engleski

Targeting additional effort for students’ success improvement: The highest effect group selection method

Goal of our research is to present a method for selection of a group of students. The selected group of students is supposed to receive additional teaching attention in order to improve their performance in the course. Guiding line used, for the group selection, is: value of an action is proportional to the benefit that produces to a customer. In our case customers are students. The selection method is based on multinomial logistic regression, Poisson-type discrete variables modelling the number of points achieved in a term exam, and transition probability matrix. We use data of a study progress during monitoring program in an undergraduate mathematical course. Demographic data and other attributes about previous performance were not included in the analyses. In the first part of the paper we present methodology, while in the second part we introduce data used for demonstration of the proposed method. At the end of the paper, an individual approach for the final selection is proposed. Criteria for selection is clear: increased probabilities of obtaining desired final grade for a student. Weighing of criteria is subjective depending on the goals of the decision makers.

learning analytics, predictive analytics, students’ success, multinomial logistic regression, Poisson distribution

nije evidentirano