Low R2 when reflective, high when formative?

[Mattias_O]

Hi - I have a couple of latent variables, where I have a weak theoretical idea whether they should be formative or reflective, theoretically I can go with either/or.

When I right-click in Smart-PLS and inverts the measurement models I find that the R2 is 0,3 when reflective and 0,6 when formative?

I guess that means that I go with formative.

But what does it mean when the difference is so large? Is there something wrong with the data? I have not previously experienced such a large difference between the modes?
:-)

[nayabeyes]

Hi - I have a couple of latent variables, where I have a weak theoretical idea whether they should be formative or reflective, theoretically I can go with either/or.

When I right-click in Smart-PLS and inverts the measurement models I find that the R2 is 0,3 when reflective and 0,6 when formative?

I guess that means that I go with formative.

But what does it mean when the difference is so large? Is there something wrong with the data? I have not previously experienced such a large difference between the modes?
:-)

I think the decision of using reflective and formative should be based conceptually.

Maybe you can read it for further clarification.

Becker, Jan-Michael, Kristina Klein and Martin Wetzels (2012): Hierarchical Latent Variable Models in PLS-SEM: Guidelines for Using Reflective-Formative Type Models, Long Range Planning (LRP), Vol. 45, Issue 5-6, 359-394.

[Hengkov]

Hi,

Reflective indicators is same with simple regression, for example
KSI -> X1
KSI -> X2
KSI -> X3
So, your result R^2 is small

Contrary, formative indicators is same with multiple regression, for example
Y1 -> ETA
Y2 -> ETA
Y3 -> ETA
So, your result R^2 high

Note: R^2 bias for number exogenous variable, so I recommend used adjusted R^2.


Best Regards,