Spatial Modeling of Fixed Effect and Random Effect with Fast Double Bootstrap Approach

The use of panel data on spatial regression has many advantages. However, testing the dependency and parameter presumption generated in becomes inaccurate when applied to regions with large numbers small units. One method overcoming problems unit sizes is bootstrap method. research aimed combine cross-section time-series data. analysis was performed extract information based observations modified by influences space or location, known as panels. influence location effects presented form weighting. Fast Double Bootstrap (FDB) modeling poverty rates Flores Island. results Hausman test show right model, which a random effect. Meanwhile, concludes dependence Island, more likely be Spatial Autoregressive Random (SAR) model. SAR effect value R2 77,38% does not meet normality assumption. FDB approach can explain diversity rate Island 88,64% meets residual assumptions. panels shows better than common