I have a variable called Concealment of estimated categories 1, 2, 3, and 4, representing (0-25%, 25-50%, 50-75%, 75-100%) respectively. I want to include this category in mixed effects models - as a fixed effect in some models and as a response in others. I am interested in Concealment's effect on a different numeric response and in other variables' effects on Concealment. (Quantifying the effect super precisely isn't necessarily super important)
What is the better variable class for this purpose in R - ordered factor or numeric?
In case it matters: My other variables are either factors or numeric running between 0-100 and -100 - 100. Most of my models are linear mixed effects models with a random effect for individual - although I understand I can't run linear models with Concealment as an ordinal response.
I've run model selection on those with Concealment as a fixed effect a few times and changing the class of this variable doesn't seem to make a marked difference in which models come out as best.
When I run the ordinal models instead of 1,2,3,4 I get L, Q, and C in the model summary for Concealment - is this representing a Linear, Quadratic, and Cubic relationship between Concealment and the response?
Thanks!!!!!!!!