Discriminant validity of 2nd Order Factor

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
mdw
PLS Junior User
Posts: 3
Joined: Sat Dec 06, 2008 8:45 pm
Real name and title:

Discriminant validity of 2nd Order Factor

Post by mdw »

I realize this type of question has probably been asked before, but having read all the replies and associated links, I'm still a little unclear on how to tackle this issue.

I have a reflective-formative 2nd order factor within a model (as per p.231 of the PLS Primer by Hair et al) whereby the LOCs are measured by reflective indicators, and the LOCs have a formative relationship with the HOC.

All values within the entire model are absolutely fine, apart from the AVE of the 2nd order construct, which remains stubbornly just below 0.4.

Am I correct in thinking this isn't a problem? Some replies suggest AVE isn't required for a HOC, but others suggest I need to get it near to or > 0.5.

I've read the LRP paper by Becker et al (2012), but am still trying to arrive at the definitive view.

Any help much appreciated!
User avatar
Diogenes
PLS Super-Expert
Posts: 899
Joined: Sat Oct 15, 2005 5:13 pm
Real name and title:
Location: São Paulo - BRAZIL
Contact:

Post by Diogenes »

Hi,

If HOCs are formative, it does not make sense compute AVE and composite reliability.
We do not expect that their LOCs will be correlated, for this reason they are used as formative.

If HOCs were reflective, we should compute their AVE and CC based on their relations with LOCs (they are presented as path coefficients, but we should interpret them as loadings, see: viewtopic.php?t=640 ).


Best regards,

Bido
bwilson
PLS Expert User
Posts: 41
Joined: Wed Apr 02, 2008 7:11 am
Real name and title:
Contact:

Post by bwilson »

Diogenes wrote:Hi,

If HOCs are formative, it does not make sense compute AVE and composite reliability.
We do not expect that their LOCs will be correlated, for this reason they are used as formative.

If HOCs were reflective, we should compute their AVE and CC based on their relations with LOCs (they are presented as path coefficients, but we should interpret them as loadings, see: viewtopic.php?t=640 ).


Best regards,

Bido
Yes. the validity should be tested within a broader nomological framework.
never report AVE or Reliabilities for formative constructs.

regards
Brad
Bradley Wilson. Ph.D.
Senior Lecturer in Advertising.
RMIT University.
School of Media and Communication.
GPO Box 2476V
Location. 9.5.20
Melbourne. Victoria.
Australia.

SEE FOR PUBLICATIONS
www.rmit.edu.au/staff/bradleywilson
mdw
PLS Junior User
Posts: 3
Joined: Sat Dec 06, 2008 8:45 pm
Real name and title:

Many thanks...

Post by mdw »

Many thanks both - problem now resolved!
mattburek
PLS Junior User
Posts: 1
Joined: Fri Feb 09, 2018 10:38 am
Real name and title: Mateusz Burek, B.Sc.

Re: Discriminant validity of 2nd Order Factor

Post by mattburek »

I encountered the same problem. I have a reflective-formative model and would like to report my numbers. Following Hair et al. (2014), I need to follow " the same measurement model evaluation criteria apply to the HOC as for any other construct in the PLS path model".

My reflective latent variables (LOCs) are all discriminant and convergent valid. Now what do I have to report for the HOC?

Regards
jmbecker
SmartPLS Developer
Posts: 1282
Joined: Tue Mar 28, 2006 11:09 am
Real name and title: Dr. Jan-Michael Becker

Re: Discriminant validity of 2nd Order Factor

Post by jmbecker »

The same as you would report for any formative latent variable.
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
Post Reply