Small Area Estimation for Repeated Survey and Positively Skewed Distributions


Poster Session

Poster no.11

Anang Kurnia (Department of Statistics, Bogor Agricultural University, Indonesia) (Speaker)
Rahma Anisa (Department of Statistics, Bogor Agricultural University, Indonesia)
Siti Muchlisoh (Statistics Indonesia)
Dian Handayani (Department of Mathematics, State University of Jakarta, Indonesia)

This paper focuses on small area estimation for repeated survey data with positively skewed distribution. Nowadays, the demand for small area statistics which utilize cross-sectional and repeated measures data is increasing. For instance, Indonesian National Socio-Economic Survey is conducted twice a year. In our study, we consider repeated survey including the same census blocks at different time point. The aim of the proposed model is to borrow strength across small areas and repeated census blocks over time in case the variable of interest does not follow a normal distribution. However, it can be modeled by log-scaled transformation.  The model parameters will be estimated by maximum likelihood method.  The characteristics of our proposed model is evaluated by simulation study. We also apply our results for real data, specifically for Indonesia National Socio-Economic Survey data.

Keywords: Small area estimation, repeated survey, lognormal distribution, national socio-economic survey