Pit Bull Attack!

by Brian Dunning
December 13, 2011

Perhaps the most horrifying story to hear on the news is a case of a child being killed by a pack of dogs, hardly anything can incite a more emotional response. We're quick to vilify the dogs; perhaps justifiably so, perhaps not. In the United States, it seems that more often than not, the dogs involved in such attacks are pit bulls. Legislation is quick to address highly emotional issues, and many states now have various bans and limitations on pit bulls. Today we're going to turn our skeptical eye onto the popular belief that pit bulls are truly as dangerous as their reputation suggests.

Defining exactly what a pit bull is is not a slam-dunk. It's not a specific breed; rather it's a collection of several related breeds. Those dogs that are unambiguously pit bulls include the American pit bull terrier, the American Staffordshire terrier, and the Staffordshire bull terrier; however the bull terrier or English bull terrier is not. Some municipalities classify the American bulldog as a pit bull "type" dog. Finally, pit bulls are usually classified as any dog having the substantial physical characteristics and appearance of pit bull breeds, which establishes the somewhat unfortunately vague precedent that you know a pit bull when you see one.

Deaths by dog attack have been thoroughly studied. Perhaps the most often cited large study was published in 2000 in the Journal of the American Veterinary Medicine Association assessing 20 years of DBRF (dog bite related fatalities) in the United States, from 1979 through 1998. During that period, 238 Americans were killed by 403 dogs. Just over half of these deaths involved pit bull type dogs and Rottweilers. It's important to note that there is always some uncertainty about breed. A lot of dogs out there are not pure bred or are mixed, and numbers for those dogs were included in the study as well. But the trends over 20 years were clear: Pit bulls are indeed responsible for the most DBRFs, though in some years Rottweilers were most deadly. German shepherds are third, huskies and malamutes are next, and it goes down from there. Pit bulls killed more than seven times as many people as Doberman pinschers, which we usually consider to be so dangerous.

The authors of the study also noted one very important weakness of such studies: they look only at the dogs themselves, and not at the owners. The example they give is that of an owner who wants an aggressive dog, perhaps as a guard dog, or as an ornament for his barbed-wire bicep tattoo. An owner who wants a scary dog, and who plans to use it in a macho or antagonistic way, is much more likely to buy a pit bull to put into this role than he is a poodle or chihuahua. Some percentage of potential dog bite scenarios are always going to be set up by aggressive dog owners; so statistically, we're always going to see a correlation between dog bites and certain breeds that were selected based on reputation, whether that reputation is deserved by the dog or not.

When lacrosse coach Diane Whipple was killed by two pit bull type dogs in San Francisco in 2001, the specific breed was a Presa Canario. Sales of these shot up, driven by people who wanted the latest and greatest bad-boy dog. They were selected by aggressive people based on reputation. Indeed, the San Francisco dogs were owned by a couple who was raising them on behalf of prison inmates trying to run a dog fighting operation from their prison cell.

The other glaring weakness of this type of study is that it doesn't take into account the relative prevalence of each of these dog breeds. Maybe pit bulls killed seven times as many people as Dobermans because there are seven times as many of them out there. So what are the numbers; are pit bulls, Rottweilers, and German shepherds sufficiently popular that numbers alone can account for the number of deaths they cause?

This is, unfortunately, a question that cannot be well answered. The only real manifest of dog breed popularity in the United States is the American Kennel Club's registry. This registry includes only dogs that owners choose to register, and is highly skewed toward pure bred dogs owned by serious dog owners. It does not include anywhere near the more than 75 million dogs living in the United States. Labrador retrievers nearly always top the AKC list, yet these were listed 12th in fatalities in the 2000 study. Rottweilers, found right at the top of the DBRF list, rank down in the teens on the AKC registrations. The relative registrations of German shepherds, on the other hand, does match their position on the DBRF list, right up in the top two or three.

We might be tempted to make the following conclusions from these few data points:

1. Labrador retrievers are very safe dogs.
2. Rottweilers are very dangerous dogs.
3. German shepherds are about average.

But whether these conclusions are true or not, we don't have enough data to confirm them, because the AKC registration data does not necessarily reflect the actual prevalence of these dog breeds out in the world.

Neither should fatalities be considered a significant factor when assessing how dangerous a certain dog breed is. Another highly cited study was published in the journal Pediatrics in 1994, and it found an average of about 20 deaths a year from dogs in the United States, compared to 585,000 dog bite injuries requiring medical treatment. Dogs inflict injuries nearly 30,000 times as often as they inflict death. Clearly, injury is where the overwhelming majority of dog aggression is, not death; and it's probably where we really should be looking to determine the relative aggression of certain breeds.

The authors surveyed nearly 1,000 reports of dog bites in the city of Denver in 1991, and restricted their results to cases where they were able to contact the owners and get complete information about the dog, its history, and the circumstances of the bite. Then, for each biting dog, they found a geographically nearby control dog, of any breed, with no biting history. Dog breeds were reported by the owners, and in cases of mixed breed, dogs were listed as whatever breed the owner considered to be dominant. Since the non-biting control dogs were a random selection from the existing breed distribution in the same region as the biting dogs, the factor of breed prevalence was effectively canceled out. These authors structured their study to give us a real picture of which breeds, or other factors, most often contribute to dog bites. And here's what they found.

Surprise: German shepherds and chow chows are the big biters. Golden retrievers and standard poodles are the least likely to bite. Dogs whose distribution among the biting and non-biting populations was not significant include chihuahuas, cocker spaniels, Dobermans, Labrador retrievers, Scottish terriers, and Shetland sheepdogs. For all other dog breeds, there was insufficient data.

But where are pit bulls in that list? When the study was done, new pit bull ownerships had been banned in Denver since 1989, so there were no pit bull bites recorded in the study. This ban was based on 20 pit bull attacks in Colorado over the preceding five years. That's four a year, out of a nationwide 585,000 a year. A class of plaintiffs called the Denver Dog Fanciers tried to overturn the ban, unsuccessfully. The court's findings included:

It cannot be proven that pit bull dogs bite more than other dogs. However, there is credible evidence that pit bull dog attacks are more severe and more likely to result in fatalities.

And:

The City did prove that [pit bulls] inflict more serious wounds than other breeds. They tend to attack the deep muscles, to hold on, to shake, and to cause ripping of tissues. Pit bull attacks were compared to shark attacks.

These points do seem to be supported by the facts. Pit bulls are involved in a disproportionately high number of fatal versus non-fatal attacks, though this number is still extremely small. Pit bulls do tend to bite and hold, displaying an amazing ability to not release their grip. This has given rise to the rumor — which is completely false — that they have some physiological ability to "lock" their jaw. There's also no truth to the story that pit bulls have uniquely large jaw muscles, or have the highest measured bite strength. Pit bulls are strong, no doubt about it; but so are many other large dogs.

Only a very few studies of dog bite force have been done, and Rottweilers seem to be the strongest found so far, with a bite force of around 1,400 newtons. Pit bulls have been measured at 1,100 newtons. For comparison, hyenas can bite with four times the force, at over 4,400 newtons. But keep in mind that these numbers are from very small pilot studies.

Here are some other factors that the Pediatrics authors found. Dogs bite more often when they're male, when they're not neutered, when they're over 20 kilograms, and when they're less than five years old. Biting dogs are more likely to live in homes with children below the age of ten, are more likely to be kept chained when outdoors, and are more likely to growl at visitors. Interestingly, obedience training, guard training, and discipline styles have not been found to have a statistically significant impact on that dog's likelihood to bite.

So here's the bottom line, based on my own analysis of the available data. If you want a safe dog, avoid chow chows and German shepherds. Golden retrievers are your best bet. Pit bulls may well be a breed to avoid, but there is not definitive data to support this. Get a female or a neutered male, small, and over five years old. The fewer children around, the less likely it is to bite.

If a dog is going to bite you though, the two breeds you least want it to be are a pit bull or a Rottweiler. They are definitely the most dangerous biters, once they decide they're going to bite you. If you see one on the street, there is not sufficient data to support any particular need for concern. Like all dogs, its owner and its environment are major factors in its level of aggression.

This is a case where the value of good science is to drive policy. Most researchers agree that breed-specific legislation — a nice term for pit bull bans — are inappropriate. No good data exists to demonstrate that such bans have had any impact. Improved enforcement of existing laws, and improved education for dog owners, are far more likely to reduce the number of dog bites, fatal or not.

By Brian Dunning
Follow @BrianDunning