Hello,
I have more of a conceptual question rather than a coding question. I am trying a multilevel moderated mediation model where I have two mediators (attn_spread and res_spread) in the relationship between geo_disp (IV) and rev_growth (DV).
I also have a moderator (abs_cap) in the path from res_spread to rev_growth. The simplified models look something like this
Mediator 1: attention spread
model_a1 <- attn_spread ~ geo_disp + X1 + X2 + X3
Mediator 2: resource spread
model_a2 <- res_spread ~ geo_disp + year + X1 + X2 + X3
Full Model
model_b <- rev_growth ~ attn_spread + res_spread * abs_cap + geo_disp + X1 + X2 + X3
I find that the main IV has a significant effect on the mediators , which is good. However, in the results of the full model, I find that the mediating variable res_spread is not significant but the interaction term abs_cap*res_spread is significant. How do I interpret this, can I still claim that res_spread is a mediator but its effect on the DV is conditional on the values of abs_cap ?
I have more of a conceptual question rather than a coding question. I am trying a multilevel moderated mediation model where I have two mediators (attn_spread and res_spread) in the relationship between geo_disp (IV) and rev_growth (DV).
I also have a moderator (abs_cap) in the path from res_spread to rev_growth. The simplified models look something like this
Mediator 1: attention spread
model_a1 <- attn_spread ~ geo_disp + X1 + X2 + X3
Mediator 2: resource spread
model_a2 <- res_spread ~ geo_disp + year + X1 + X2 + X3
Full Model
model_b <- rev_growth ~ attn_spread + res_spread * abs_cap + geo_disp + X1 + X2 + X3
I find that the main IV has a significant effect on the mediators , which is good. However, in the results of the full model, I find that the mediating variable res_spread is not significant but the interaction term abs_cap*res_spread is significant. How do I interpret this, can I still claim that res_spread is a mediator but its effect on the DV is conditional on the values of abs_cap ?
