A deflated indicators approach for estimating second-order reflective models through PLS-PM: an empirical illustration |
| |
Authors: | M Nitti E Ciavolino |
| |
Institution: | Department of History, Society and Human Studies, University of Salento, Italy |
| |
Abstract: | The paper provides a procedure aimed at obtaining more interpretable second-order models estimated with the partial least squares-path modeling. Advantages in interpretation stem from the separation of the two sources of influence on the data. As a matter of fact, in hierarchical models effects on manifest variables (MVs) are assigned to both first-order (specific) factors and second-order (general) factors. In order to separate these overlapping contributions, MVs are deflated from the effect of the specific latent variables (LVs) and used as indicators of the second-order LV. A case study is presented in order to illustrate the application of the proposed method. |
| |
Keywords: | high-order construct model PLS-PM psychological assessment deflated indicators OQ-45 |
|
|