Multiple DVs: estimate jointly or independently?

[hariskr]

If I have a model with multiple dependent variables (DVs), will there be a difference in either path coefficients or standard errors if I estimate the model with these DVs independently vs. jointly? For example, if I have 2 IV with 2 DVs, what are the ramifications of estimating the model twice (once for each DV) vs. once with the path to both DVs active?

[cringle]

I am not sure if I understand your question. The results will change whenever you change the PLS path model since you change the context for the model estimation.

Best regards
Christian

[hariskr]

To clarify: should I be getting different path coefficient estimates by using method 1 vs method 2, as per the attached image?

[cringle]

Yes, definitely. The context and thus the LV scores of A changes in Model 1 and Model 2. As a result, the coefficients are different. You are looking in completely different models.

Best
Christian