Formative (Typ II) 2nd Order Constructs (Endogenous) with CS

[Julianaskrent]

Dear all,

I want to model a second order construct in the following way: relationship between indicators and first order is reflective, relationship between first order and second order is formative. (Typ II).

It is an endogenous construct. I have paid attention that the number of indicators for all three dimensions are the same. Therefore, I believed that I should be fine applying the "Hierarchical Component"- Method (repeated indicators).

I attended a seminar with Jörg Henseler and Christian Ringle in June of this year. In their handout-material on slide 98 such a 2nd Order Construct is modeled with the Hierachical Component Method (repeated indicators). The graphic refers to the article by Jarvis/MacKenzie/Podsakoff (2003).

Now, I have read in this forum here the following two discussions:

viewtopic.php?t=660&highlight=second+order

viewtopic.php?t=719&highlight=second+order

in which Professor Bido points out, that an endogenous Typ II second order construct should be modeled with the composite score.

I have tried to read Professor Bido's paper on this topic (http://mee-pls.googlegroups.com/web/Var ... %20PLS.doc ), but as I don't speak Portuguese, I have problemens understanding it.

Can somebody explain me, how to proceed? How do I get the scores for the first order? Do I calculate a mean for every 1st-order LV? I tried it, but got an error warning while running the algorithm...

I am looking forward to your help
Regards
Julia