Thailand rice production analysis: Alternative approaches without P-value

Abstract

© Published under licence by IOP Publishing Ltd. This paper was contributed to seek a supporting evidence for the use of spatial analysis in panel data estimations. Yearly panel data regarding rice production and crucial factors, as well as quantities of rice planting areas, fertilizer usages, rice farmer families, and precipitations were observed as a time-series trend during 2009 to 2017. Methodologically, to avoid traditionally statistical assumptions like probabilistic significant test and to improve sensibility of model estimations, Bayesian statistics and inferences were applied to the entire methods of this paper. Stationary testing was accomplished by employing the Bayesian-ADF unit method. Additionally, the findings estimated via Bayesian panel regression models that indicated the spatial dummy added into the models also influenced both rice products and related factors. Thus, rice production needed to be regionally investigated and proposed critical issues should be practically considered.