No, You Probably Can’t Guess That Dog’s Breed — and a New Study Explains Why It Matters
Your phone rings. When you answer, your friend says, "Hi! I just got a new dog!"
And naturally, you respond, "That's great! What kind is it?"
Every dog lover in the world has had this conversation, or at least one along the same lines. But if the dog in question is a mixed breed, chances are good that the answer to that question will be somewhat (if not totally) inaccurate. In many cases, it really doesn't matter; you were simply curious. But in other circumstances, the answer can be the difference between life and death for a dog.
A recent study by the Maddie's Shelter Medicine Program at the University of Florida has shown that dog experts of all kinds have one thing in common when it comes to identifying predominant breeds in mixed-breed dogs: Their assessments are stunningly unreliable. This is confirmation of what many in shelter medicine already knew to be true, but the findings are far-reaching and affect everything from apartments and insurance companies that have breed-specific regulations to rescues trying to match pets to families. In addition, this inability to recognize what breed a dog is endangers dogs facing breed bans and can hinder properly identifying lost and found dogs in shelters.
To learn more about the study and its applications, we spoke to Dr. July Levy, DVM, a professor of shelter medicine at the University of Florida and leader of the dog breed identification study.
What Breed is Dog 13?
Top Responses: Border Collie, German Shepherd Dog, Basenji, Great Dane, Boston Terrier
DNA Results: 25% German Shepherd, 25% Staffordshire Bull Terrier, 13.36% Weimeraner, 7.29% German Wirehaired Pointer
What Breed is Dog 88?
Top Responses: American Bulldog, American Staffordshire Terrier, Boxer, Dalmatian, Argentine Dogo
DNA Results: 25% Plott Hound, 25% Boston Terrier, 25% German Spitz, 11.68% Saluki
What Breed is Dog 2?
Top Responses: Labrador Retriever, American Staffordshire Terrier, No Predominant Breed, Border Collie, Pointer (includes English Pointer)
DNA Results: 50% Catahoula Leopard Dog, 25% Siberian Husky, 9.94% Briard, 5.07% Airedale Terrier
What Breed is Dog 68?
Top Responses: German Shepherd Dog, No Predominant Breed, Rhodesian Ridgeback, Anatolian Shepherd Dog, Belgian Malinois
DNA Results: 25% American Staffordshire Terrier, 25% French Bulldog, 25% American Foxhound, 22.13% Belgian Tervuren
What Breed is Dog 57?
Top Responses: Labrador Retriever, German Shepherd Dog, No Predominant Breed, Golden Retriever, Anatolian Shepherd Dog
DNA Results: 25% Beauceron, 25% Siberian Husky, 25% American Staffordshire Terrier, 12.73% Schipperke
What Breed is Dog 71?
Top Responses: Chihuahua, No Predominant Breed, Labrador Retriever, Border Collie, Schipperke
DNA Results: 25% Miniature Pinscher, 25% Brittany Spaniel, 25% Chinese Crested, 12.83% German Spitz
What Breed is Dog 64?
Top Responses: Beagle, No Predominant Breed, German Shepherd Dog, Corgi (including Cardigan, Pembroke), Basenji
DNA Results: 50% Russell Terrier, 50% Plott Hound
What Breed is Dog 99?
Top Responses: Labrador Retriever, No Predominant Breed, Pointer (includes English Pointer), Brittany, Foxhound (including American, English, Treeing Walker Coonhound)
DNA Results: 25% American Staffordshire Terrier, 25% Collie, 21.41% Black Russian Terrier, 19.86% Norwegian Buhund
What Breed is Dog 7?
Top Responses: Labrador Retriever, American Staffordshire Terrier, No Predominant Breed, German Shepherd Dog, Staffordshire Bull Terrier
DNA Results: 25% Irish Water Spaniel, 25% Siberian Husky, 25% Boston Terrier, 8.33% Bull Mastiff
What Breed is Dog 55?
Top Responses: Doberman Pinscher, Labrador Retriever, No Predominant Breed, Rottweiler, Dachshund (including miniature, standard, smooth coat, wirehair, longhair)
DNA Results: 25% Great Dane, 25% Schipperke, 12.5% Chow Chow, 12.5% Collie
What Breed is Dog 79?
Top Responses: No Predominant Breed, American Staffordshire Terrier, Boxer, Beagle, Collie
DNA Results: 25% Doberman Pinscher, 25% Wirehaired Dachshund, 12.5% Samoyed, 12.5% Miniature Schnauzer
Why Visual Identification Matters
The UF study isn't the first to tackle the topic of breed identification, but it's certainly the most comprehensive, and it all began when Levy and her team began investigating which dogs were most at risk for ending up in a shelter, and then which were most at risk of euthanasia because they weren't placed in a home. "One of the issues is breed identification," says Levy. "Certainly some breeds are more popular for adoption than others." As those involved in animal rescue know, Pit Bull–type dogs don't tend to be the pick of the litter. But how much of that has to do with the dogs themselves and how much has to do with the name of the breed with which it is associated?
Aside from helping more mixed-breed dogs find homes, showing how unreliable breed identification is serves several other purposes. For instance, if a family is using breed identification to find a dog that's a good fit based on that breed's typical behavior, or picking a puppy that's expected to stay within an apartment's size regulations as it grows, knowing the limitations of a visual assessment could keep a dog from being returned after adoption, Levy explains.
"It also comes up in lost and found searches," Levy says. "A lot of times, lost and found searches are based on breed, so an owner might come in [to a shelter] and describe their missing dog as a Lab mix, but the shelter might’ve recorded that black dog as a Pit Bull mix or a Shepherd mix. If the search is done purely by breed, they might not match at all, even though that dog is at the shelter or described in a found poster."
Looks Can Be Deceiving
To study this more rigorously, the team enrolled four different Florida animal shelters, selected 30 Pit Bull–type dogs at each, and asked several staff members to guess the predominant breeds in those dogs, then compared those guesses to DNA results. "We found that based on DNA analysis, staff missed quite a few Pit Bull–type dogs in their breed identification, and also identified non–Pit Bull dogs as Pit Bulls," Levy says.
The study recently went nationwide. "We showed pictures of the dogs to over 5,000 dog experts nationwide," Levy says. "They were asked to identify the predominant breeds of these dogs, and once again we found that only a small minority could pick out even one of the breeds that was identified in these dogs by DNA. We’ve shown repeatedly and confirmed in our study that you really can’t look at a dog and know reliably whether it’s a Pit Bull or not, at least in many cases. And yet life and death decisions are being made based on those visual assessments."
What's the Next Step?
So now that we know that visual breed identification is unreliable at best, what can we do about it? That's tricky.
"We’re frustrated by our inability to come up with a better idea. Based upon our confirmation of the inaccuracy of identifying breeds in a national sample of dog experts, we can now clearly say that guessing breeds is not reliable, but we haven’t come up with an alternative," Levy says. "It would be nice for the next step to be leaders with an interest in this area convening some kind of summit to put together protocols or suggestions for a better way to identify dogs."
In the meantime, some shelters are taking measures to try to identify dogs by methods other than breed identification. "They're describing things everyone would agree on, like calling it a black dog or a male dog," Levy says, admitting that even that has the possiblity for error, particularly with more complicated colors.
Still, it's a start. Will this change the way you describe mixed-breed dogs, or the way you view Pit Bull–type dogs?