Improvement ratio/amount

The improvement that you want a demonstration test to detect. Depending on the specified distribution, Minitab plots the probability of passing the demonstration test against improvement as either an improvement ratio or an improvement amount.

Improvement ratio—the true value divided by the minimum value you want to demonstrate. For example, an electronic components supplier has improved the design of a particular glass capacitor. The lifetime of the existing capacitor follows a Weibull distribution with a 5th percentile of 50 hours. They consider a 5th percentile of 150 hours to be a significant improvement in capacitor life. Hence, they want a demonstration test for the newly designed capacitor to pass if the improvement ratio is 150/50 = 3. The POP graph displays a plot of the probability of passing the demonstration test against the improvement ratio for the  Weibull, exponential, lognormal, and loglogistic distributions.

Improvement amount—the minimum value you want to demonstrate subtracted from the true value. For example, a manufacturer improves the design of a thermostat used inside a gas water heater. The lifetime of the existing thermostat follows a normal distribution with a 5th percentile of 1825 hours. They consider a 5th percentile of 2010 hours to be a significant improvement in thermostat life. Hence, they want a demonstration test for the newly designed thermostat to pass if the true improvement amount is 2010 - 1825 = 185 hours. The POP graph displays a plot of the probability of passing the demonstration test against the improvement amount for the smallest extreme values, normal, and logistic distributions.