forum.smartpls.com

Hi,

I have a question about the calculation of the index of moderated mediation (Hayes, 2015).

If I have a moderated mediation model and all constructs in the model are latent variables measured with multiple items. In order to test the significance of the conditional indirect effect of the mediator, how do we calculate the index of moderated mediation? I understood the formula but have difficulty finding the data for the calculation.

After running bootstrapping (5000 subsamples) with the model, under the "final results"->"specific indirect effects"->"sample" button, I can find the coefficients of simple effect and moderating effect" for all 5000 bootstrapping cases. However, in order to calculate the data series of the index of moderated mediation, I need the latent variable scores of the moderator. To match with 5000 bootstrapping coefficients, I need all 5000 latent variable scores of the moderator. But I can't find the 5000 bootstrapping dataset in SmartPLS.

Could anyone tell me where I can find the latent variable scores of the moderator for 5000 bootstrapping cases? And the latent variable scores applied to the index of moderated mediation calculation should be standardized or non-standardized?

Thanks in advance!

LD

This paper should explain how to do it in PLS-SEM: https://dl.acm.org/doi/abs/10.1145/3505639.3505645

Thank you very much! I will review the paper you recommended.

The index of moderated mediation in SmartPLS, you can get from Specific indirect effect table at row : X*W->M->Y (X : IV, M : Mediator, Y:DV, W:moderator). From that table, we can get p value and Confident Interval as well.
you can prove Hayes formula :
w (Index of moderated mediation) = p2*p5
p2 : path coefficient M -> Y
p5 : path coefficient X*W -> M

Hope it will help