Bonferroni

Frequently asked questions about PLS path modeling.
Post Reply
MariaLunde
PLS Junior User
Posts: 6
Joined: Wed Dec 19, 2018 2:23 pm
Real name and title: Maria Lunde

Bonferroni

Post by MariaLunde » Fri Dec 21, 2018 8:34 am

Hello,

I am about to do a multigroup analysis as the MICOM analysis confirmed that the data is appropriate for the analysis.
However, I read about the Bonferroni correction in Advanced issues in partial least squares structural equation modeling.

How do I deal with the Bonferroni correction when doing a multigroup analysis between 4 segments?

I am very much looking forward to your help.

Kind regards,
Maria

jmbecker
SmartPLS Developer
Posts: 908
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Bonferroni

Post by jmbecker » Fri Dec 21, 2018 11:04 am

When you do pairwise group comparisons with 4 groups you have in total 6 comparisons (i.e., 1-2, 1-3, 1-4, 2-3, 2-4, 3-4).

According to Bonferroni you need to achieve a p-value lower than alpha/n, with n being the number of comparisons. In your case (with 6 comparisons), you would need a p-value lower than 0.05/6=0.00833 for an alpha of 0.05.

Bonferroni is usually the most conservative correction method. It thereby avoids type I errors (false positives), but might inflate type II errors (false negatives). But it is also the easiest method to apply.
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

MariaLunde
PLS Junior User
Posts: 6
Joined: Wed Dec 19, 2018 2:23 pm
Real name and title: Maria Lunde

Re: Bonferroni

Post by MariaLunde » Fri Dec 21, 2018 11:34 am

Thanks for your reply, I really appriciate your comments. However, I might be a bit slow (I am very new to PLS), because I thougt that I should do a multigroup analysis as follows (if four categories):

Compare group 1 with group 2 (p value should be <0,05 or >0,95 if there are significant differences)
Compare group 1 with group 3 (p value should be <0,05 or >0,95 if there are significant differences)
Compare group 1 with group 4 (p value should be <0,05 or >0,95 if there are significant differences)
Compare group 2 with group 3 (p value should be <0,05 or >0,95 if there are significant differences)
Compare group 2 with group 4 (p value should be <0,05 or >0,95 if there are significant differences)
Compare group 3 with group 4 (p value should be <0,05 or >0,95 if there are significant differences)

Have I misunderstood?

Maria

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: Bonferroni

Post by cringle » Sun Dec 23, 2018 1:59 pm

This is not a PLS-specific but general concern. Please search for articles, books, and webpages that address the multiple testing issues. In multiple testing situations, we need to adjust the accepted probability of error level. When using the Bonferroni correction to account for 6 comparisons, the probability of error of 5% changes into 0.83%. In accordance, you need to adjust the considered probability levels (e.g., <0.0083 and >0.9917).

Best
CR

MariaLunde
PLS Junior User
Posts: 6
Joined: Wed Dec 19, 2018 2:23 pm
Real name and title: Maria Lunde

Re: Bonferroni

Post by MariaLunde » Thu Dec 27, 2018 8:39 am

Thanks a lot, this clarified my understanding:)

Post Reply