No Defense: Shelters are Trying to Salvage Temp Tests

There is not a single temperament test used to evaluate shelter dogs that is reliable in predicting behavior. Over the last several years, a wide body of research has shown this again and again with one study concluding, based on over 25 years of data, that there is “no evidence that any canine behavior evaluation has come close to meeting accepted standards for reliability and validity.”

While shocking, it should not be surprising since these tests are built on a “fatally flawed” premise: “that the provocations used at a single time during a dog’s stressful experience in a shelter will predict future behavior at a different time and place.”

Despite those significant limitations, some shelters in Australia are trying to salvage those tests. A new paper in an upcoming issue of the Journal of Veterinary Behavior — written by individuals associated with the shelters still championing temperament testing — argues that they should still be done, so long as they are not used to “pass/fail” dogs and the results are considered as part of broader observations. Their claim is unconvincing for several reasons.

First, most of their reasons for calling for continued use of temperament testing have nothing to do with whether the tests actually work. For example, the authors argue that temperament testing reduces “the possibility of litigation” and being held liable if a dog bites. Providing a false seal of approval that is also unnecessary, shelters have other tools to defend against liability without unreliable tests that cost dogs their lives. In New York, for example, state law prohibits an SPCA or humane society that holds an animal control contract from being held civilly liable for adopting a dog. Moreover, given the unreliability of the tests, adopting out dogs after they have been evaluated may actually increase liability because of the implicit certification that the dog is safe.

The authors also argue it lessens the guilt of staff who order dogs to be killed. They state that these evaluations, flawed as they may be, “takes the difficult decision of whether a particular animal should live or be euthanised [killed] away from: staff,” saving them the emotional “stress and strain” of condemning an animal to death. It is a blatant rejection of ethics to propose that flawed, unreliable tests that result in dogs being wrongly killed should be used to lessen feelings of guilt.

Moreover, there is no evidence that it won’t have the opposite effect, as truly caring staff struggle with condemning dogs to death based, in part, on the results of temperament testing that are unreliable and lack predictability. The authors, however, argue that predictive validity can be improved if “staff are adequately trained” and acknowledge “that existing tests are imperfect and require continuous research and development.” But the tests aren’t just “imperfect.” They don’t have a small error rate as the word “imperfect” suggests. They are, to quote the study that looked at 25 years of data, “no better than a coin toss.” In fact, some of the tests are wrong as much as 84% of the time.

In addition, a staff member who is encouraged to kill a dog under such circumstances is being encouraged to devalue dogs, is failing to understand the behavior in the context of stressful, unfamiliar housing combined with recent trauma (including separation from their families), and is not treating each dog as an individual who calls upon our conscience, our resources, and our creativity to solve problems, rather than eliminate them through killing. As such, training staff on fundamentally flawed and unreliable tests take them further from, not closer to, a real understanding of dog behavior. And staff “stress and strain” are more likely reduced through success at rehabilitating traumatized dogs, rather than relying on gimmicks that result in failure.

Finally, the authors argue that the tests will provide more information about the dog to help provide better placements. But this information can be gathered in other ways: staff observations, volunteer observations, historical narratives from prior homes, and more, without putting dogs at undue risk.

The study authors admit to conflicts of interest: the lead author’s studies are funded by a shelter and one of his co-authors works for one.

The paper, “In Defense of Canine Behavioral Assessments in Shelters: Outlining Their Positive Applications,” is here.


Have a comment? Join the discussion by clicking here.