Problem:invalid path model , the focal construct

[lorie_bouchard]

Dear people,

I need an advice I want to test my research measure with PLS.

the measurement is made of 4 reflective first order measure and the weighted composite of these form the second-order latent variable. However, I did not use a reflective item to measure the focal construct or any item for that matter. In consequence, I get a message from PLS as an invalid model.

Is there a way to remedy to this ? What can I do?

I did some reading, I am knew to PLS. As a solution, I thought of using the hierarchical model proposed by Wold (1982) explained by Tenenhaus et al. (2005). It is as follow:

"Forestimating these LVs § , Wold (1982) has proposed the hierarchical model
defined as follows:
• A new block X is constructed by merging the J blocks X1; : : : ;XJ into a superblock.
• The superblock X is summarized by one LV §.
• A path model connects each exogenous LV §j to the endogenous LV §." (Tenenhaus et al., 2005, p.48)

In your opinion, is it a viable solution?
Thank you very much

Lorie