Yenni Kurniawati, Anik Djuraidah, I. Made Sumertajaya, Anwar Fitrianto, Mira Meilisa, Nonong Amalita, Dina Fitria, Zamahsary Martha
The spatial-temporal model includes spatial autocorrelation and temporal autocorrelation in linear predictors. In general, the method used is the Conditional Autoregressive (CAR) which allows for residual spatio-temporal autocorrelation. Based on the Bayesian hierarchical modeling framework, the prior in this study used AR order 1. This study uses data on the number of child laborers in each province on the island of Sumatra in 2014-2017. The research aims to estimate the parameters of the ST. CARar model determines the influential factors and tests the effect of spatial-temporal autocorrelation on the ST. CARar model. The results showed that the variables that influence the number of child laborers on the island of Sumatra are the Open Unemployment Rate (OUR), Dropout Rate (DOR), and Illiteracy Rate (IR), with an estimated value of 0.44, -0.38, and -0.04 respectively. The relative risk value of the unemployment rate is 1.56, meaning that increasing OUR on the island of Sumatra during 2014-2017 risks an increase in the number of child laborers by 56%. While DOR and IR provide risks of 0.68 and 0.96 respectively for the number of child workers. This is due to the influence of the spatial-temporal autocorrelation on the model shown by the values of rho. S and Rho. T which are equal to 0.30 and 0.54 in the ST. CARar model. © 2026 Author(s).
Department of Statistics, Universitas Negeri Padang, Padang, Indonesia; Department of Statistics, Institut Pertanian Bogor, Bogor, Indonesia; Department of Engineering, Universitas Muhammadiyah Sumatera Barat, Bukittinggi, Indonesia
