Model SmartPLS with moderator variables.

Questions about the implementation and application of the PLS-SEM method, that are not related to the usage of the SmartPLS software.
Post Reply
Francisco
PLS Junior User
Posts: 9
Joined: Fri Sep 15, 2017 8:03 am
Real name and title: Francisco Manuel

Model SmartPLS with moderator variables.

Post by Francisco »

Hi, I'm doing a research project. I am new to structural equation models and especially to SMARTPLS software. I need urgent help and would greatly appreciate the guidance of an expert. I take this opportunity to thank you for the work you do and that I am discovering little by little.
I enter to evaluate my model: The objective of my research is to analyze the relationship between the entrepreneurial environment (a variable composed of 12 indicators) and the entrepreneurship (2 variables composed of 1 indicator each). In addition, all this is moderated by a categorical variable referring to a group of countries of 9, 27 and 24 countries respectively.
What would be the next steps?
I have done an analysis of permutations and I can observe that the assumption of compound invariance is not fulfilled.
How can I solve that?.
I have been able to verify that there are significant differences for the TAP variable and there are no such differences for the APR, although observing the Path coefficients if we can see that there are differences, why does this happen?
Also, I need to know:
How can I study the influence of the two entrepreneurial variables in each of the conditions (indicators) of the environment ?.
Should I perform the PLS algorithm and bootstrapping to analyze the measurement model and the structural model for the general model ?.
Where can I look at the influence it has had in including the moderator variable over the original model?

Anything that can help me, I would appreciate it enough.
User avatar
cringle
SmartPLS Developer
Posts: 818
Joined: Tue Sep 20, 2005 9:13 am
Real name and title: Prof. Dr. Christian M. Ringle
Location: Hamburg (Germany)
Contact:

Re: Model SmartPLS with moderator variables.

Post by cringle »

Let's try to look into this.
Hi, I'm doing a research project. I am new to structural equation models and especially to SMARTPLS software. I need urgent help and would greatly appreciate the guidance of an expert. I take this opportunity to thank you for the work you do and that I am discovering little by little.
Thanks for your kind feedback!
I enter to evaluate my model: The objective of my research is to analyze the relationship between the entrepreneurial environment (a variable composed of 12 indicators) and the entrepreneurship (2 variables composed of 1 indicator each).
Does this mean that entrepreneurship is some sort of higher-order model? Why don't you use entrepreneurship with two (e.g., formative) indicators.
In addition, all this is moderated by a categorical variable referring to a group of countries of 9, 27 and 24 countries respectively.
You can use this categorical variables as moderator variable or - if you have enough data - to form groups and to run a multigroup analysis.
What would be the next steps? I have done an analysis of permutations and I can observe that the assumption of compound invariance is not fulfilled. How can I solve that?
There is no straight forward way to solve this. The result simply says that the constructs are different per group and, thus, the differences in the path coefficents may steem from measuring two different things per group.
I have been able to verify that there are significant differences for the TAP variable and there are no such differences for the APR, although observing the Path coefficients if we can see that there are differences, why does this happen?
Also, I need to know:
How can I study the influence of the two entrepreneurial variables in each of the conditions (indicators) of the environment ?.
Should I perform the PLS algorithm and bootstrapping to analyze the measurement model and the structural model for the general model ?.
Where can I look at the influence it has had in including the moderator variable over the original model?
Unfortunately no idea. I woudld try running a moderator analysis. You find more details in the moderator chapter of this book:
https://www.smartpls.com/documentation/ ... s-sem-book

Alternatively, why don't you simply include the categoritcal variables as single item construct control variable into your model.

Best regards
Christian
Post Reply