I tried to dummy code my categorical covariate (nationality: 8 groups), however when I run the mediation analysis (with the amount of bootstrap samples set to 5000) I get e.g. "NOTE: Due to estimation problems, some bootstrap samples had to be replaced. The number of times this happened was: 4418". The latter number varies per analysis, though it is always >4000. This "error" does not happen with a dichotomous covariate (2 groups).
My question then is: how can one properly control for categorical covariates with more than 2 groups in PROCESS? Or is this perhaps impossible in this program?
Thanks in advance.
Add categorical variables to covariates and run the test. If the significance values of categorical values are insignificant, that shows that those variables have no effect on the model or in other words, controlled ones.
As per your question, you need to place bootstrap samples at 4000 and then run. As i have done it with 6 groups at maximum.
I recommend my comment, Jamovi is an open source program more robust than PROCESS, easy to use and includes standardized beta effect estimation
https://www.researchgate.net/post/Has_anyone_used_JASP_and_Jamovi