P values and CI Bias Corrected not tally

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
Kim78
PLS Junior User
Posts: 9
Joined: Wed Sep 05, 2018 2:46 am
Real name and title: Tan Kim Lim, PhD Candidate

P values and CI Bias Corrected not tally

Post by Kim78 »

Hi pals,

Not sure if this is a new topic. But when I recently did a bootstrapping, I realised that the P-values shows significant, but CI bias-corrected crosses "0". Like to seek advice on what could be the problem and how should I interpret this result ? Appreciate guidance on this.
jmbecker
SmartPLS Developer
Posts: 1282
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: P values and CI Bias Corrected not tally

Post by jmbecker »

The p-value is based on the t-value and thus makes distribution assumptions about the variability of the coefficient (t-distribution). You will get corresponding confidence intervals if you choose Studentized Bootstrap. Studentized confidence intervals are symmetric and will give you the same result as the p-value based assessment.
However, your coefficients sampling distribution may not be symmetric. It might be slightly skewed or peaked, etc. In those cases the non-parametric bias-corrected percentile or bias-corrected and accelerated (BCa) intervals might be better.
Generally, it is hard to know which interval method is best, but the studentized confidence interval is the least used in PLS to avoid making distributional assumptions.
Dr. Jan-Michael Becker, BI Norwegian Business School, SmartPLS Developer
Researchgate: https://www.researchgate.net/profile/Jan_Michael_Becker
GoogleScholar: http://scholar.google.de/citations?user ... AAAJ&hl=de
Kim78
PLS Junior User
Posts: 9
Joined: Wed Sep 05, 2018 2:46 am
Real name and title: Tan Kim Lim, PhD Candidate

Re: P values and CI Bias Corrected not tally

Post by Kim78 »

Thank you Dr Becker for the guidance.
Post Reply