MEDIATION ANALYSIS

Feel free to post your requests for new SmartPLS features and other suggestions.
Post Reply
saadinho7
PLS Junior User
Posts: 1
Joined: Fri Feb 22, 2019 9:17 pm
Real name and title: ALAOUI SAAD PHD STUDENT

MEDIATION ANALYSIS

Post by saadinho7 » Mon Mar 04, 2019 6:16 pm

Greetings
My model consists of 3 independent variables, 2 dependent and mediating variables and 2 dependent variables.
My 3 independent variables have effects on both mediating variables. And there is a relationship between my two mediating variables.
My problem is at this level: when I test my hypotheses with the presence of the relationship between my two mediating variables. The hypotheses concerning the relationship between my independent variables and the first mediator variable are validated while the hypotheses concerning the second mediator variable are rejected.
When I remove the relationship between my two mediating variables, all the hypotheses are validated. This confirms a full mediation (indirect relationship) of the first mediator variable between the independent variables and the second mediator variable.
My question is this:
Can we partially test the model (before the mediation relationship) to validate the rejected hypotheses? or not?
The hypotheses that have been rejected because of mediation; are they considered rejected or validated (before mediation)?
Thank you for your feedback and advice

Post Reply