Bayesian Survival Mixture Model on Years of Schooling in West Papua Province

Abstract

Education could be considered as one of the basic pillars to determine the performance indicator of a respective region. Year of schooling is one of the education indexes, which becomes the government's target in the 9-year compulsory education program. This index illustrates the importance of knowledge and higher-level skills. Meanwhile, West Papua Province as one of the youngest provinces in Indonesia is challenged to improve the quality of human resources, particularly in the underdeveloped regions. Therefore, it is important to identify the variables which influence the years of schooling in the West Papua province. Statistically, the type of data such as length of time is frequently used to be the survival analysis. Nevertheless, the distribution pattern of the response variables is difficult to be analyzed. For that reason, this study applied mixture model on years of schooling. Mixture model estimation leads to the complex statistical problems with a number of parameters. Bayesian methods accomplish the estimation through the simulation process of Markov Chain Monte Carlo (MCMC). The survival mixture model was formed based on the status of county. Rural areas were evidenced to give the contribution of years of schooling distribution more than urban area up to 59.87 percent. The opportunity to obtain formal education at least to junior high school in urban areas was greater than rural area had, yet it went down faster in year 12-th or in senior high school level. In general, the factors which influenced the years of schooling in urban and rural areas turned out to be different.