Hi there,
i was reading many journals about the multigroup analysis of the path coefficents in a multigroup analysis but nothing about any R-Sqaured Comparison? Is there any reason? Why cant i compare the R-Squared of my dependent constructs not with the "same" R-Squared of the same model only with different data?
SMartPLS3 supplies a MGA considering the path coefficents.
i have 2 equal models (one group with female data, the second with male data) and it would be lovely if its possible to compare more than the path regression weigjhts.
Is here anyone who can tell me if there any journals about that R-Squared comparison or why this comparison isnt working or valid.
Cheers, thank you and i hope to learn something from you guys
Dear,
I do not know if I was clear his doubt but usually the goodness of different methods on a set of data is compared.
Sometimes we have multiple datasets of the same thing, and in that case if the comparison is valid. This gives us information on the variability of the phenomenon ofr the proposed methods.
Dear Levent,
I guess that is because the aim of doing multi-group analysis is to see if there are any significant differences in the parameters for the different (pre-assigned) groups. So you could say that there are significant differences between the female and male groups. That being said, at least Smartpls 3+ also reports on R squared for each group and tells if differences in R squared between the groups are significant. So you can also reflect on the explanatory power of each group of data, although a slightly different consideration.
Best regards,
Thorhildur Jetzek
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling