November 11, 2016
double tap picture to expand gallery
In a small room in a Pennsylvania shelter, Dixie, a tricolor hound, wags her tail nervously as she embarks on what might be the most important 20 minutes of her life.
While one shelter worker videotapes the session, another leads Dixie through a series of tests, evaluating her willingness to make eye contact, to be touched all over her body, to have her paws slightly squeezed. The staffer brings out some squeaky toys to judge how Dixie acts when excited. But Dixie isn’t feeling playful. She seems eager to please, but also anxious and confused. When she’s introduced to a barking, jumping dog, Dixie sidles away, refusing to engage with the unruly visitor.
She shows more enthusiasm after a bowl of food appears. When a fake hand on the end of a pole attempts to slide the bowl away, Dixie just eats faster. But later, when the staffer tries to tug a rawhide chew from her mouth, the hound resists. There’s no growling or raising of hackles, but she clearly doesn’t want to surrender her treat.
For every reaction, Dixie is rated on a 1-5 scale, with 1 signifying the lowest potential for aggression. She scores a 1 on everything except food aggression. A note at the end of the video explains that because she was rated a 3 on the food behavior assessment, Dixie will be listed as best placed in a household without children under 10.
It seems reasonable, even laudable, that a shelter performs this type of due diligence before placing dogs into homes. At the same time, for dogs like Dixie in a high-intake, open-admission shelter, a restriction like this could easily mean the difference between adoption and euthanasia.
And that’s a problem, say Gary Patronek and Janis Bradley in a new study, “No Better Than Flipping a Coin: Reconsidering Canine Behavior Evaluations in Animal Shelters,” that will be published in the Journal of Veterinary Behavior.
A veterinarian with Tufts University’s Center for Animals in Public Policy, Patronek once worked for the Animal Rescue League of Boston, where he helped develop, implement and validate behavior exams. Bradley, a veteran dog trainer with the National Canine Research Council, once administered such tests. They now make a compelling argument that standardized behavior evaluations are simply unreliable and don’t serve their intended purpose—to make better adoption matches and ensure public safety.
Behavioral test models were developed with the best intentions. They “were seen as a way to do a better job, to remove some subjectivity and to have objective criteria” when evaluating a dog for adoption, Patronek says. At the same time, they were thought to relieve emotional stress on staff by transferring the burden of life-and-death decisions to the dog’s performance against a formal checklist. “It was seen as a more scientific way to make those decisions,” he explains.
Yet dog behavior is far from an exact science, and while an estimated 25 percent of shelters today use formal behavior exams, none of the test models has been scientifically validated, says Patronek.
Kenny Lamberti, acting vice president of companion animals for The HSUS, remembers how in the 1990s, formal canine behavior assessments were viewed as a “gold standard” practice in shelters. But over time, people began to have doubts, especially when test results determined euthanasia decisions. Lamberti, who worked as a dog trainer for 20-plus years, realized that many of his clients’ dogs would flunk a test if they were in a shelter, and yet they were living happily and safely in homes. “If I was to do any evaluation on my dog in my house, where he’s comfortable, he’s going to respond really differently than if I did it in shopping mall plaza,” Lamberti says. “Stress has a profound impact on results.”
Others in the field have recognized the difficulty of evaluating dogs in a high-stress environment but believed the solution lay in finding ways to improve the tests’ accuracy, says Patronek. “We never stepped back and asked a fundamental question: Was this ever feasible in the first place?”
In their study, Patronek and Bradley applied statistical models used to calculate the accuracy of medical diagnostic testing and behavioral risk assessment in humans. Based on the results, they conclude that devising a test that will reliably predict how dogs will act in future adoptive homes is “vanishingly unlikely.”
The prevalence of serious aggression in dogs is low, Patronek points out, and most dogs do fine in their homes. “The public probably makes better choices than we give them credit for, and dogs are better than we give them credit for.” At the same time, he feels shelters should accept the limits of their powers. “Where was it ever said that shelters … could guarantee something when we’re dealing with a living animal?”
This doesn’t mean shelters should adopt out any dog regardless of his temperament, Patronek says. But most dogs with a history of serious aggression, or who can’t be safely handled by staff, are typically weeded out of the adoption pool before behavior tests are administered. Applying tests designed to provoke a threatening response in the dogs who haven’t displayed problem behaviors is at best a waste of time and resources, Patronek says, and at worst a tool that leads to needless deaths.
Instead, Patronek and Bradley recommend that shelter staff evaluate dogs by interacting with them in normal ways—such as walks, doggie playgroups, games and training. “These activities are likelier to identify any additional dogs whose behavior may be of concern, will enrich dogs’ lives and minimize the adverse impact of being relinquished and confined to a shelter,” they write.
Evaluating dogs’ suitability for adoption “is still going to involve some tough decisions, but you can’t rely on unvalidated science,” Patronek says. “We need to move common sense and dog sense back into that process.”