Original practice · fully worked
Original variant: covariance of two ordered readings
Two independent calibration readings are uniformly distributed between 2 and 8. Let L be the smaller reading and H the larger reading. Calculate Cov(L,H).
- A
1 square unit - B
4 square units - C
6 square units - D
24 square units - E
25 square units
Answer and calculation
After scaling to two unit-uniform readings, the ordered pair has means 1/3 and 2/3, while the product of the minimum and maximum has mean 1/4. Its covariance is 1/36, and scaling both readings by 6 multiplies covariance by 36, giving 1 square unit and choice A.
Setup
Setup
Scale the readings to independent unit-uniform variables and order them.
Xi=2+6Ui,Ui∼Uniform(0,1) A=min(U1,U2),B=max(U1,U2) L=2+6A,H=2+6B Model
Model
Use the marginal order-statistic means and the fact that ordering does not change the product.
E[A]=31,E[B]=32 AB=U1U2,E[AB]=E[U1]E[U2]=41 Compute
Compute
Compute covariance on the unit scale, then apply the two affine scale factors.
Cov(A,B)=41−(31)(32)=361 Cov(L,H)=6(6)Cov(A,B)=36(361)=1 Answer
Answer
The covariance of the smaller and larger readings is 1 square unit.
1 square unit(A)