forum.smartpls.com

Dear all：
I would be very grateful if you tell me the difference between the path coefficients of smartpls and the standardized regression coefficients of spss. Are they the same？

Hi,
yes, they are the same.
If you want to test:
- run the one model in SmartPLS, for example: (x1, x2, x3 --> y)
- report / PLS / calculation results / Latent variable scores
- copy these scores to a spreadsheet / import them to SPSS
- Run the regression analysis (structural model).
- compare the beta values with path coefficients.

Best regards,
Bido

Diogenes wrote:Hi,
yes, they are the same.
If you want to test:
- run the one model in SmartPLS, for example: (x1, x2, x3 --> y)
- report / PLS / calculation results / Latent variable scores
- copy these scores to a spreadsheet / import them to SPSS
- Run the regression analysis (structural model).
- compare the beta values with path coefficients.

Best regards,
Bido

Dear professor，
Why are they not equal when I test? Following is my test steps, there are five measured variables for each latent variable in my model. I take the mean of the five measured variables. Then I run the linear regression of the means and get the standardized coefficients. The standardized coefficients are not equal with the path coefficients. Is there any wrong with my steps?

Hi,

The LV scores from SmartPLS are not simple mean of the indicators, they are weighted by the outer weights).

Suggestion to compare regression in a simple way:
- use just one indicator by LV
- run the model in SmartPLS, for example: (x1, x2, x3 --> y)
- Run regression analysis (x1, x2, x3 --> y) in SPSS.
- compare the beta values with path coefficients.

Best regards,
Bido

using latent variable scores/factor scores is common in research and especially in factor analysis and regression.

c.f., My PLS Handbook chapter which outlines the process of doing this.

Good luck
Brad