I'm running multivariate mixed effects linear meta-analytic models in R. One of my predictor variables is categorical, but each observation generally has multiple categorical values. I'm not sure how to best incorporate this variable as a fixed effect in the model.
To make it more concrete, the variable is "activity type" and the data look something like this:
Observation A: recreation, hunting
Observation B: hunting
Observation C: recreation, roads
... etc. There are 7 categories in total across >150 observations.
I could include each activity type as its own binary variable (perhaps w/ interaction terms to account for multiple activities), but I would like some way to assess the overall effect of "activity type" if possible. Is there a way to have a matrix as a predictor variable?
All suggestions welcome!