Special method is required for analyzing censored data with spatial dependence. Using linear regression will results in invalid parameter estimations, normality assumption violations, and obscure the model interpretation. Spatial Tobit regression model is used to analize the data of internet usage in Java. MCMC Gibbs sampler method with Bayesian inference approach was used for parameter estimation. As a result, internet usage in Java Island is influenced by the percentage of population living in urban areas, the percentage of population graduated from senior high school, the average length of school, the percentage of households with mobile phones, and the percentage of villages receiving cell phone signal.
                        
                        
                        
                        
                            
                                Copyrights © 2017