forum.smartpls.com

You are talking about higher order latent variable:
- If the first order LVs have about the same number of indicators, use the repeated indicators approach (see Wetzels et al., 2009, p.181)
- If the number of indicators are very different one of another, it is better to use de two step approach (see Wilson, & Henseler, 2007).

WETZELS, M. et al. Using PLS path modeling for assessing hierarchical construct models: guidelines and empirical illustration. MIS Quarterly, v.33, n.1, p.177-195, March 2009. [p.181]

Wilson, B., & Henseler, J. (2007). Modeling reflective higher-order constructs using three approaches with PLS path modeling: a Monte Carlo comparison. In Australian and New Zealand Marketing Academy (ANZMAC) Conference (pp. 791–800). Retrieved from http://doc.utwente.nl/91758/1/BWilson_2.pdf

Best regards,

Bido

All constructs need indicators in PLS. Therefore you need to use approaches that try to identify the second-order constructs, which are usually either the repeated indicator approach or the two-stage approach.

You can find information on this issue in the following publications:
Becker, J.-M., K. Klein, and M. Wetzels (2012). Formative Hierarchical Latent Variable Models in PLS-SEM: Recommendations and Guidelines, Long Range Planning, 45 (5/6), 359-394.
Kuppelwieser, V., and M. Sarstedt (2014). Applying the Future Time Perspective Scale to Advertising Research, International Journal of Advertising, 33(1), 113-136.
Ringle, C. M., M. Sarstedt, and D. W. Straub (2012). A Critical Look at the Use of PLS-SEM in MIS Quarterly, MIS Quarterly, 36 (1), iii-xiv.
Hair, J. F., Sarstedt, M., Ringle, C. M., & Gudergan, S. P. 2018. Advanced Issues in Partial Least Squares Structural Equation Modeling (PLS-SEM). Thousand Oaks, CA: Sage.

How we can interpret the meaning of latent variable scores? What does these scores represent?

PLS is a composite-based method (in contrast to CB-SEM, which is a factor-based method). That means it treats latent variables as composites (i.e., a weighted linear combination of its indicators/manifest variables, as explain earlier in this threat). These latent variable scores or composite scores are the central quantity that define the behavior in composite-based methods such as PLS. They represent the latent variable an approximate their behavior. When using the unstandardized latent variables scores from the IPMA you can take those values as direct estimates of the real value for the latent variable. For example, if you have a 2 on a scale from 1 to 7 on a latent satisfaction variable that would imply a dissatisfied customer.

Dear JMBecker,

Can you please guide how many decimal places to take (or round) for Latent Variable Scores to be used as indicators in the model?
(since raw data is integer)

The more the better. SmartPLS calculates with double precision that is 16 decimal places.