two-stage second order approach question, urgent!

Frequently asked questions about PLS path modeling.
Post Reply
tonylu1987
PLS Junior User
Posts: 1
Joined: Tue Nov 29, 2016 5:49 am
Real name and title: Mr. Tony Lu

two-stage second order approach question, urgent!

Post by tonylu1987 » Tue Nov 29, 2016 6:36 am

Hi all,

I am a new researcher and recently I am working on a model that has 3 second-order constructs (reflective-formative, type ii), within which 2 are endogenous variables and 1 is exogenous variable. I followed a mix of repeated approach and two-stage approach.
I draw a model with all variables and constructs including both first order and second order constructs and generate the latent variable scores for all the first order constructs. Then I use the first order latent variable scores as the indicators and link them to the second order constructs in a formative way and run the model again.
My problem is, in the first stage where I adopted a repeated approach, weights of first order constructs on the second order constructs are all positive and significant, however, when i move to the second stage, in which I use the latent variable scores of the first order constructs as the indicators to the second order constructs, some of the indicators (i.e. first order constructs presented by the latent variable scores)' weights turned out to be insignificant and negative, although the path coefficients are all as expected.

My question is, is it a big concern if some of the indicators' weights turned out to be insignificant? If it is a big concern, what else can I do? any other recommended approaches other than the one I adopted to test my model, given the complexity of my model?

Thank you very much in advance and I really need this to be solved very urgently!

Tony

User avatar
cringle
SmartPLS Developer
Posts: 797
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Re: two-stage second order approach question, urgent!

Post by cringle » Wed Mar 01, 2017 8:36 pm

You may have a collinearity issue on the second stage. Do you really need to run a higher-order construct approach or can't you just relate the lower order constructs with each other (that's what I usually do since PLS has no complexity limitations).

Best regards
Christian

Jennifer H.
PLS Junior User
Posts: 1
Joined: Thu Jun 07, 2018 2:52 pm
Real name and title: Jennifer Hendricks, Phd candidate

Re: two-stage second order approach question, urgent!

Post by Jennifer H. » Thu Jun 07, 2018 3:17 pm

Dear Christian Ringle,

I have a similar problem as the above described, as three of my seven first-order construct scores have insignificant weights. Is it really necessary to consider this when I have done the repeated indicator approach in a first stage and checked for all quality criteria at this stage?

I am looking forward to your answer as I have to find a solution for this problem very soon... Thank you very much in advance!

Best,
Jennifer

muriloz
PLS Junior User
Posts: 1
Joined: Sat Jun 23, 2018 12:48 pm
Real name and title: Murilo Zamboni Alvarenga

Re: two-stage second order approach question, urgent!

Post by muriloz » Sat Jun 23, 2018 12:55 pm

Hello,

Same problem here. I have to report the measurement model that i had in the first stage or in the second stage?


I am asking because in the first stage are OK, but in the second not (no VIF problem).

Post Reply