Application of Log-Linear Model for Four Dimensional Contingency Table in Analysis of Students Anxiety in Mathematics

Abstract

The study applied log-linear mode for four contingency table in analysis of anxiety in mathematics among Nigerian students. The objectives are: to compare the log-linear models and select the best model, to determine whether mathematics anxiety level among university students is related to gender, course of study and mathematics performances using log-linear model and to examine the association  between  genders,  department  of  the  students,  their first  semester  mathematics performance  using  the  log-linear  model.  The population of the study is comprised of the undergraduate university students in Delta State, Nigeria. A total of one hundred and eighty (180) respondents constitute the sample size of the study and the data were collected using structured schedule. The independent model stated as;  log( μijKl) = λ + λi(M) + λj(A) + λK(G) + λl(D) was  found to be the best model out of the models under consideration as it has the least AIC and BIC values. And the null hypothesis H0 which states that there is no relationship among the variables under study was not rejected as the p-value is greater than the critical value as revealed by the goodness- of-fit test statistic table for the best model.  This indicates that mathematics anxiety among university students is independent of their gender, department and performance in mathematics.