Skip to content
Prev 27955 / 29559 Next

"spdep": check whether a spatial model fully controls for spatial correlation

On Sun, 22 Mar 2020, Gary Dong wrote:

            
Economics Letters 101, 282-284, referenced from ?spatialreg::Hausman.test, 
you should rather conclude that both models are critically misspecified, 
as the OLS/SLX and SEM/SDEM coefficients differ more than the test expects 
(null hypothesis no big differences, alternative big differences).
impacts() of the fitted SLX/SDEM model, without simulation as the standard 
errors of the impacts are calculated by linear combination. Typically use 
the summary() method of the impacts as otherwise documented. But since the 
model is badly misspecified, it looks as though you'll need to go back to 
your input data anyway, so impacts are premature.

Roger