Consistent Bootstrapping

Frequently asked questions about PLS path modeling.
Post Reply
lucy
PLS Junior User
Posts: 2
Joined: Tue Dec 12, 2017 6:02 pm
Real name and title: Luciana Alves Rodas Vera

Consistent Bootstrapping

Post by lucy » Tue Dec 12, 2017 6:45 pm

Dear SmartPLS Forum,

I have just done the following in Smartpls 3 for one specif sample of my work:

1) Ran the consistent PLS algorithm. The results are ok.
2) Ran the consistent Bootstrapping (5000 subsamples). The results of all my T-values have returned as 0.000.

I don't understand, because I tested the same model for two other samples, I did the same process and the results are ok (the T-values are different to 0). But for this specif sample, the results of T-value returned as 0,000...
After these results, I tried to use the normal Bootstrapping for this specif sample, after using the consistent PLS algorithm. The results of T-Value didn't return as 0,000...But I don't know if it's correct this procedure...
What do you recommend?

Thank you so much

jmbecker
SmartPLS Developer
Posts: 1065
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Consistent Bootstrapping

Post by jmbecker » Tue Dec 12, 2017 9:11 pm

You may search the forum for this topic. It has been intensively discussed before.

In brief: consistent PLS is very strict about its assumptions of a factor model for reflective construct. If this is not fulfilled it may produce inadmissible solutions. In this regard, it is less robust that the normal PLS. This manifest, for example, in instable bootstrapping performance, producing large outliers for the estimated parameters that heavily effect the estimated bootstrap standard error. This is similar to Heywood cases in CB-SEM.
Dr. Jan-Michael Becker, University of Cologne, SmartPLS Developer
Researchgate: https://www.researchgate.net/profile/Jan_Michael_Becker
GoogleScholar: http://scholar.google.de/citations?user ... AAAJ&hl=de

Post Reply