Averaging approach #2 (stratified pooling)


If we want to take into account the possibility that one class might be different from another, we need to calculate the difference in scores between the two types of students students within EACH CLASS before we calculate the overall average difference between their test scores. We find the overall average difference by adding the differences together and dividing by 3 (ie, the total number of classes).

Here’s what we find when we calculate the differences between scores that way:

  Students with work experience Students without work experience Difference  
Class Average Score # of students in class Average Score # of students in class  
A 83 15 80 5 83-80= 3
B 70 5 79 15 70-79= -9
C 92 10 86 10 92-86= 6

Average Difference

       

3-9+6= 0

     0/3= 0

This calculation tells you that on average, the difference between the student scores is zero. In other words, test scores for students without work experience are the same as scores for students with work experience.