A school has students and teachers. In the first period, each student is taking one class, and each teacher is teaching one class. The enrollments in the classes are and . Let be the average value obtained if a teacher is picked at random and the number of students in their class is noted. Let be the average value obtained if a student was picked at random and the number of students in their class, including the student, is noted. What is ?
- A)
- B)
- C)
- D)
- E)
Answer
B
Key insight
t is the plain average 20; s is size-weighted, (50^2 + 2*20^2 + 2*5^2)/100 = 33.5, so t - s = -13.5.
Solution
For , each of the teachers is equally likely, so is the ordinary mean of the five class sizes:
For , each of the students is equally likely, and a student in a class of size reports . There are students who report , who report , and who report :
Therefore .
The answer is .
Why this works
Averaging over classes weights each class equally; averaging over students weights each class by its size, so large classes count more. This "size-biased" average is always at least the plain average (they agree only when all classes are equal), which tells you in advance that is negative. The same effect explains why "average class size as experienced by students" exceeds the school's official average.
The trap
Assuming s equals t because 'the average class size is 20', ignoring that a random student is far more likely to be in the big class.
Common mistakes
- Assuming s equals t because 'the average class size is 20', ignoring that a random student is far more likely to be in the big class.
- Getting the sign backwards and answering , choice (D); the question asks for , and is the larger one.
Techniques
Set up the equation/formula and compute; no special trick needed