The data that was presented in the Ottawa Citizen article, Measurement Canada's data, has what we call in the statistical world a data skew from what you would expect from a normal distribution. So you would expect that 50% of the time the pump would fail or wear in favour of the consumer and 50% of the time in favour of the retailer.
In the data that was presented--I've heard two different numbers--the Ottawa Citizen said 74% rather than 50% were in favour of the retailer. Measurement Canada has since told me the number is 65%. So it's something more than 50%. So there is a skew of somewhere between 15% and 24%, depending on that data.
The question is why, and it's a good question.