Endogeneity in PLS-SEM
Posted: Mon Jun 21, 2021 11:52 am
Hello everyone,
I'm trying to use the Gaussian copula approach to detect endogeneity issues. I followed Hult, G. T. M., Hair Jr, J. F., Proksch, D., Sarstedt, M., Pinkwart, A., & Ringle, C. M. (2018). Addressing endogeneity in international marketing applications of partial least squares structural equation modeling. Journal of International Marketing, 26(3), 1-21.
My model has multiple endogenous variables. I read Park, Sungho, and Sachin Gupta (2012), “Handling Endogenous Regressors by Joint Estimation Using Copulas,” Marketing Science, 31 (4), pp. 567–86. with the consequence that I know that it is required to create the copula for each individual endogenous variable and adding all copulas to the original model. Unlike the example of Hult, et al. 2018 my research model contains 6 variables. Four of them are antecedents with each one arrow to either of the two mediator variables (full mediation). This means only the two mediator variables have arrows to the to target variables. But now my problem:
Do I have over 20 copulars or just 3? (Please compare the CUSL, COMP, CUSA, LIKE example to understand how I counted over 20 copluars for the whole model or just 3 if only the mediators and endogenous variables are relevant for the endogeneity test).
I would appreciate help with this baseline problem, thanks!
I'm trying to use the Gaussian copula approach to detect endogeneity issues. I followed Hult, G. T. M., Hair Jr, J. F., Proksch, D., Sarstedt, M., Pinkwart, A., & Ringle, C. M. (2018). Addressing endogeneity in international marketing applications of partial least squares structural equation modeling. Journal of International Marketing, 26(3), 1-21.
My model has multiple endogenous variables. I read Park, Sungho, and Sachin Gupta (2012), “Handling Endogenous Regressors by Joint Estimation Using Copulas,” Marketing Science, 31 (4), pp. 567–86. with the consequence that I know that it is required to create the copula for each individual endogenous variable and adding all copulas to the original model. Unlike the example of Hult, et al. 2018 my research model contains 6 variables. Four of them are antecedents with each one arrow to either of the two mediator variables (full mediation). This means only the two mediator variables have arrows to the to target variables. But now my problem:
Do I have over 20 copulars or just 3? (Please compare the CUSL, COMP, CUSA, LIKE example to understand how I counted over 20 copluars for the whole model or just 3 if only the mediators and endogenous variables are relevant for the endogeneity test).
I would appreciate help with this baseline problem, thanks!