Does "random effect" used here for the clm function (in ordinal package in R) the same as "fixed effect"?

Question

I am trying to understand the the clm function to use it for a mixed effect ordered logistic regression. Referring to the following package

In the following data, the variable 'judge' can take 9 different values. Nine judges each assessed wine (the 'rating' variable) from two bottles from each of the four treatment conditions, hence there are 72 observations in all.

I am focusing on random effects and fixed effects for a mixed effect model. Given the data, the variables "temp" and "contact" are to be considered for random effects and the variable "judge" should be considered for fixed effects. However, this package used the phrase random effect with respect to the "judge" variable. This is really confusing. Is there something wrong in the way I am reading it?

 clmm2(rating ~ temp + contact, random=judge, data=wine)

Answer 1

There is either something wrong with what you are reading or something wrong with how you are reading it (and you would not be alone! Random and fixed effects are confusing and some statisticians, such as Gelman, prefer not to use the terms at all).

Here, just based on the variable names, it is almost certain that temp and contact are fixed effects and judge is a random effect. I am guessing that the judges in your sample were a subset of all possible judges and that you are not particularly interested in the effect of these particular judges. You need to include "judge" because each judge had multiple observations and, therefore, the errors are not independent.

On the other hand, temp and contact sound like fixed effects. These are what you would include, even if each judge had only one observation.