LOGISTIC MIXED MODELLING OF DETERMINANTS OF INTERNATIONAL MIGRATION FROM THE SOUTHERN ETHIOPIA: SMALL AREA ESTIMATION APPROACH

Abstract

The main objective of this study is to investigate socio-demographic and economic characteristics of a household on international migration and to estimate small area proportions at district and enumeration area level. Migration status refers to whether a household has at least one member who ever migrated abroad or not. A total of 2288 data are collected from sixteen randomly sampled districts in Hadiya and Kembata-Tembaro zonal areas, Southern Ethiopia. Several versions of the binary logistic mixed models, as special cases of the generalized linear mixed model, are analyzed and compared. The findings of the study reveal that about 39.4% of the households have at least one international migrant, and the rest 60.6% have no such migrants. Based on analysis of the generalized linear model and stepwise variable selection, four predictors are found to be significantly related to household migration status at 5% significance level. These are age, occupation, and educational level of household head and family size. Then twelve mixed models are analyzed and compared. The best fitting model to the data is found to be the logistic mixed regression model consisting of the six predictors with age nested within districts as random effects. Area or district specific random effect has variance of 1.6180. The district level random variation founded on final model with six predictor variables about the presence of migrant in the households such as the variation between districts is 33% and variation within the district is 67%. From analysis of the final model, it is found that the likelihood of a household of having international migrant increases with head's age and family size. An increase of family size by one person increases the log odds of having migrant by 0.131 indicating that large family size is one of the determinants for migration in the study area. The migration prevalence varies among the zones, the districts and the enumeration areas. Household characteristics: age, educational level and occupation of head, and family size are determinants of international migration. Community based intervention is needed so as to monitor and regulate the international migration for the benefits of the society.

Keywords: GLM, GLMM, Migration, Mixed Logistic, Small Area Estimation