Fitting Mixed Effects Models With Big Data