Study: Judging Dogs by Breed is a Poor Way to Anticipate Behavior

It turns out, based on recent research, that we as the human race have been unfairly stereotyping dogs. In other words, canine profiling.

Breeds are, of course, most recognized for physical aesthetics, such as the bat-like ears sported by chihuahuas, the curly coats of poodles and the elongated hot dog body shape of dachshunds, otherwise known as wiener dogs; but they’re also associated with certain behaviors. The American Kennel Club describes border collies as “affectionate, smart, energetic” and beagles as “friendly, curious, merry,” while Shih-Tzus are often described by their owners as high-strung and not so kid-friendly.

But genetic information from more than 2,000 dogs paired with self-reported surveys submitted by dog owners sheds new light on the matter and reveals a dog’s breed is actually a poor predictor of behavior. In fact, the types of breed only explains about 9 percent of the behavioral differences between dogs, according to a recent study in the journal Science.

“Everybody was assuming that breed was predictive of behavior in dogs,” said geneticist Elinor Karlsson of the University of Massachusetts Chan Medical School in Worcester in a news briefing. But “that had never really been asked particularly well.”

One study in 2019 considered whether genetics might explain collective variations between breeds and found that genes could explain some of the differences, but Karlsson and her colleagues wanted to learn how much breed types could predict the variation in the behaviors of individual dogs.

To study behaviors at the individual level, Karlsson and her research team needed to collect genetic and behavior data from a lot of dogs — so, they developed Darwin’s Ark, an open-source database to which more than 18,000 pet owners responded to surveys about their dogs’ traits and behavior.

The dog owners involved in the study were asked over 100 questions about observable behaviors, which the researchers grouped into eight “behavioral factors,” including human sociability — how comfortable a dog is around humans — and biddability — how responsive it is to commands.

The researchers also collected genetic data from 2,155 purebred and mixed-breed dogs, including 1,715 dogs whose owners answered the Darwin’s Ark questions and also sent in saliva swabs from their pooches. The inclusion of mixed-breed dogs, or, mutts, helped the team better analyze how ancestry affects behavior while removing the purebred stereotypes that could affect the way the dog is treated — and thus behaves.

Studying mutts also made it easier to decouple traits from one another, explained Kathleen Morrill, a geneticist in Karlsson’s lab. “And that means on an individual basis, you’re going to have a better shot at mapping a gene that is actually tied to the question you’re asking.”

Then the team combined the genetic and survey data for the individual dogs to identify genes associated with particular traits and found that the most transmissible behavioral factor for dogs is human sociability, and that motor patterns, like howling and retrieving, are generally more transmissible than other behaviors.

Kathryn Lord, an evolutionary canine biologist in Karlsson’s lab, said during a press briefing on the study that before modern breeding started, only in the last couple of hundred years or so, dogs were selected for the functional roles they could provide, such as hunting or herding. Today, such characteristics still show up in breed groups. Herding dogs, for example, tend to be more biddable and interested in toys. Then, within breed groups, individual breeds are more likely to have certain motor patterns: hence, Retrievers are more likely to retrieve.

Yet, the team discovered even though breed types are associated with certain behaviors, they are not reliable predictor of individual behavior. Which means, wile retrievers are less likely to howl, some owners said their retrievers howled often; greyhounds rarely bury toys, although some do.

The study confirms what people have long observed, that breeds differ on average in behavior, but there’s a lot of variation within breeds, says Adam Boyko, a canine geneticist at Cornell University, told Science News.

The research team was surprised to find a dog’s size had even less of an effect, virtually none, as it turns out, on an individual’s behavior. Boyko points out that small dogs — yappy small dogs — may often behave worse than large dogs, but instead of all that yappiness being built into their genetics, “I think it’s that we typically tolerate poor behavior more in small dogs than we do in big dogs.”

Curtis Kelley, a dog trainer at Pet Parent Allies in Philadelphia, told Science News breed provides a loose guideline for what kind of behaviors to expect, “but it’s certainly not a hard-and-fast rule.

“A puppy will show you who they are at eight weeks old…It’s just our job to believe them,” he said.

“Dogs,” Kelley added, “are as individual as people are.”