Title: Estimation of regression coefficients when fixed effects and random effects are correlated
Authors: Chun-Shu Chen - National Changhua University of Education (Taiwan) [presenting]
Yung-Huei Chiou - National Changhua University of Education (Taiwan)
Abstract: In spatial regression models, when covariates and unobservable random effects are correlated, past researches have shown that ignoring this relationship would have significant influences on the estimation of regression coefficients but how to modify them remains an active research topic. To solve this problem, an idea of restricted spatial regression is used to ensure that the unobservable spatial random process is orthogonal to covariates. Then, an adjusted generalized least squares estimation method is proposed to estimate regression coefficients, resulting in estimators that perform better than conventional methods. Statistical inferences of the proposed methodology are justified both theoretically and numerically.