forum.smartpls.com

1) For the Cross-Loadings in SmartPLS 3, click on Discriminant Validity and you will find them. 2) How did you set the values for the Bootstrapp in SmartPLS 2? Did you set the sample size correctly? Some users had used wrong sample sizes in SmartPLS 2 in the past or were confused which number is cor...

First, you need to have at least two variables predicting your dependent variable to include an interaction term. The interaction term is, like in regression, based on a product term (with different approaches to generate the product term: Two Stage, Product Indicators, Othrogonalizing). If you only...

You will find such a result in the IPMA under Total Effects.

Actually, I don't see any problem with a formative construct as dependent variable. In fact, claiming to avaoid formative constructs in endogenous positions would be like claiming to avoid endogenous variables in PLS altogether.

1) Do you have the latest version (3.1.6)? There was a bug in version 3.1.5 regarding indirect and total effects. 2) Which results are different? In most of the algorithms there should be no difference between the two versions. 3) I know of people who have used wrong data settings and got different ...

Are you using the latest version (3.1.6)??? There was a known bug with indirect and total effects in an earlier version.

We are working on this topic and we have a preliminary version running for the PLS algorithm. However, it gets complicated in combination with some of the other algorithms and we do not want to release it in a premature stage,

The problem is now resolved with the new release 3.1.6

With the latest release (3.1.6), we also added more validations for the data files.
For files you already imported, there is now a "Re-Analyze" button.
If there are validation errors, these will be listed.

If you use factor-weighting scheme and Mode A you should generally not need regressions (RGP function) but only correlations/covariances. An important aspect to keep in mind is that both the indicators and latent variable scores are always standardized. To achieve this the weights under your step 1....

Hi, On the one hand, the latent variables scores (or composite scores) often follow a normal distribution if you have multiple indicators. This is because of the central limit theorem that states that a combination of a large number of random variables, each with a well-defined expected value and we...

Another useful article that describes the method for formative second-order constructs is Becker, J. M., Klein, K., & Wetzels, M. (2012). Hierarchical latent variable models in PLS-SEM: guidelines for using reflective-formative type models. Long Range Planning, 45(5), 359-394. From your picture ...

Hi, it is not possible to have constructs without indicators in PLS. You need indicators to estimate the construct. If you want to model something like a second-order construct you could use the repeated indicator approach. However, your final dependent variable does not even look like a typical sec...

Beside saving the data in the correct format, it is important that every row and every column have the same number of values stored. It is not allowed to have spaces, i.e., you must have a numerical value in every cell, except for the first row, where you have your indicators.

To make it clear. What was reported in Smart 2 under communality is the same as the AVE. Hence, we only report AVE in Smart 3.