Path Model (SEM)

[PKA111]

Hello,

Can anybody please explain how to interpret the "Path Model" results?
In path models the structure is predefined (LV, MV and the Path), If not sure of the path and the grouping of LV and MV is there any way to figure out which MV should be part of a particular LV?

Thanks

 

[noetsi]

I have worked with path analysis in SEM but I don't know what you mean by LV and MV. We never used that terminology.

 

[Lazar]

I also work extensively with SEM and have no idea what those terms mean. At a guess I assume you mean manifest variable and latent variable. In which case you should be guided by theory and a careful consideration of whether the latent variable is formative or reflective.

 

[PKA111]

Hello Noetsi and Lazar,
LV would be Latent Variable and MV would be Manifest Variable.
I have been using PLS PM for some time now and am quite aware of the reflective model or formative model, am more concerned with the result interpretation. PLS PM gives you many results to check (GoF, Cronbach’s alpha, DGrho, path coefficients , cross loading etc.,)
My quest is more to understand the direct and indirect effects of the independent variables, and the selection process of manifest variable (rather grouping certain manifest variables in to a latent variable).
In the diagram (attached) Price has 3 dimensions, Brand has 10 and Product has 6 dimensions. I have categorized as such based on the experience with category and brand, what I intend to know is
1) Is there any particular method to know which manifest variable (rightfully) would belong to a particular latent variable?
2) How to interpret the direct and the indirect effect? (usually I convert the coefficients into percentage for convenience, would this be the right way?)
3) Can we measure the effect of one manifest variable on two or three latent variable? (Example:- Can I measure the effect of masculine/dynamic effect on brand as well as product)


Thanks