Educators often monitor students' performance in class to make students aware of their weaknesses. Analysis on the relationship between educational settings and student performance can be useful in predicting student performance in a class. In addition, this analysis is able to help identifying the key indicators that may affect the students' final grade. In this paper, we present the initial work on the development of a predictive model that can predict student performance in a class to assist lecturers in improving student's learning process. We identified the predictor variables that can be used in our predictive model. The predictor variables of this model are based on attributes from different educational settings such as coursework marks, psychosocial factors and Course Management System (CMS) log data. These variables are collected from an advanced programming course in an institute of higher learning in Malaysia. This study provides a theoretical model that shows how data from different educational settings can contribute in the prediction of student's final grade. The results indicate that coursework marks has the most significant positive relationship with the student's final grade followed by total number of materials downloaded from CMS. © 2012 Infonomics Society.