I am relatively new to this forum (and statistics) so apologies if this is a naive question.
I am exploring whether a categorical variable (low, medium and high traits) moderates the impact of relationships between three predictor variables and an outcome variable (all continuous).
The original regression model (the three predictors on the outcome) is a well tested theoretical model that has been shown to be robust with lots of different data and studies. A meta-analysis even said it was robust to moderating influences of demographics, etc.
However, the papers I have read did not say how the moderation analyses were done. Should all of the interaction effects be included in the model, along with all of the predictor variables? Or should separate models be run, with each predictor variable and its interaction?
Thanks in advance