Original practice · fully worked
Original variant: calibrate a sensor-error tail
A sensor's measurement error X is uniformly distributed on an unknown interval [L,U]. Its 10th percentile is -3 units, and its 70th percentile is 9 units. Calculate the probability that X exceeds 5 units.
- A
0.250 - B
0.333 - C
0.500 - D
0.750 - E
0.900
Answer and calculation
The two quantiles reveal a 20-unit support from -5 to 15. The threshold 5 is its midpoint, so the upper-tail probability is 0.500 and choice C is correct.
Setup
Setup
Express both stated quantiles in terms of the unknown lower endpoint and the support width W=U-L.
L+0.10W=−3 L+0.70W=9 Model
Model
Subtract the equations to determine the width, then back-substitute to recover both endpoints.
Compute
Compute
The lower endpoint is -5 and the upper endpoint is 15. For a uniform variable, the tail probability is the upper-tail length divided by the full width.
L=−3−0.10(20)=−5,U=−5+20=15 Pr(X>5)=15−(−5)15−5=2010=0.500 Answer
Answer
The probability of an error above 5 units is 0.500.
0.500(C)