OK, we have a scientific study versus a document from an organization which is against pitbulls - which one do we believe?
Even if we are not sure, let's look how those numbers are created:
- "Compiled by the editor of ANIMAL PEOPLE from press accounts since 1982" - so they take press articles as the source for their numbers. This is a huge bias. First, its not reliable, second, you have a selection bias. It's the same when you would use this kind of data to determine crime rates based on nationality. Most newspaper either print selectively (more likely a story is made if a pitbull bites someone compared to another "non-dangerous" dog), or leave the breed out in cases where its not one of the "dangerous" dogs.
- For many incidents, the dog escapes after the bite. Afterwards, asked for the race, the choice pitbull is probably more likely than others, which is another bias
- There are no reliable information about the breeds in the US, so the % of dogs numbers are just wild guesses. These are, however, pretty relevant to put any absolute figure into context.