forum.smartpls.com

The constructs of perceived parental support and self efficacy are so highly correlated (as evidenced by a very large standardized path coefficient), it is very likely that the models suffers from a multicollinearity problem. Maybe even some of the indicators of these constructs are very highly corr...

If your effect size is very small, your constructs has no substantial contribution to the construct. It does not explain much variance in the target construct. You its effect can be significant (i.e., different from zero). This is the typical difference between relevance and significance. There migh...

First and foremost AVE and reliability are similar but two different concepts. Second, some of the reliability criteria improve if you have many indicators (e.g., Cronbachs alpha). Thus, many medium indicator give good Cronbachs Alpha, but weak AVE. Third, the composite reliability criterion usually...

What exactly is the problem?

If you have no measure at all for your goal construct it will be hard to build a meaningful model, because you have no information about the part that interests you.

One way to decrease the variability and to increase t-values is to increase sample size. If you sample size is too small and your model too complex (you seem to have complex higher order constructs with repeated-indicators) you will get instable results. Another problem could be insufficient identif...

The reason is simple:
Some researchers believe they are useful and have to be reported. Others believe that they are not very useful and we do not know enough about their properties in PLS-SEM. Hence, they would not require you to report them.

Without knowing your model and data, it is hard to tell the problem

First, you should never ignore empirical evidence. You should report it and craft an argument for or against you hypotheses. Should lower-order constructs show discriminant validity? Yes. If they are not distinct, it would not make sense to model them as separate constructs. You should combine them ...

1.A If you carefully read the article you will understand why and when you have to calculate the numbers yourself. You always have to do it, when you use the repeated indicator approach, because the repeated indicators are not the actual measurement of a higher-order construct that you want to evalu...

As I have already answered to other posts in this forum, it is not possible to calculate all fit measures for all models. For some models it is simply not possible to calculate NFI and Chi-Square as well as d_G. In most cases, these are models with repeated indicators, for example, when you model hi...

For PLS models the assessment of model fit is generally not very established yet, and thresholds are not commonly defined. Some researchers also questions the usefulness of model fit assessment in PLS models as they are mainly predictive. You find some more information on our website: https://www.sm...

Usually, you also include the paths from Y1 to M2 and M1 to Y2 to show that you actually have a serial mediation and that the mediation is not only through M1 or M2. If you do not specify these additional paths you may find a serial mediation because you have specified an incomplete model. Thus, you...

This is a common problem. If you search the forum, you will find several discussions of this issue. Also we have compiles some explanations on our website: https://www.smartpls.com/documentation/ ... s-problems

Please read: Sarstedt, M., Hair Jr, J. F., Cheah, J. H., Becker, J. M., & Ringle, C. M. (2019). How to specify, estimate, and validate higher-order constructs in PLS-SEM. Australasian Marketing Journal (AMJ), 27(3), 197-211.