Higher dimensional log-linear model and its application

• Main Authors: Usman , Mustofa, ., Widiarti, Ruswandi, Rudi, Elfaki, Faiz Ahmed Mohamed, Daoud, Jamal Ibrahim
• Format: Article
• Language: English
• Published: Sci.Int.(Lahore) 2015
• Subjects: QA Mathematics
• Online Access: http://irep.iium.edu.my/45781/1/paper_0_Mustofa_et_al_2015.pdf; http://irep.iium.edu.my/45781/; http://sci-int.com/index.php

: In this study the model of higher dimensional log-linear model is applied to four categorical variables in Education. The data are collected from the alumni data of University of Lampung, from 2010 to 2013 and about 9060 alumni involved. In this study, the variables of interest are: Length of Study with three categories (<4.5 years; 4.5- 5.5 years; and >5.5 years), Field of Study with three categories (Sciences, Social Sciences, and Education), Sex with two categories (Male, and Female), GPA in scale 0 to 4 with three categories ( <3.0, 3.0-3.5, and >3.5). In this study the aims are going to find the best model to explain the relationship among the factors. By using hierarchical Log-linear Model Analysis and backward method it was found that the best model for the data with three variables interactions in the model are: Length of Study*SEX*GPA, Length of Study*Sciences*GPA, and Sex*Sciences*GPA.