Survey results are often accepted unquestionably without considering the representative database, the validity of the results, or even, whether or not the findings make sense. Because it is written down, it must be the gospel truth. Quite frankly though, I have yet to read even a single published survey on the breed incidence of biting, which would receive a grade better than F in a kindergarten midterm examination. More disturbing, once these hopelessly unreliable and unrepresentative surveys are edited, expurgated, bowdlerised and summarized to an hundred words or less for popular consumption by our ant-brained tabloid media, the remaining newsprint is hardly fit for paper-training puppies. It would not be so bad if the surveys were merely frivolous. Unfortunately such ill-conceived “results” sow the seeds for breedism, which, fostered by hard-wired, worker-reporters in the media, flourish into undeserved bad reputations for some breeds in particular and for all dogs in general.

Some years ago, I reviewed a score of surveys (on the relative incidence of biting in different breeds in the United States), which had been published over the past twenty years. Please do not pay heed to the results — they are all but meaningless — and are only included to illustrate just how silly so-called researchers can be at times. Averaging the results of all twenty studies, produced the following overall ranking of dogs involved in bite incidents: way ahead of all other breeds and occupying undisputed first place — number 1. The German Shepherd, followed by 2. Cocker Spaniel, 3. Poodle, 4. Collie, 5. Dachshund, 6. Labrador Retriever, 7. Dobermann Pinscher, 8. Miniature Schnauzer, 9. Springer Spaniel, and 10. Old English Sheepdog.

Breed Bias

All the surveys were strongly biased towards the more popular breeds.  Difficult though it may be to believe, none of the twenty studies even considered breed popularity and the effect it might have on the number of bites by each breed. Of course one can expect more bites from cockers, poodles, Labradors, shepherds, beagles, miniature schnauzers and Dachshunds, since these represented seven out the top ten most popular breeds in the United States for that time period. Until the results have been controlled for differential breed popularity, it is absolutely unfounded to state that shepherds, cockers and poodles are more likely to bite than other breeds of dog. Such statements are asinine twaddle.   

Body Size Bias

An important question about any survey concerns the source of the data. Most dog bite surveys (including eighteen of those averaged above) are based on public health reports of reported dog bites. However, from my bite-case histories, it is frighteningly apparent that less than 1 out of 20 bites are actually reported. Indeed, the vast majority of bites occur in the home and usually the dog lives with, or at least knows, the bite victim, who is most often a child, or a man. Indeed, when bitten by the family dog, few people are going to report the bite to public health, medical, or police authorities. Hence, reported bites are an extremely unrepresentative database, since they reflect only a fraction of the bites that actually occur. In addition, public health surveys are strongly biased towards the type of bite that is likely to be reported, i.e., severe bites and/or incidents in which the victim did not know the dog, for example when the bite occurred on public property with the dog running at large. Severe bites are most commonly caused by large and/or powerful dogs. Also, the profile of a free-ranging dog is a medium-sized to large dog, usually a mixed breed. Thus, public health surveys based on reported bites are strongly biased towards large and powerful breeds and towards mixed-breeds.

Controlling for Bias

To control for bias against breed popularity and large body size, I divided the mean number of reported bites for each breed by the number of dogs of that breed registered by the American Kennel Club and by the average weight of that breed in pounds. Even so, the results still did not make sense: German shepherds still remained way (WAY) in the lead, followed by Labradors, huskies and terriers, with the rest of the pack all in a bunch. The results were not what I expected. I had discussed bite-incidence with many dog professionals and most would agree that: all breeds of dog bite people; little dogs are apt to bite more frequently (largely because they get away with it), but larger dogs cause more damage. So how come the results of these studies did not reflect this obvious common sense?

Erroneous Breed Classification

The reason for the illogical bite distribution was even more fundamental than I had first thought. The major confounding of all these studies lies with the original breed assignation of the biting dog. Since mixed-breeds are the most common type of dog, presumably they should be at the top of every bite survey. However, of the studies I reviewed, mixed-breeds were ranked #1 in only one survey and placed in the “top ten” in only five others. Mixed breeds were inexplicably absent from the remaining 14 studies, in which there was not even a category for bites by mixed-breeds. Unbelievable! So how on earth were the substantial number of mixed-breed bites classified? "Who got the mongrel vote?" Prototypical Dog - that's who! Presumably bites by mongrels were classified as the closest resembling pure-breed.

Now, if four dissimilar pure-breeds e.g., Lhasa Apso, Cardigan Corgi, Saluki and  Newfoundland, were allowed to interbreed willy nilly, within just two generations, the mixed-breed offspring would largely resemble shepherds, Labradors, huskies and terriers types, even though these dogs have had no shepherd, Lab, husky, or terrier blood in their veins for hundreds of generations. This phenomenon is known as atavism, whereby the interbreeding of highly selected strains causes a reversion to the wild type, i.e., wolfy-lookalikes (shepherds and huskies) and prototypical dogs (Lab-pointer types). Thus, many mongrel bites are erroneously classified as shepherd, Labrador and husky bites because the dogs looked like those breeds, i.e., the dogs were classified by phenotype (looks) rather than genotype (geneology). All black dogs were classified as Labradors. Large dogs with sticky-up ears were called shepherds, but sticky-up ears and a curly tail would put the dog in the husky category. Similarly, little dogs were invariably called cockerpoos or terriers and the most sophisticated classification of all, small brown Heinz 57's were invariably and expertly categorized as shepherd/terrier crosses. Wow! A bona fide first filial shepherd/terrier hybrid! How on earth did they tell?

In so many studies the very data is basically questionable and therefore, the results are utterly invalid. GIGO — Garbage in - garbage out. No amount of fancy statistics can make a survey respectable if the original data base is a bunch of baloney. Again, if the fallout from these slipshod studies were not so serious, the whole thing would be laughable.

The weak link in the chain lies in the identification of the breed of the biter. For example, when a child is bitten by an unknown dog, it is usually the child and the doctor who arrive at a breed diagnosis... I'm sorry, but this has to be a case of the blind leading the blind. No one can convince me, that even the cutest and most intelligent of children along with the most competent of physicians could ever agree on the correct classification of a Wirehaired Pointing Griffon bite, for example, that is if they could even distinguish a Griffon from a hole in the ground. Most likely, the bite would be categorized as a German shepherd/American cocker bite.

Several years ago, I conducted an informal survey of ten eight-year-old children and ten physicians. I showed all twenty subjects 146 black and white photographs from my faithful five shilling, 1965 Revised Edition of The Observer's Book of Dogs. Surprisingly, (well, maybe not so surprisingly), American eight-year-olds recognized more breeds that did American physicians. However, even more to my surprise, neither group correctly identified more than twenty breeds of dog!

Thus, potential biting-breed categories must be limited to those few breeds familiar to the victim and the physician. And of course, the most recognizable breeds are likely to be the most popular (yet another bias of popularity), the most unique and those made famous by the arts and the media. When a purebred or mixed-breed biter is incorrectly identified, it is most commonly miss-labeled as one of the more well-known breeds. For example, in the sixties and seventies, a number of public health studies ranked collies as the most biting breed. During that period of course, thanks to Lassie, collies were extremely popular, easily recognizable and known by all. It is extremely unlikely that collies were biting significantly more than other breeds; rather there were oodles of collies, and every large, hairy biting dog was given the collie classification. Similarly, in the seventies, black and tan biters were called Dobermans, whereas nowadays they are called Rottweilers.

Breed Reputation

Sometimes the arts and the media create a different kind of reputation; different breeds become infamous rather than famous. Generally, large and powerful breeds, rapidly increasing in popularity acquire bad reputations. It is hard to say precisely how and why, but once established, unwarranted reputations grow in leaps and bounds, magnified by further unrepresentative media coverage.

For years, the media trashed German Shepherds and Doberman Pinschers, then Rottweilers and more recently, pitbulls, the Tosa  Inu, Fila Brasilieros and Dogo Argentino. Very few people are even familiar with the last three breeds but we'll soon know what they look like though, courtesy of the media. And just how different is the generic “pitbull” from the American Pitbull Terrier, American Staffordshire Terrier, Staffordshire Bull Terrier, Bull Terrier, etc? Soon we can expect similar vendettas to be leveled against Chows, Akitas and Shar Peis. Then, what next? Retrievers for Heavens sake?

Fighting Dogs

The whole “fighting-dog” argument is one big red herring and too silly for words. Albeit an illegal activity, fighting dogs, by definition, have to be non-aggressive towards people, otherwise they might injure the handlers and judge in the ring. Also, most breeds of dog were, at one time or another, raised and trained to hunt and/or kill one life-form or another. Some were even bred to kill dogs and some bred to kill humans. In fact, most of the “mastiff-type” dogs, (dogs with a characteristically short muzzle, pronounced stop and hangy-down ears), were thought to be descended from the giant Molossian and Hyrcanian mastiff war dogs that were bred, raised and trained to kill and eat people!. Since mastiff-type dogs comprise the majority of modern day breeds, including nearly all working dogs, herding dogs and gundogs, ban these breeds and we'll have no dogs left.

Breed X

To be fair to the media, with the notable exception of sports and weather, by definition, “news” is totally unrepresentative — it has to be. If the events were truly representative of everyday living, they would not be newsworthy. Who, for example, would be astounded to hear, that for yet another day in the United States, over three hundred million people have not been savaged by breed X? On the other hand, if a dog remotely resembling breed X were to merely look grumpy, let alone growl and snap, it would most likely be reported on the front page in headlines higher than dog's hocks.

Once a bad breed reputation is established, whether warranted or not, a vicious circle quickly develops, whereby differential and unrepresentative media coverage increases the recognition of breed X, which in turn increases the likelihood that breed X bites will be reported (and therefore included in surveys) and also, increases the likelihood that bites from mixed-breeds and other purebreds will be misclassified as breed X bites. This in turn, increases the media coverage of bad breed X. At which point breed X becomes a hot prospect for some pea-brained jerk, who intentionally wants to create a bad dog. More fuel for the media.

Eventually, the media-manufactured reputation in itself becomes a primary and direct cause of dog bites. Public scare campaigns foster a fear of dogs in general and of breed X in particular. And without a doubt, fear, or uneasiness towards dogs is the most potent, provocative cue to cause dogs to feel apprehensive and uneasy around people and maybe even defend themselves. To be a little more specific, when people are scared of a dog, they are much more likely to be bitten. Thus, sensational media-coverage is also a major contributory factor for dog bites. For example, a gaggle of fleeing, flailing and screaming children, who think they have just encountered big bad breed X of television notoriety, is the very stimulus that might excite and incite a lesser-proofed dog to give chase and nip and maybe bite. Whereas, if the children were not afraid, most probably the dog would have just ignored them, passed them by and happily gone about his daily doggy business of sniffing and peeing.

The Danger of Dogs

Certainly, dogs can be dangerous; they frequently bite, very infrequently maim and only extremely rarely, kill people. Certainly, each death is a heart-rending tragedy. But it is nonetheless essential to maintain perspective and to consider the danger of dogs relative to other dangers.

The incidence of dogs biting people is extremely high. In the United States, there are usually over one million reported dog bites each year and since only one out of every twenty bites are actually reported, the annual toll exceeds twenty million bites. However, very few dog attacks leave the victim maimed. Rather, most incidents involve a only a single nip and very few bites even puncture the skin. Very few people are actually killed by dogs and even fewer children.

In the United States, an average of twenty people, half of them children, are killed by dogs each year. This is indeed terrible, but more children are killed annually by farm animals. More children are killed in playgrounds, drowned in swimming pools, or choked on balloons.  And last year in the United States, 2000 children were killed — not by dogs — but by their parents. Each year in the US, 2000 children are killed by their parents. This is the problem each year.  And of course, many, many, many more children are killed by guns and by cars. The dog bite problem is serious but we must always keep it in perspective.

The Solution
The solution is not to ban specific foreign makes of automobile. Instead we ban cars that are not road-worthy, we teach people how to drive and we penalize irresponsible drivers. Similarly, banning specific foreign breeds of dog will not prevent dog bites. Instead, as a short-term solution to protect the public from dangerous dogs, it is only common sense to confine and/or muzzle dogs that have nipped or bitten (i.e., keep them off the streets) and although I love dogs, I really think it is kindest to all concerned to euthanize dogs that have maimed or killed. As a long-term solution though, dog owners must be taught, how to teach their dogs bite inhibition (so that they cause no damage) and how to socialize dogs to people (so that they feel no need to bite People). The Lassie myth — relying on the dog to magically develop a dandy demeanor and a terrific temperament backfires all too often. Good breeding is great but once dog is born, socialization and training are the only ways to better behavior and temperament.  Bite inhibition, socialization, handling, gentling, confidence-building and proofing should all be proactive training exercises — part and parcel of routine canine husbandry. Regardless of breed and breeding, we know it works. So, let’s do it.

To hear more of Dr. Dunbar's insights check out his live appearance schedule, he's likely coming to a town near you!

This article is based on Dr. Dunbar's Behavior column in the October 1989 issue of the American Kennel Gazette. Reprinted with the permission of the author and the American Kennel Club.