|
|
Moderated Mediation with Categorical IV |
|
Message/Author |
|
|
Hello! SEM newbie and could really use some help!! I want to test a moderated mediation model with a categorical IV with 3 levels (catIV), two parallel continuous mediators (M1,M2), two continuous DVs (DV1,DV2), and one binary covariate (cov1). I created dummy variables for catIV, making the level 3 the reference level (dummy_level1, dummy_level2). Are these needed? Also, I don’t know how to set up the binary covariate? Can someone review what I've done below and give any guidance? USEVARIABLES ARE catIV M1 M2 DV1 DV2 cov1 !dummy_level1 !dummy_level2 ; MISSING ARE ALL (999); ANALYSIS: ESTIMATOR = ML; MODEL: DV1 ON M1 M2 catIV; DV2 ON M1 M2 catIV; M1 WITH M2; DV1 WITH DV2@0; MODEL INDIRECT: DV1 IND catIV; MODEL INDIRECT: DV2 IND catIV; OUTPUT: STDYX; SAMPSTAT; RESIDUAL; MODINDICES; |
|
|
Q1: Yes; they replace catIV. Q2: Nothing special. Input looks ok but there is no reason to have DV1 WITH DV2 fixed at zero. |
|
|
Thank you and thank you! Sorry if these questions are elementary... Would I add the binary cov1 in the model section like below, or does it go somewhere else? DV1 ON M1 M2 catIV cov1 |
|
|
When I include my binary covariate in my ON statement as describe above, I get a message saying cov1 is a binary variable but is being treated like a continuous variable. What is the right way to add this variable into the analysis? |
|
|
Without seeing the full output, I would venture a yes. |
|
Back to top |
|
|