Hi!
I am new to PLS. I have a few questions and would really appreciate any help or comments:
1. I did scale reliability tests and Principle component tests in SPSS. Based on these results, I got an aggregated score for each construct in each case. Now I am using PLS to test my overall research model. Can I use the construct scores I got from SPSS tests for LV instead of using the multiple indicators?
2. I got much high R Square when I tested portions of my research model using multiple regression using SPSS than the R square I got from PLS analysis. Why is that?
Thank you!
Suling
Question about PLS vs. Multiple regression
-
- PLS Junior User
- Posts: 2
- Joined: Thu Jul 12, 2007 9:41 pm
- Real name and title:
-
- PLS Expert User
- Posts: 139
- Joined: Wed Jul 05, 2006 1:43 pm
- Real name and title:
- Location: Coimbatore, India
pls vs regression
First of all EFA is only a trial check of constructs. You need to check for validity of constructs to confirm them. When you can use separate indicators, why do you want to use regression.
You can also take the scores of constructs from pls and try out regression, you will get the same results as in PLS.
Furthur in regression you have only one dependant variable and a few independant variables, not a path model. Path model coefficients are always different from normal regression coefficients, and are a more reliable indicator of relations in an interdependant model.
You can also take the scores of constructs from pls and try out regression, you will get the same results as in PLS.
Furthur in regression you have only one dependant variable and a few independant variables, not a path model. Path model coefficients are always different from normal regression coefficients, and are a more reliable indicator of relations in an interdependant model.