American Journal of Theoretical and Applied Statistics 2013; 2(6): 255-267 Published online December 20, 2013 (http://www.sciencepublishinggroup.com/j/ajtas) doi: 10.11648/j.ajtas.20130206.24 Investigating predictors of examination result data using logistic regression (A case study of Imo State Polytechnic, Umuagwo, Imo State, Nigeria) Opara Jude 1 , Esemokumo Perewarebo Akpos 2, * , Iheagwara Andrew Ihuoma 3 , Okenwe Idochi 4 , OSUJI GEORGE A. 5 1 Department of Statistics, Imo State University, Owerri Nigeria 2 Federal Polytechnic Ekewe, Yenagoa, Bayelsa State, Nigeria 3 Department of Statistics, Imo State University, Owerri Nigeria 4 Department of Statistics, School of Applied Sciences, Rivers State Polytechnic, Bori, Rivers State Nigeria 5 Department of Statistics, Nnamdi Azikiwe University, PMB 5025, Awka Anambra State Nigeria Email address: judend88@yahoo.com (O. Jude), contactperes@yahoo.com (E. P. Akpos), andyiheagwara@yahoo.com (Iheagwara A. I.), nwonda@yahoo.com (O. Idochi), george.osuji99@yahoo.com (OSUJI G. A.) To cite this article: Opara Jude, Esemokumo Perewarebo Akpos, Iheagwara Andrew Ihuoma, Okenwe Idochi, OSUJI GEORGE A.. Investigating Predictors of Examination Result Data Using Logistic Regression (A Case Study of Imo State Polytechnic, Umuagwo, Imo State, Nigeria). American Journal of Theoretical and Applied Statistics. Vol. 2, No. 6, 2013, pp. 255-267. doi: 10.11648/j.ajtas.20130206.24 Abstract: This study tends to analyze the school examination results (scores) of 300 randomly selected students of Imo State Polytechnic, Umuagwo near Owerri, Imo State, Nigeria who offer English Language and Mathematics as general courses, using the binary logistic regression model with the aim of examining how some factors (variables) in secondary school level contribute to the performance of the students in the Polytechnic. The analysis is performed on the basis of the explanatory variables viz; gender, type of secondary schools, category of secondary schools, board of examinations and location of secondary schools, where scores of students in English Language and Mathematics are assumed to be the response variables. Applying the method of Correspondence Analysis revealed that there exist a significant correlation between board of examinations and location of schools, which made the analysis to be into two stages. The first stage is based on using English Language and Mathematics as a response variable with gender, type of secondary schools, category of secondary schools, and board of examinations as the explanatory variables. The second stage, on the other hand, English Language and Mathematics is the response variable, while gender, type of secondary schools, category of secondary schools, and location of schools are the explanatory variables. The odds ratio analysis compares the scores obtained in two examinations viz English language and Mathematics. The result of the analysis revealed that females are always showing best performances in Mathematics than English examination in all the two stages carried out in this paper. The study also showed that performances of students from girls’ schools are found to be the best in English Language course examination than those of students from boys; secondary schools. Furthermore, the study revealed that government schools always show better performance in English course examination than in Mathematics. Keywords: Odds Ratio, Wald Statistics, Logistic Regression Model, Correspondence Analysis 1. Introduction Neither two students’ nor two schools are identical. Students’ differ in gender, culture, religion, language, home environment, financial status of parents etc., whereas schools differ in size of students, quality of teacher, infrastructure, location of the school, aid provided by the government etc. Obviously performance of the students measured in terms of scores or grades obtained by them in examination varies from student to student and school to school. The variability in scores is a function of social climate which has to be studied and analyzed scientifically. The history of analyzing the students’ performance is as old as history of education. However formal presentation of analysis started around early thirties of the 20 th century. The performance measure corresponding to different