ROFLMAO! I don't know if you know it, but that is just the best joke I've heard in a long time!
You're taking the number of deaths in one set of years, then deciding that it remains the same in a different set of years to change 21% into 14.7%. It's hysterical! By that logic, since the number of deaths from other dogs would also remain static, there must be a fair number of dog caused deaths that were caused by no dog. Amazing! I think I'll prove that Fords are only involved in .01% of all car crashes by taking the number of car crashes involving a Ford in 1920 (let's say 200), using that number in the total number of car crashes in 2004 (let's say 200,000) - and there we go, Ford only crashes 200 times a year, so they are only in .01% of all crashes - so Ford is the safest car in the whole world!
I'm just curious - was that deliberate, or did you really think you were telling the truth? I think it's too obvious to be accidental, but I am curious.