The Relevance of Student Differentiation in the Modern System of Continuing Education

The rapidly evolving digital economy creates a need for a transformation of the education system. In modern conditions, the requirements for both the student and the teacher are changing. The teacher, in order to improve his/her professional skills, needs to enhance and take courses in continuing education. At Ufa State Petroleum Technical University, courses for faculty on modern educational technologies are organized. Each student is given the opportunity to acquire competencies for the independent formation of a course in their discipline in the educational environment. When implementing the program of the refresher course, one should take into account the known heterogeneity of the student population. The authors put forward the following research hypothesis: similar sets of students in the course can be divided into stable clusters that form stable patterns with homogeneous characteristic features. The study examined a sample of students of continuing education courses of 187 people. Initially, the study of the relationship between the sample parameters using standard statistics did not lead to obvious results. It was hypothesized that there were differences in the behavior of students of different age, sex, etc. Data for analysis requires preliminary processing of the initial set. The set of all students can be conditionally divided into subgroups, a typical representative - the pattern - is then singled out. Identifying a typical representative makes it possible to distinguish his/her typical behavior or result of a certain activity. Further, the sample was divided into clusters in the STATISTICA package. The clustering methods used were k-means and EM. This choice is due to the following advantages of the methods: relative ease of use; clustering visualization; convergence of the algorithm (impossibility of looping). The number of clusters is not known in advance; the EM algorithm is poorly applied with weak differentiation of the initial set into clusters (comparison with the results of the k-means method); the results are volatile. To verify the adequacy of the results obtained, clustering was performed using the maximum likelihood method. The results obtained for clusters of the highest differentiation in this case are almost identical. Thus, the approach proposed by the authors to identify sustainable groups allows forming patterns of course participants. By identifying clusters, teachers that conduct training are given the opportunity to highlight the characteristic features of these patterns. The results of the proposed approach can be used in predicting the success of training, as well as for choosing a methodology for controlling the impact on students by the teacher.

Keywords

Authors

References