and validity. To analyze the validity of the test, point biserial correlation coefficients were found - these are correlation coefficients of some task with the student's individual score. After that, the total validity coefficient was calculated. As the calculations showed, the total validity coefficient is 0.53. This is a fairly high indicator, which cannot be said about some point biserial correlation coefficients, namely, for tasks No. 2 and 9. Therefore, these tasks must either be modified or replaced by others. Analysis of the discrimination coefficients of test tasks showed that their values satisfy the requirements of experts.