Cluster analysis of data of score rating system (based on the subject of “Mathematics”)

Score rating system is considered a rather effective assessment tool. At the same time, the analysis of a data file of a rating is very labor-consuming. Data processing by standard statistical methods often leads to the removal of average indicators that do not meet the requirement of an individual approach to students. Individual approach is of particular importance in differentiating clusters of students with low progress on a motivation scale at junior university courses in conditions of difficult demographic situation when there is a need to preserve the number of students. The aim of the paper is to differentiate the revealed clusters of initial group of students on a success degree (credit score-rating) by cluster analysis. The hypothesis is formulated: the score-rating data in dynamics of its formation has hidden information on the existence (or absence) of tendencies on transformation of the students’ clusters. The identification of such tendencies will allow us further to define the degree of stability of the similar clusters as characteristics of students’ motivation and degree of success of the training process. The results show that when the division into the clusters of “successful” and “unsuccessful” students is preserved, we can observe a rather stable cluster of “transition” students, which is difficult to reveal by standard statistical methods of the analysis. The cluster of “unsuccessful” students in its turn possesses its own structure of clusters similar to the structure of clusters of initial group. In terms of Pedagogy, this fact can confirm the hypothesis about the existence of dynamics in formally created clusters and identify the potential increase in a number of “successful” students. The revealed structure of a “transition” cluster is the target audience for the teachers to help these students to improve their results and transfer from the cluster of “unsuccessful” to the cluster of “successful” students.