回复楼上，改做了LR和Wald 检验
xsmle Lngdp Score Goveff SecondaryIndustry Investment Lnpopulation stateowned ,fe model(sdm) wmat(W) type(both) nolog noeffects #sdm模型
est store sdm_a
xsmle Lngdp Score Goveff SecondaryIndustry Investment Lnpopulation stateowned ,fe model(sar) wmat(W) type(both) nolog noeffects #sar模型
est store sar_a
xsmle Lngdp Score Goveff SecondaryIndustry Investment Lnpopulation stateowned ,fe model(sem) emat(W) type(both) nolog noeffects #sem模型
est store sem_a
lrtest sdm_a sar_a #LR test spatial lag
lrtest sdm_a sem_a #LR test spatial error
Wald 检验：
xsmle Lngdp Score Goveff SecondaryIndustry Investment , wmat(W) model(sdm) robust nolog effects fe r #在LR之后
(xsmle y x1 x2 x3 x4)
test [Wx]Score = [Wx]Goveff = [Wx]SecondaryIndustry = [Wx]Investment = [Wx]Lnpopulation = [Wx]stateowned =0 #test [Wx]x1 = [Wx]x2 = 。。
testnl ([Wx]Score =-[Spatial]rho*[Main]Score ) ([Wx]Goveff =-[Spatial]rho*[Main]Goveff ) ([Wx]SecondaryIndustry =-[Spatial]rho*[Main]SecondaryIndustry ) ([Wx]Investment =-[Spatial]rho*[Main]Investment ) ([Wx]Lnpopulation =-[Spatial]rho*[Main]Lnpopulation ) ([Wx]stateowned =-[Spatial]rho*[Main]stateowned )
#testnl ([Wx]x1 =-[Spatial]rho*[Main]x1 ) ([Wx]x2 =-[Spatial]rho*[Main]x2) ([Wx]x3 =-[Spatial]rho*[Main]x3)