Dear Fora,
Great thanks in advance for your helpful replies!
I have a nested model. The baseline model consists of 7 latent variables and 1 observed continuous variable (moderator); there are 10 direct paths and 2 curvilinear paths. I’d like to assess (post-hoc) if the addition of 5 moderation effects explains incremental variance in the model. The moderation effects are modeled to influence the relationship between different predictor (3) and criterion variables (3). *All latent variables are interval variables. The same variable is used as the moderator for each of these 5 relationships.
I’m aware of the F-test and that it is used to make comparisons between models. But I’m not sure how to assess if the addition of the moderation effects explains additional variance here. All moderator effects are not significant (but 2 are significant) but to me it makes sense to test all at once since this is a post-hoc examination and then trim the model.
Can you please advise how I can compare the two models to determine if the addition of the moderation effects explains incremental variance?
Thanks again!