Stupid Question: Control Variable vs Multigroup Analysis

[arcticarrow]

HI sorry for this stupid question but I cannot find the answer.

if I want to test X --> Z, what's the difference between Control Variable (eg. Having gender as a single construct pointing to Z), and Multi-group analysis to see if the path_coefficient are singnificantly (t test bootstrapping) different for X --> Z across male and female.

THanks.

[Hengkov]

Hi,

Variable control has the function to control the DV variables (such as demographic variables). Whereas in PLS-MGA control as moderating variables.

Regards,

[arcticarrow]

Hi THanks for this, but I am still very confused, in particular,

In testing the hypothesis X --> Z, I am not doing MGA analysis here because Y is a latent variable measured by 6 reflective indicators.

My question is, should I Model Y as a moderating effect (product indicator) or should I just add an arrow from Y, pointing to Z and see if the R2 increase a lot (in this case it would be a control variable?)

So I am confused with Moderator and COntrol variable. Thanks

[Hengkov]

Hi,

Why you ask "Y" role as a moderating or control to someone else? Obviously knew it was you, in accordance with the model you get up, what you want to test and how your research goals. Control variable is demographics, rare latent form.
Variable control function to control the effects of extraneous variables, while moderating variables you must know.

Best regards,