This interval is very wide because there was considerable variability in the results from one store to the next. We recommend further consideration of the new packaging. If there are no large cost disadvantages, the data suggest that we should proceed. B) Technical issue: Does the treatment effect depend on the past average sales?