I am using the following mediate command in Stata18:
mediate (FundedDum GenderDum Loan_AvM Irregular TermLog TagTopDum1 AvLendM i.Sector_code, logit) (PersonalPlus, logit) (IndirectDum)
The outcome variable is a dummy variable equal to 1 if the loan was funded. The treatment variable is a dummy variable equal to 1 if the loan was made through a financial intermediary. The hypothesis is that loans made via this route remain attractive to lenders because they are small and for personal use. Thus, the mediator is PersonalPlus, a dummy variable representing personal-use loans.
The outcome of the above command is provided below.
My question is:
Does this model specification make sense, given that the general tendency is not to use a financial intermediary, so the direct effect is negative? However, these loans are still attractive to lenders seeking to finance small and personal loans, resulting in a positive indirect effect.
What is the interpretation of this?
mediate (FundedDum GenderDum Loan_AvM Irregular TermLog TagTopDum1 AvLendM i.Sector_code, logit) (PersonalPlus, logit) (IndirectDum)
The outcome variable is a dummy variable equal to 1 if the loan was funded. The treatment variable is a dummy variable equal to 1 if the loan was made through a financial intermediary. The hypothesis is that loans made via this route remain attractive to lenders because they are small and for personal use. Thus, the mediator is PersonalPlus, a dummy variable representing personal-use loans.
The outcome of the above command is provided below.
My question is:
Does this model specification make sense, given that the general tendency is not to use a financial intermediary, so the direct effect is negative? However, these loans are still attractive to lenders seeking to finance small and personal loans, resulting in a positive indirect effect.
What is the interpretation of this?
