t values in model results

Frequently Asked Questions about the SmartPLS software. Please refer to this section first, if you have problems with the software.
Post Reply
tht
PLS Junior User
Posts: 4
Joined: Mon Nov 07, 2005 5:59 pm
Real name and title:

t values in model results

Post by tht »

Hi,

I wonder how the t-values for the path coefficients given in the regular model outcomes section (not the bootstrap) is calculated? I thought that PLS does not calculate such parameters due to its distribution-free character? I also found that these values differ from the n=200 bootstrap estimates.

Thanks Thorsten
thorsten.giesa
PLS Junior User
Posts: 4
Joined: Wed Nov 16, 2005 1:41 pm
Real name and title:

Post by thorsten.giesa »

I would be interested in an answer to that question as well, especially as I am also looking for t-values for the FIMIX-PLS Path coefficients!

Thanks,

The other Thorsten ;-)
User avatar
cringle
SmartPLS Developer
Posts: 821
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Post by cringle »

Hi,

these are the t-values of the “regular” (multiple) regression in the structural model.

We just developed those values for internal test – if they go along with the distribution-free character of SmartPLS should be discussed in the forum “FAQ (Methodology)“.

As stated in other post, we did not include the cross-loadings in this version by mistake. Instead we added these t-values… - this will be corrected in the next SmartPLS release.

About the FIMIX-PLS t-statistics – those values are not calculated. As stated in our PLS’05 paper on FIMIX-PLS (Ringle/Wende/Will 2005, Customer segmentation with FIMIX-PLS, in: Aluja, T./Vinzi, V. E./Casanovas, J./Morineau, A./Tenenhaus, M. (Hrsg.): PLS and related methods : proceedings of the PLS'05 international symposium, Barcelona 2005, pp. 507-514), these t-values are calculated by aggregate as well as a priori segmented PLS-model-estimation (combined with bootstrapping).

Suggestions (with detailed information on the correct calculation) for generating “FIMIX-PLS t-statistics” are welcome and can be implemented in one of the upcoming releases (please post this in “Requested Features“).

Best regards
Christian
tht
PLS Junior User
Posts: 4
Joined: Mon Nov 07, 2005 5:59 pm
Real name and title:

Thanks!

Post by tht »

Thanks much!

Thorsten
Post Reply