Model for Predicting Summative Performance from Formative Assessment Results of Engineering Students

The present paper proposes a new model for predicting the summative assessment marks (SAM) of engineering students based on their formative assessment marks (FAM). Often, instructors at higher education institutions have to provide learning to very large class sizes. Therefore, different approaches are sought to alleviate such constraints. One of the recent interests is the possibility of using FAM results to estimate or predict summative performance of students. In this study, a model was developed using data of modules lectured to undergraduate civil engineering students over a period of 10 years. The data sets used consisted of 18 assessment events of various modules involving 852 students. The linear and power type models were formulated through data fitting and validation. Statistical error indicators showed the power function model of the form SAM = P(FAM)X-Q with P = 3.2 and Q = 0.7 to be superior, as the model gave realistic predictions.