Indirect Effect

[anhthyngo]

Is there a way to decompose the Natural Indirect Effect (NIE)?

Suppose I have mediator (M), an independent variable (X), and dependent variable (Y).

I think in your package, the NIE is the combination of the effect from X -> M + M -> Y, done through neEffdecomp().

Is there a way in medflex to extract the individual effects of X -> M and M -> Y?

[ducpq91]

You can just run the regression for each of the model above, and take their corresponding coefficients from summary().