The experiment is repeated. This time, the operator forgets to apply the coatings, so each of the 20 wear measurements is actually made on uncoated gears. Is it likely that one or more of the formulations will appear to reduce wear, in that H_{0} will be rejected?
Yes. Example 6.20 shows that the probability is 0.64 that one or more of the coatings will appear to reduce wear, even if they are not actually applied.