Modelling the Grade Point Average (G.P.A.): A Case study of the Postgraduate students of the University of AJK

Abstract

Considerable research has been undertaken on the grade point average (GPA) of
the students. In the present study, an attempt is made to forecast the GPA by
fitting a polynomial regression model on the GPA of the Masters level students of
the University of Azad Jammu and Kashmir, Muzaffarabad, Pakistan. The data
was found to be acceptable for the regression modelling after testing the
assumptions. The Best subset, backward elimination and stepwise regression
procedures were adopted to fit the model. Good of fit of the models is measured
by the coefficient of determination, i.e. R p 2 , R adj 2 , MSE and Mallow’s p C etc.
The model Yˆ  3.63 + 0.186X1 - 0.124X4 + 0.0246X6 with R p 2 , R adj 2 , MSE
values 71.1%, 70.6%, and 0.033 respectively is found to be the parsimonious
model. The results indicated that the three variables, i.e. study hours at home
(X1), sleeping hours (X4) and qualification of father (X6) significantly affect the
GPA of the Masters level students and provide sufficient information to forecast
the GPA of post graduate students of the said University.