Randomly sample 50 schools with probability proportional to the number of students in the school. Within each selected school, randomly sample 10 students. Pretend for a minute that you used SRS and calculate the mean and standard error of baseline test scores (total_bl) in your sample. Now account for the fact that you first selected schools and then selected students, and recalculate the mean and standard error of total_bl in your sample. Ignore any finite population correction.