CTA-PLS using the Two-Stage Approach

PLS is broadly applied in modern business research. This forum is the right place for discussions on the use of PLS in the fields of Marketing, Strategic Management, Information Technology etc.
Post Reply
Jennifer H.
PLS Junior User
Posts: 4
Joined: Thu Jun 07, 2018 2:52 pm
Real name and title: Jennifer Hendricks, Phd candidate

CTA-PLS using the Two-Stage Approach

Post by Jennifer H. » Wed Nov 20, 2019 10:16 am

Hello,

I would like to use CTA-PLS for specifying a higher-order construct. In "Advanced issues in PLS" (Hair et al. 2018, p. 54) there is a recommendation for employing the two-stage approach in this case. As the tow-stage approach uses the standardized latent variable scores, I am not sure if the interpretation of the CTA-PLS results is the same with this approach? Attached you find my results and my conclusion.

"As more than one of the construct’s tetrads is significantly different from 0, the null hypothesis is rejected and a formative measurement specification between the first- and the second-order constructs is confirmed (Hair et al. 2019a)."

Thank you very much in advance!

Best,
Jennifer
Attachments
Unbenannt.jpg
Unbenannt.jpg (184.28 KiB) Viewed 116 times

User avatar
cringle
SmartPLS Developer
Posts: 812
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Re: CTA-PLS using the Two-Stage Approach

Post by cringle » Sat Nov 23, 2019 11:42 am

Sounds reasonable. Just run the two-stage approach as explained here:
https://www.sciencedirect.com/science/a ... 8219301223

Then, just run the CTA-PLS on the second stage.

Jennifer H.
PLS Junior User
Posts: 4
Joined: Thu Jun 07, 2018 2:52 pm
Real name and title: Jennifer Hendricks, Phd candidate

Re: CTA-PLS using the Two-Stage Approach

Post by Jennifer H. » Sat Nov 23, 2019 12:58 pm

Dear Professor Ringle,

Thank you very much for your help and the literature reference! I know this paper and found it very useful for specifying higher-order constructs! I actually employed the embedded two-stage approach and ran the CTA as you recommend, I was just wondering if my interpretation of the results is correct because based on the latent variable scores, the results look quite different from those in a regular CTA analysis...

Again, thank you very much for your great support!

User avatar
cringle
SmartPLS Developer
Posts: 812
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Re: CTA-PLS using the Two-Stage Approach

Post by cringle » Sat Nov 23, 2019 1:26 pm

The numbers in the resutls look very larger indeed. You probably did not correctly copy, paste, save and import the latent variable scores.

jmbecker
SmartPLS Developer
Posts: 1059
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: CTA-PLS using the Two-Stage Approach

Post by jmbecker » Tue Nov 26, 2019 2:51 pm

Actually, you should better use the unstandardized latent variables scores from the IPMA as indicators on the second stage to conduct the CTA. Usually, it does not make much of a difference, whether you use unstandardized or standardized latent variable scores as indicators, but the CTA works on covariances so you will only get reasonable results from unstandardized data.
Just run the IPMA on your first stage model.
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

Jennifer H.
PLS Junior User
Posts: 4
Joined: Thu Jun 07, 2018 2:52 pm
Real name and title: Jennifer Hendricks, Phd candidate

Re: CTA-PLS using the Two-Stage Approach

Post by Jennifer H. » Tue Nov 26, 2019 7:09 pm

Dear Professor Becker, dear Professor Ringle,

Thank you very much for your valuable feedback! I conducted the first analysis again and finally identified the errors! I also ran the IPMA and used the unstandardized latent variable scores as indicators, and the CTA results regarding the higher-order construct specification remain the same. I am very grateful for your advice and for being able to do it correctly.

Post Reply