I run a regression Satisfaction=f(value) with SPSS and obtained a Rsquared=0,58.
I run the same regression with PLS and obtained Rsquared=0,999.
The sample is made of 63 cases.
Do you think is it reasonable such difference in Rsuqared? Did I do something wrong?
thanx
Fabio
PLS vs. SPSS - regression analysis
-
fabio.colacchio
- PLS Junior User
- Posts: 7
- Joined: Thu Mar 22, 2007 4:47 pm
- Real name and title:
-
cringle
- SmartPLS Developer
- Posts: 818
- Joined: Tue Sep 20, 2005 9:13 am
- Real name and title: Prof. Dr. Christian M. Ringle
- Location: Hamburg (Germany)
- Contact:
Hi,
you probably did include the interception term. This is wrong - the last step of PLS uses (in the formative mode) a multiple Regression of the manifest variables on the latent variable scores and thr R² is close to one. This is always the results if you replicate this step in EXCEL, SPSS, R etc.
Cheers
Christian
you probably did include the interception term. This is wrong - the last step of PLS uses (in the formative mode) a multiple Regression of the manifest variables on the latent variable scores and thr R² is close to one. This is always the results if you replicate this step in EXCEL, SPSS, R etc.
Cheers
Christian
Prof. Dr. Christian M. Ringle, Hamburg University of Technology (TUHH), SmartPLS
- Literature on PLS-SEM: https://www.smartpls.com/documentation
- Google Scholar: https://scholar.google.de/citations?use ... AAAJ&hl=de
- Literature on PLS-SEM: https://www.smartpls.com/documentation
- Google Scholar: https://scholar.google.de/citations?use ... AAAJ&hl=de
-
fabio.colacchio
- PLS Junior User
- Posts: 7
- Joined: Thu Mar 22, 2007 4:47 pm
- Real name and title:
Hi Christian,
I m not sure this is my case.
Both Value and Satisfaction are latent variables. Each of them has 1 manifest indicator (obtained by summated scales). Furthermore when I run the same regression in PLS with the full sample (N=144 cases) Rsquares is equal to 49,8. When I select the sub-sample it jumps to 0,99. Can it be related to sampling bias?
I m not sure this is my case.
Both Value and Satisfaction are latent variables. Each of them has 1 manifest indicator (obtained by summated scales). Furthermore when I run the same regression in PLS with the full sample (N=144 cases) Rsquares is equal to 49,8. When I select the sub-sample it jumps to 0,99. Can it be related to sampling bias?