Dear All,

In this topic, I'd like to highlight a problem in mediating analysis when there are several independent variables; the problem emerges from different researchers' points of view. Some researchers indicated that the analysis should be carried out using multi regression and with Sobel test, while others applied path analysis or handled with multi independent as linear regression, and yet others mentioned partial multi regression, SEM, and....etc.

My problem stems from the fact that I have several independents, one dependent, and one mediator. Furthermore, there is a correlation between the independent variables, which means that when I undertake the mediating effect test, whatever it is, the influence of the other independent variable should be eliminated. So, which approach is less error-prone or biased, and which method of analysis should I use? Any references would be much appreciated.