Hi all:
I can't figure out how to test mediation using bootstrapping in Stata. I used the sem command to create a model that I need to test for mediation. My model has three latent variables, two with three indicators and one with two indicators and two covariates. I have copied the code that I used to create the model below.
I ran the model, and then I tested it for mediation using the estat effects command. Using this command, it appears that I do have a fully mediated model. But my advisor prefers that I use bootstrapping and 95% confidence intervals.
I've used this example http://www.ats.ucla.edu/stat/stata/f..._mediation.htm to test for mediation in a path model. And I tried looking for help using the bootstrap command and the estate bootstrap command. Can someone point me in the right direction?
code that created my model:
sem (PM -> PM1, ) (PM -> PM2, ) (PM -> Re_Subst, ) (PM -> Re_PA_MR, ) (Re_Subst -> SU_MR1st_BY_Mean, ) (Re_Subst -> SU_MR2nd_BY_Mean, ) (Re_Subst -> SU_MR3rd_BY_Mean, ) (Re_Subst -> Re_PA_MR, ) (Re_PA_MR -> PA_MR1st_By_Mean, ) (Re_PA_MR -> PA_MR2nd_By_Mean, ) (Re_PA_MR -> PA_MR3rd_By_Mean, ) (Age_W1 -> PM, ) (Total_Rel_Length_MR -> Re_PA_MR, ), method(mlmv) latent(PM Re_Subst Re_PA_MR ) nocapslatent
I can't figure out how to test mediation using bootstrapping in Stata. I used the sem command to create a model that I need to test for mediation. My model has three latent variables, two with three indicators and one with two indicators and two covariates. I have copied the code that I used to create the model below.
I ran the model, and then I tested it for mediation using the estat effects command. Using this command, it appears that I do have a fully mediated model. But my advisor prefers that I use bootstrapping and 95% confidence intervals.
I've used this example http://www.ats.ucla.edu/stat/stata/f..._mediation.htm to test for mediation in a path model. And I tried looking for help using the bootstrap command and the estate bootstrap command. Can someone point me in the right direction?
code that created my model:
sem (PM -> PM1, ) (PM -> PM2, ) (PM -> Re_Subst, ) (PM -> Re_PA_MR, ) (Re_Subst -> SU_MR1st_BY_Mean, ) (Re_Subst -> SU_MR2nd_BY_Mean, ) (Re_Subst -> SU_MR3rd_BY_Mean, ) (Re_Subst -> Re_PA_MR, ) (Re_PA_MR -> PA_MR1st_By_Mean, ) (Re_PA_MR -> PA_MR2nd_By_Mean, ) (Re_PA_MR -> PA_MR3rd_By_Mean, ) (Age_W1 -> PM, ) (Total_Rel_Length_MR -> Re_PA_MR, ), method(mlmv) latent(PM Re_Subst Re_PA_MR ) nocapslatent
