Direct effects and Indirect effects of mediator in multinomial logistic regression
Dear all,
I am doing a study on the effect of social network on housing channel choice for immigrants. The exogenous variable includes gender, age, immigrants' income, and the distance of immigrants' hometown to current city. Immigrants mobilize different sources including social network and internet during their housing process. Here I define two types of social networks: local network and hometown network. I also ask immigrants to choose the most important housing channels among: local housing agent, friend, and Internet.
I'd want to test the effect of two type of networks on their housing channel choice. Here networks are also endogenous，and can be thought as mediators. I use GSEM to run this multinomial logit regression. However, I don't know how to test the direct effects and indirect effects in GSEM. Here is the diagram for your refenrence. Is there any suggestions? Thanks a lot!
I am doing a study on the effect of social network on housing channel choice for immigrants. The exogenous variable includes gender, age, immigrants' income, and the distance of immigrants' hometown to current city. Immigrants mobilize different sources including social network and internet during their housing process. Here I define two types of social networks: local network and hometown network. I also ask immigrants to choose the most important housing channels among: local housing agent, friend, and Internet.
I'd want to test the effect of two type of networks on their housing channel choice. Here networks are also endogenous，and can be thought as mediators. I use GSEM to run this multinomial logit regression. However, I don't know how to test the direct effects and indirect effects in GSEM. Here is the diagram for your refenrence. Is there any suggestions? Thanks a lot!
