T-statistic difference in Path Coefficient and Outer Loading

Frequently asked questions about PLS path modeling.
Post Reply
rizka.khairunnisa
PLS Junior User
Posts: 1
Joined: Tue Feb 13, 2018 1:32 am
Real name and title: Rizka Khairunnisa, S.ST

T-statistic difference in Path Coefficient and Outer Loading

Post by rizka.khairunnisa » Mon Feb 19, 2018 2:54 am

If we do bootstrapping, we can see T-statistic value in Path Coefficient and in Outer Loading tab. What are the differences between them?
Based on literatures I read, T-statistic value in Path Coefficient describes whether the the hypothesis is rejected or accepted. Then, what are described by the T-statistic value in Outer Loading?

Thank in advance.

User avatar
cringle
SmartPLS Developer
Posts: 804
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Re: T-statistic difference in Path Coefficient and Outer Loading

Post by cringle » Wed Feb 21, 2018 3:24 pm

Exactly what you say: the t value of the outer loading (which usually are very high). But you should rather use the 95% bias-corrected and accelerated (BCa) bootstrap confidence intervals for significance resting (and report those results).

Regards
Christian

elenag2
PLS Junior User
Posts: 1
Joined: Thu Feb 07, 2019 4:32 pm
Real name and title: Elena Gorokhova

Re: T-statistic difference in Path Coefficient and Outer Loading

Post by elenag2 » Thu Feb 07, 2019 6:49 pm

I would like to elaborate on this question. When optimizing the model - what is more important: to get more significant path coefficient for IV-LV path or significant indicators for LV? In my model, both are on the border of significance threshold and by increasing the one I decrease the other...
/Elena

jmbecker
SmartPLS Developer
Posts: 932
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: T-statistic difference in Path Coefficient and Outer Loading

Post by jmbecker » Mon Feb 11, 2019 2:03 pm

First, it depends on what type of measures you have. For reflective measures you always want high and significant loadings of all your indicators to ensure the reliability of your measurement of the constructs.

Second, you want to test the significance of your paths, i.e., find out if one construct has an influence on another as hypothesized by your theory.

Finally, you should not fudge with the significance. Don't optimize your model to get more or better results. Model according to your theoretical considerations and interpret the results.
Dr. Jan-Michael Becker, University of Cologne, SmartPLS Developer
Researchgate: https://www.researchgate.net/profile/Ja ... v=hdr_xprf
GoogleScholar: http://scholar.google.de/citations?user ... AAAJ&hl=de

Post Reply