Multinomial Logistic Regression on the Effect of Drug Abuse among University Undergraduates

A number of social, economic and health problems have been associated with drug abuse (Khalsa et, al, 2008). Education is one of such sectors affected. Education in faith-based universities at times has a challenge in meeting parents’/guardians’ expectation which is usually high, despite that students of these universities are equally exposed to some of the problems affecting public university students. In this work, a statistical analysis to study the effect of drug abuse on students’ academic performance measured by the Cumulative Grade Point Average (CGPA) in a faith-based university is performed using multinomial logistic regression. Multinomial logistic regression is particularly applicable to this study because one of the dependent variables is multinomial. It was based on the profile of students developed on data collected through survey from students of the university. Descriptive analysis as well as tests of independence was carried out, supporting the multinomial logistic regression. The result showed that drug abuse is one of the factors that affect students’ CGPA. Students suggested that peer pressure is one vital reason for drug indulgence and by counseling, this can be reduced. It was suggested that drug abuse can be included as an explanatory variable in studies on student’s academic performance.Mathematics Subject Classification: 62J02Keywords: Multinomial Logistic Regression, Drug Abuse, Faith-based Universities, Academic Performance, independence.