A GLM Model for Predicting Students’ Academic Performance in Further Mathematics: A Multiple Logistic Regression - International Journal of Academic Research in Business, Arts & Science | IJARBAS

Issue: Issue: 2, , Volume: Volume: 1,, Year: Year: 2019.

A GLM Model for Predicting Students’ Academic Performance in Further Mathematics: A Multiple Logistic Regression

Date of Publication : 09, Aug, 2019

Author: Damianus Kofi Owusu (MSc)

Co Author: Gabriel Asare Okyere (PhD) , Isaac Akpor Adjei (MPhil)

Academic success in further mathematics is perceived to have a significant contribution towards the development of science and technology in any nation. The main objective of this study is to identify the major determinants of students’ academic success at WASSCE further mathematics. Two public schools in Hemang Lower Denkyira District constituted the study area with a sample of 84 students. Data for 2016/1017 WASSCE academic year group tracked from the documented records of the two selected schools in the District with 1:1 gender ratio were used for the study. Explanatory factors were; Mock examination, Age, Student Residence, School Location, and Gender on the target variable WASSCE grade. A three-stage stratified cluster sampling technique was used to select the two schools as well as classes and subject area under study. IBM SPSS Version 21 was used to analyse the data for the purpose to modelling multiple logistic regression with a binary response ‘WASSCE grade’ (upper grade/lower grade) against the systematic component of linear combination of predictor variables at 95% Confidence Interval. The study adopted the ex-post-facto research design. The model correctly classifies 77.4% of the overall cases indicating its prediction accuracy (robustness) and the extent to which it accurately predicted students’ WASSCE grades. AUROC = 0.823, and Hosmer and Lemeshow test had p=0.540 > 0.05 indicating goodness-of-fit of the model. Three (mock examination, school location, and gender) out of five predictors made significant contribution to model with no multicolinearity among the predictors. These variables are the major determinants predicting students’ WASSCE grade/performance in further mathematics. The study concluded that mock examination is reliable and thus should be centralized and supervised by the area education officer to make it more stringent, and there should be equity in resource allocation where both the less and well-endowed schools are equally supplied with teaching-learning resources, and finally, the education in its contents, designing and application should be gender solicitous.

 

DOI: 10.5281/zenodo.3364275

Sharing is caring!

Keywords : Academic Performance, Logistic Regression, At-Risk Students, Odds Ratio,

Serial: 1 Download Page: 1- 23