Modelling of the highest level of children's education in the family using Multilevel Ordinal Logistic Regression Analysis

Abstract

Multilevel ordinal logistic regression analysis is a type of regression that modelling data with hierarchical structure and has an ordinal response variable. The Indonesian Family Life Survey (IFLS) produced data with hierarchical structure from individuals nested within sub-district. This research had three goals: first, to select the best model of the highest level of children's education in the family; second, to explain factors that affecting the highest level of children's education; third, to describe the characteristics of the highest level of children's education categories. The result showed that the best model was the random intercept model. The estimation variance of between sub-districts was 0.2314. The explanatory variables that have significant effects in the model were mother's education, number of household members, and number of household members who graduated from senior high school and status of residence. People who had high probability to go to university were coming from family, which the mother had at least senior high school background, most of family member graduated from senior high school, and they lived in Bogor city. In the other hand, people who had high probability to had only elementary school were coming from family which the mother had elementary school background, no members of family who graduated from senior high school, and they lived in Bogor Regency.