forum.smartpls.com

Hi,

I have three moderating and/or two mediating variables in my study. Should I take the impact of these impacts as one construct,(moderating as one construct and mediating as one construct), or should I take a separate construct for each moderator and mediator variable? Or should I test each at one time?

May I have your valuable thoughts on this?

Thank you

Ruwan

Usually you want to keep them separated but all include them in one model.

However, sometimes it also may make sense to estimate different models. It really depends on what you want to test and what you want to understand.

Thank you very much for your comment.

My model predicts the investors' behaviour.

For example, If I want to test the moderation impact of investment amount and the macroeconomic factors, should I run the model with one moderation at a time or both simultaneously? Because it provides different results.

My other concern is,

Since I predict the behaviour I use RMSE or MAE value to decide the prediction ability. How can I use RMSE values of different models for comparison purposes? For example, if I want to compare the investors with lenders.

Thank you very much, and I appreciate your valuable guidance.

Ruwan

Hi,

Azzlin here. I have a question for running moderator analysis in smart pls 4. I have 4 moderators in 1 model.Is it right to run all moderators at once time in smart pls 4 since not appropriate to run all in smart pls 3 previously?

Another one question,Do I have to run pls algorithm instead consistent algorithm for moderating analysis as my model included all reflective constructs since smart pls 4 running two stage approach?Can results turn diferrently according to it?

Thank you in advanced!