In my survey, participants evaluated a set of paintings on liking (my dependent variable), object recognition (my independent variable 1) and colour (my independent variable 2). Each participant evaluated 10 paintings which belong to 10 different types (e.g., cubism, classic, abstract, etc.).
Firstly, I conducted a regression to see how my independent variables explain my dependent variable, taking into account the nested data: participants (as factor 1, or ID) and painting type (as factor 2, or Type). I used lmer function in R.
However, I suspect that the relations between Liking and independent variables vary as a function of paintings' type. That is, var1 and var2 can be positively related to liking of abstract paintings, negatively - to liking of cubism paintings, and unrelated - to liking of classic paintings, etc. To test this on a full scale analysis, I need to calculate the moderation effect of paintings' type on the link between liking and my independent variables. The question is: how can I do that? The paintings' type is not a continuous variable, it's categorical - in involves 10 different categories of paintings.
Thanks for your advice!
