Hi! In the case of the variables we are referring to, it depends on what the theory states. The theory that being a foreigner/citizen increases or decreases community acceptance and happiness index or these two variables achieve different values depending on the sample group analyzed: foreigner or citizen. For the first case foreigner/citizen would be a independent variable, while for the second case it would be a variable used to create groups in the sample in order to to address heterogeneity. The same thing is valid for the gender. So, the hypothesis you are testing on your PLS-SEM model depends on the theory and, not that much on statistics. From my point of view, both foreigner/citizen and gender should be used to create groups and perform multigroup analysis in order to test the relationship between community acceptance and happiness index. But, you should rely on the theory revision you performed.

Statically, we can use both foreigner/citizen and gender as independent variables, moderators or for multigroup analysis in a PLS-SEM. However, it is not recommended the use of variables with only one indicator because of the problems in assessing reliability. There are also some other issues in using categorical variables.

So, check first the theory. If your proposal is well grounded in the theory you can also try other statistical methods to tests your hypothesis. You can calculate the latent variables scores for community acceptance and happiness index with the help of PLS-SEM Algorithm and then use PROCESS for mediation or linear regression. All these analyses can be performed in SmartPLS4.

Yes, both the moderator and the independent variable can be categorical in Partial Least Squares Structural Equation Modeling (PLS-SEM). In PLS-SEM, categorical variables can be included as independent variables or moderators by using dummy coding or other forms of categorical encoding. Here’s a brief overview of how it works:

1. Dummy Coding of Categorical Variables: If the categorical variables have multiple categories, they can be dummy-coded (typically converted into binary variables) to include them in the PLS-SEM model. This process creates separate binary indicators for each category (e.g., if you have a categorical variable with three groups, you would create two dummy variables to represent it).

2. Interaction Terms for Moderation: When testing moderation with categorical variables, interaction terms are created between the dummy-coded variables. This allows you to assess whether the effect of the independent variable on the dependent variable changes across levels of the moderator.

3. Smart PLS and Other PLS Software: Tools like Smart PLS allow for the inclusion of categorical moderators and independent variables by creating interaction effects through coding or interaction terms.

It’s important to ensure that the dummy variables are correctly specified, as the interpretation of interactions with categorical moderators can become complex depending on the number of categories involved.

Thank you so much Vasilica Maria Margalina Shahzad Khalid

Hi Stuti! Please, check this tutorial created by Kwaja Fawad about categorical independent variables https://www.youtube.com/watch?v=Bpx5gWfjHmQ