Student Satisfaction of ETraining based Cluster Analysis: K-means & divisive hierarchical

Abstract

Student satisfaction is considered as one of the university education service quality indicators. We aimed to figure out how satisfied the students were with their attendance, learning environment, visuals materials, facilities, teacher skills and attitudes. A questionnaire with a total 41 items grouped in three were taken from 502 first-year, pre-medical students at the Mongolian National University of Medical Sciences. In the first group of questionnaire consisted of 10 items about how effective and interesting the lesson was; the second group 19 items were regarding to the teacher skills and attitudes, and the last 12 questions were asked how students participated in the lesson. Divisive Hierarchical Clustering method, the eclust function of the R program was employed. The satisfaction was assessed on five levels: “excellent”, “good”, “average”, “bad”, and “very bad”. In the study, the k-means method was 0.1-0.4 and the divisive hierarchical factor was 0.88-0.91. Therefore, we formed three clusters for classroom assessment (I), three for teacher evaluation (II), and two for student participation (III). The average grade is 33% in cluster I, 61.4% is good in cluster II, and 5% is bad in cluster III. The teacher evaluation model has a good grade of 63.9% in cluster II, an average grade of 31.8% in cluster I, and a poor grade of 4.1% in cluster III. The good grade in cluster I was 69.9% and the average grade was 30.07% in cluster II. The analytical study demonstrated that the divisive coefÏcient is more significant cluster model. The 61.4% was good in the subject evaluation model, 69.1% - good in the teacher evaluation model, and the student active and participation model was very good (31.2) and good (38.8).