HTMT >1

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
elambour
PLS Junior User
Posts: 2
Joined: Sun Oct 16, 2022 9:24 pm
Real name and title: Eric Lambourdiere Dr.

HTMT >1

Post by elambour »

Hi everyone,
I have exactly 38 hours with Partial Least Squares Structural Equation Modeling.
I did performed a bootstrapping HTMT and I have relation that equals 1.003. I don't know what to do now.
Best regards,
Eric
PLSSEM.png
PLSSEM.png (121.84 KiB) Viewed 751 times
jmbecker
SmartPLS Developer
Posts: 1281
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: HTMT >1

Post by jmbecker »

The HTMT is the ratio of the between construct correlations to the within construct correlations of items. In a proper (reflective) measurement model, the between construct correlations are smaller than the within construct correlations (i.e., items with a construct should correlate higher than items of different constructs).
If HTMT is larger than 1, it indicates that your between construct correlations are larger than within construct correlations. Thus, it points to problem with discriminant validity and measurement operationalization in general. It could be caused by wrongly assigned indicators, constructs that are not reflective but formative (where within correlation are often low), improper construct definition (maybe both construct are actually part of a larger concept), and many more (or a combination thereof).

Hence, you really want to go back and thoroughly review the measures of both constructs and their definition to find the underlying problem.
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
elambour
PLS Junior User
Posts: 2
Joined: Sun Oct 16, 2022 9:24 pm
Real name and title: Eric Lambourdiere Dr.

Re: HTMT >1

Post by elambour »

Hi Dr Becker,
Thank you so much for this clear and specific answer.
I'll follow your suggestions and get back to the data and the model as well.
I wish you an awesome day.
Best regards,
EL
Post Reply