Application of a Multiple Choice Model with Ordered Alternatives to Analysis of Objectivity of Knowledge Test Results

To assess the objectivity of the results of knowledge testing, these results are compared with some reference results, the objectivity of which is not questioned. For example, a survey conducted by an experienced teacher can be considered a benchmark for automated testing results. Methods of statistical analysis are used as comparison tools. Strictly speaking, classical regression analysis is not applicable to such problems due to the discrete nature of generally accepted assessments of knowledge. Therefore, a multiple choice model with ordered alternatives is used. This article contains the statement of the problem, its solution by the method of maximum likelihood and the results of the first studies.