Replies

Blogs are a great way to refute studies.

I read the blog, and will note here the most obvious problem with the blog. They rightly note that the study clearly indicates that it only covers a small percentage of all dog bites, namely those that are reported, and for which breeds are identified by people who have the training and education to do so.

Then, they also correctly point out that popular information suggests that there are many more bites than what are contained in the report -- not surprising since the report said it was a subset.

But then they attempt to take the numbers from the report, treat them as complete (for pit bulls), and use them in a numerator where the denominator is the total popular reported bite count:

If Clifton’s pit bull numbers are correct, and no more than 49 of the 6,000 or so hospitalizations due to severe dog bites in the U.S. each year are a result of pit bull bites or attacks, then pit bulls and pit mixes are responsible for less than one percent of those hospitalizations.

That's not even a clever deception, it's just silly.

The main argument seems to be that we simply don't have a good database of dog bites, and therefore pit bulls can't be worse than others.

Which begs the question -- if you argue that there is no accurate information, how could you possibly KNOW that pit bulls were no worse than others?

But if you treat the clifton study as a sample of the whole, the question isn't whether it's the whole, it's whether their sampling method has any rational basis, and if so, what the sample error might be.

Unfortunately, the blogger didn't directly deal with that, at least not in a methodical way:

Clifton’s report never mentions that there is a huge discrepancy between actual hospital records and press accounts of dog attacks --- between relatively objective data, in other words, and highly subjective reporting and editing with an eye to selling papers.

Why would the hospital records contain accurate reports of the actual breed type for a dog bite? I presume at some point the police would have the dog, and there might well be a rabies test, but there's no reason to believe a hospital would be any more likely to bring a dog breed expert and include their conclusions in hospital records (which are about human injuries).

In fact, I'd guess the police reports would actually be a better source than hospital records; the police would be more likely to get a dog expert, not necessarily to get breed, but because they might have to have the dog examined for disease and to have them evaluated.

Moreso, police reports would be public records, so we should be able to get them for all reported dog bites.

Of course, that's what newspapers do -- they get police reports, and report what those reports say. Yes, we all know media sucks, but their JOB is to get accurate information and report on criminal activity, which would include dog bites.

Lastly, they argue that pit bulls are the "most popular breed" in the country, based on an extremely small sample of dogs IN SHELTERS:

This might be a good place to mention that the pit bull is one of the most popular breeds [or types] in the country. Using shelter numbers as a very rough means of estimating the number of pit bulls [registered and unregistered] in the general population, even low estimates end up in the millions. A board member of the California Animal Control Directors Association [CACDA] told me in 2005 that only labs and lab mixes are more common in California shelters. On sites like this, out of a total U.S. population of over 70 million dogs you’ll find estimates of 3 million to 10 million pit bulls.

Completely ignoring that dogs in shelters are NOT an accurate reflection of the dog population, and that a common reason for a dog to be in a pen is that a person bought the dog, but the dog was untrainable, or too violent to keep around. You might actually argue that the high numbers in shelters supports the "too violent" theory.

Worse, taking their UPPER LIMIT (10 million) and accepting their total number (70 million), pits would make up less than 20% of the total, NOT a majority (yes, they said "most popular", but that's meaningless when comparing actual numbers, and it's clear they wanted to suggest that the reported deaths and maimings could be explained by popularity).

To remind people, the clifton report suggested 40% of the deaths were caused by pit bulls, which even in the flawed logic of the blog make up only 16% of the population of dogs.