I need to calculate robust standard errors for a mixed effects model with a large dataset (200000 observations).
I've been trying to use the vcovCR function from the clubSandwich package but after 2 hours I get an error saying "cannot allocate vector of size 221.2 Mb". I think this is related to the size of my dataset as it works fine when I reduce the dataset to 10000 observations. I tried calculating robust standard errors by copying a method online but the first step was to generate a matrix and the size of this was 337 gb! Are there any other options for calculating robust standard errors for a mixed effects model with a large dataset?