ABSTRACT
How do you know that the problem actually declined? Most problems vary in intensity, even when no one is responding to them. For example, there are nine vehicle thefts in a particular city centre, on average, but seldom are there weeks with exactly nine thefts. Instead, 95% of the weeks have between 7 and 11 thefts, and in 5% of the weeks less than 7 or greater than 11 thefts are recorded. Such random variation is normal. Clearly, a reduction in vehicle thefts from an average of nine per week to an average of six per week is within the norm, and might be due to randomness alone. In other words, the response to the problem may have had no impact. If the change in the problem is small, random variation is a plausible explanation. If the number of thefts had dropped to about zero per week, with an occasional week with one theft, we could be reasonably certain that this was not a chance result; it is abnormal. How can we know if the change in the problem is within normal variation?
