Dr. Karen Overall: ‘Bad tests are killing good dogs’

April 8, 2014

bigstock_Proud_Labrador_Retriever_2914606Are behavior assessments in shelters getting it right? One expert says no.

At the 2013 NAVC Veterinary Conference, Karen L. Overall, MA, VMD, PhD, DACVB, CAAB, editor of the Journal of Veterinary Behavior, spoke about the science of behavior assesments in shelters as part of the event’s shelter medicine track.

“All my dogs are rescue dogs,” she told the audience, “and I am absolutely committed to those left behind.”

Dr. Overall said established methodologies of assessing behavior, such as those used in studying the mating displays of birds or to assess animal behavior among wildlife, are seldom applied to the assessment of dogs and cats in shelters.

She said ideally, assessments should have the qualities of:

  • Standardization
  • Reliability
  • Validity (in the scientific sense of having gone through the validation process)

In the shelter assessment efforts she’s seen, she said, “Data is being gathered, but it’s not useful data…. There are standards; we should be using them!”

Her advice: “Before you construct an evaluation, talk to a statistician. It will cost around $300 to have a statistician help you design a meaningful evaluation, better than a bunch of iPads that are collecting a bunch of crap that doesn’t mean anything…. What’s out there is not sufficient for the need at this time.”

Dr. Overall recommends taking video of an assessment from two different perspectives, and not to rely on real-time examination and observation of body language and position.

She expressed concern with the “Assess-a-Hand” test, saying:

If there were one thing I could eliminate from shelter assessments, it would be the “Assess-a-Hand,” which is responsible for thousands if not millions of animal deaths over the years.”

The concern here is that if you are looking for dogs and cats to euthanize, this test fits the bottom line. And that’s what the Assess-a-Hand does.

Another problem Dr. Overall discussed: Testing too close together, when the dogs haven’t had a chance to calm down yet. This will confound results, she said. “Many people doing these tests are teaching dogs to be reactive. They’re making them worse.”

Dr. Overall said she frequently hears objections that shelters lack the resources to validate their assessment tests, but she pointed out that researchers in one study “did this test with two handlers, two ropes – because they couldn’t even afford regular leashes – two open spaces, and one enclosed space. Don’t tell me you don’t have the resources to do better than you’re doing.”

“Shelter animals should be sheltered,” she concluded, saying that the use of unvalidated tests will result in the deaths of dogs who could be safely adopted.

Filed under: behavior and training,no-kill — Christie Keith @ 5:33 am


  1. Here’s my favorite video on Assess-a-Hand.

    I’m late to the party but I’m so glad Pet Connection’s back!

    Comment by rheather — April 8, 2014 @ 6:48 am

  2. ^^ Ladies and gentleman, my friend and newest dog-in-law, Linda!

    One thing I don’t see here is an emphasis on actually qualifying the personnel who do the assessment.

    A weekend class in “reading dog body language” does NOT an expert make.

    I was once at a SAR conference where a very good, sound instructor gave a solid one-day class on dog language.

    The next day the students, all of them SAR handlers and some who fancied themselves trainers, dispersed through the other classes.

    My teammate was there with her enormous, gentle, mixed-breed Clifford The Big Red Dog.

    Some of the handlers were apparently a bit intimidated by his size, which was the only thing remotely intimidating about him.

    All day they kept helpfully interpreting everything he did and did not do for his handler. And it was all dominance, except when it was “stress.” They parroted back things the instructor had said but as complete gibberish. Each had fixed on one or two shiny buzzphrases from the class and they were merrily applying it to *everything.* New toy.

    As my teammate said, “Everyone here is so helpful and wants to share what they know, even when they’re wrong.”

    I am absolutely certain that the same thing happens when shelter staff and volunteers get the same training.

    But they are making decisions about dogs that can be life or death, for the dogs, and even potentially for human beings.

    There is no fast way to fix this.

    There is no substitute for years of genuine experience working with dogs, specifically training dogs and addressing behavior issues.

    I have seen and been part of the exodus of experienced results-based trainers (both pros and serious, dedicated, smart hobbyists) and breeders from working with animal shelters. We’ve been pushed out and sometimes outright dismissed. Our methods are not PC, or the fact that we breed animals is moral anathema. We are inadequately submissive to edicts from bureaucrats.

    Shockingly, the 20-year-old goggie wuvvers and bored “I’ve owned dogs all my life” retirees who ARE welcomed are less clear-sighted and effective.

    But they quickly learn the correct shiny buzzphrases and parrot them back. And they don’t question that practices and equipment that are, in fact, highly distressing to the animals, are the most excellent positive and humane practices and equipment.

    I’ve found a rescue run by sensible humans for which I volunteer. But I am still sad for the local shelter dogs who are on their own.

    Comment by H. Houlahan — April 8, 2014 @ 12:01 pm

  3. I am a statistician & methodologist, a social & behavioral scientist by profession. I also am founder of an animal wefare organization. I would be HAPPY to provide pro bono consultation and advice on study design for any such efforts to truly assess and evaluate such behavioral assessments. In my own field, we routinely have to develop coding schemes for observing and evaluating human behavior. And we have to be as concerned with things like inter-rater reliability and the problem of “drift” incoming and observing behavior. Then add the issue of predictive validity with respect to outcomes of interest.

    Comment by Carol Tutzauer — April 8, 2014 @ 6:42 pm

  4. Dr Overall is a conundrum for me.

    I have found myself agreeing with her stances on such topics as this as well as her position on sterilization via gonadectomy, if her publications in some recent veterinary magazines are any indication. It’s nice to have someone with weight in the dog community coming out against standard practice.

    I still cannot get past her vitriol for certain training tools and techniques which many hold near and dear, and if this is my one sticking point with her, then the enemy of my enemy is still not, and will never be, my friend.

    Comment by Viatecio — April 9, 2014 @ 3:20 am

  5. I ?? Linda.

    Re Viatecio’s comments above, I don’t know a lot about Dr. Overall, except to say that she’s right on this one issue. The paranoid freakout by Sue Sternberg — whose apparent pathological fear of actual dogs is legendary — is being used as an excuse by other “shelters” to execute otherwise healthy animals. I don’t know how you can justify that.

    Comment by Rob McMillin — April 9, 2014 @ 12:13 pm

  6. BTW, that first line was supposed to read, “I ♥ Linda”, but apparently your comment submission form doesn’t like UTF-8.

    Comment by Rob McMillin — April 9, 2014 @ 12:15 pm

  7. How can I – or why can I not – share articles from PetConnection on Facebook?

    Comment by Giselle — April 9, 2014 @ 4:11 pm

  8. Great read, common sense.

    Comment by Doberman Rescue Ontario — April 9, 2014 @ 5:22 pm

  9. Giselle, you should be able to see social media options at the top of each post. If you’re not using the latest version of some browsers they may not show (I know from experience!).

    Comment by Gina Spadafori — April 10, 2014 @ 9:30 am

  10. When I did assessment or a follow up for food aggression I never used an assess a hand. I had far fewer fails with my own hand and never was bitten. I reversed other people thoughts on dogs being food aggressive by administering the test with my bare hand. That being said it still had little validity. Just fewer dead dogs.
    I would hope that more research and science would yeild a better testing process. I’m will to show anybody who wants to learn the bare handed process.

    Comment by Jamie Benton — April 10, 2014 @ 12:12 pm

  11. As someone who relies professionally, as an animal artist interested in depicting behavior (not just critters standing around), on my ability to accurately observe wildlife and do what is necessary to understand what I saw, I find it so refreshing that someone else *gets* that what wildlife biologists do applies quite relevantly to dogs and cats, who are also animals (a point that seems to be lost on too many people).

    Comment by Susan Fox — April 10, 2014 @ 8:27 pm

  12. I agree with some of Dr. Overall’s views and I disagree with others. I feel the same about Sue Sternberg. But let me tell you people who think every dog should be saved, it would be nice in a perfect world. Have you seen families torn apart over an aggressive dog that was obtained as a puppy to be a nice family pet? A dog that was raised well but now has bit someone or ate the family cat? Have you seen the overcrowding and bad conditions so many shelters are in? They do their best but do not have the space or money to properly care for the dogs in their care. Have you owned a dog with very serious behavior issues? A dog that that cannot go near but a few people and is a constant worry to keep happy and safe? I would love to see shelters that are empty, no dog euthanized because of behavior or space but we are far from that so those that love to be cruel to other people yet profess to love dogs so much put some effort into educating yourself. Visit the overcrowded, underfunded shelters in the south, visit a family that bought/adopted a puppy with genetic issues and now have a dangerous dog, traumatized people and maybe even a lawsuit. We may need better education and methods in the shelters but many shelters need more money first. They cannot keep a dog no matter the temperament if they do not have a tiny space to put it nor food to feed it. So do they euthanize dogs that may have future issues. Could the dog be worked with? Possibly by the right person. Does the average family looking for a dog want that-no. Should they have to want to adopt a puppy or dog that has issues the may or may not end up being serious-no they should not. You really don’t know what problems may arise when a dog is in a home for a while and I would hope everyone who gets a dog expects to put effort into training and working through some issues of course, no one human or animal is perfect but there is a line and sorry but I am on the side of doing my best to keep families safe and happy and let me tell you I have sacrificed a lot for my problem dog. She has serious genetic issues that showed up at a young age. She looked a lot like some of the puppies that Sue Sternberg would recommend euthanizing. Maybe not as bad as some of them. I decided to keep her, I would not adopt her out and thank God I did. Honestly though I love her to pieces I often question my decision as my quality of life has suffered trying to keep hers good. I have also learned (live in small town) that a lot of her relatives have bit, killed a dog and so on. I took my dog as a rescue because she was also deaf. These are Labs that are coming from a breeder being sold to families that want a loving family pet. So keep criticizing Sue Sternberg for doing her best to try and prevent nightmares like this from happening. I applaud her courage. I wish I could be so strong. Instead I lie with s lot of stress trying to save the world.

    Comment by Cheryl — April 11, 2014 @ 6:34 am

  13. Cheryl, excellent points, and I agree. I come from a dog training & shelter background, and I don’t think there is any one answer for assessing pets. I have endured heartbreak over euthanizing perfectly good, sound pets for lack of space and resources, and worked with tiny rescues who meant well and wanted to “save them all” , but placed dogs in inappropriate homes and overwhelmed new families..

    I am often amazed at the lack of support and understanding that animal shelter/rescue organizations have for each other. Sometimes it is downright nastiness. The issue is complex with a wide range of variables and really what we, as “humane” focused people, should be doing is learning from each other, and offering support, encouragement, and the sharing of ideas.

    I don’t agree with all of Sue Sternberg’s ideas, but she IS placing safe dogs in “average” homes, and is motivated to develop some protocols to assist other shelters in doing the same.

    In my experience, some of the no-kill shelters are doing far worse for homeless animals then Sternberg ever has through hoarding, inappropriate placements, selective enrollment, and shaming of shelters serving communities. No-Kill is an ideal, but for many communities it is currently unobtainable.

    The shelter/rescue world is as volatile as religion or politics.

    Comment by Christi — April 11, 2014 @ 11:08 am

  14. Excellent response by Cheryl. The majority of adopters are families looking for safe, reliable family pets. They are not looking for behavior problems, especially potential aggression issues. Knowing whether a dog may have resource guarding issues is important information to have when making adoption decisions. Does it mean every dog that displays some signs of resource guarding should be euthanized, no. Does it, though, influence what type of home the dog is best suited for, absolutely! Not all animal shelters have the resources, both financially and personnel, to house animals long term and provide them with behavior modification. As Sue Sternberg has always said, through shelter evaluations, we are trying to determine how a dog is going to behave in a home, while living in an animal shelter. There are no crystal balls. I’m more concerned about the shelters that have no evaluation process in place at all. The Assess-a-Hand is only one tool in the box to try to determine a dog’s overall temperament and behavior. What is done with the information obtained is the decision of each individual organization based on their resources, policies, adopter support, and clientele.

    Comment by Jennifer — April 11, 2014 @ 2:04 pm

  15. Your awsome !! Totally agree !!

    Comment by Deb Lucas — April 11, 2014 @ 4:19 pm

  16. Great love it and her !! Met her 2 yrs ago for my bella

    Comment by Deb Lucas — April 11, 2014 @ 4:20 pm

  17. I have had purebred dogs for 29 years and did rescue for 10+ years as well as working at and/or managing boarding kennels for 17 years.

    Most shelter and rescue evals are based on Sue “kill ’em all” Sternberg’s. The same woman who told us here in New England that “There are no adoptable dogs in shelters in New England” and that we needed to get all of our “nice well adjusted dogs” from the South

    Most dogs who have never been kenneled before do very poorly in kennels/shelters. They do not show their normal behaviors or temperament there. A dog that can be a fantastic family dog that adores kids and tolerates dogs may fail flat out fail terribly a shelter/rescue eval because they are very soft, sensitive and overwhelmed.

    Shelter and rescue evals done at shelters will rarely be 100% accurate. Dogs who pass these evals will be bland generic dogs (Sue Sternberg’s preference) with no allowance made for individual breed traits. Breed rescues doing evals at shelters SHOULD have an eval tailored for their breed, independent breeds test different than velcro breeds. Dog aggressive breeds test different than breeds bred to work in groups, hunting breeds cat test different than non hunting breeds and so forth- and this is why I made a Shiba Inu (a breed I had for 10 years) Eval, they test differently while still being “normal and safe”.

    My breed for 29 years is the Rottweiler. NO Rottweiler with a real true Rottweiler temperament will ever pass a shelter eval if done to Sue Sternberg’s standards and in Rott rescue we have come to grips with the fact that many lovely obedience, and strong working prospects WILL be killed because of evals. There is nothing to be done to change this because the average pet home is not suitable for a “real Rottweiler” and if you have a breed with a strong temperament inherit you do need to acknowledge that only the most generic “happy family dogs” will pass a shelter eval and either accept that OR have experienced foster homes available to house the more breed correct dogs until and IF a suitable adoptor comes along

    Am I advocating adopting or rescue pulling aggressive or dangerous dogs? Am I advocating placing dogs with a bite history? NO never ever.
    BUT sometimes a rescue needs to look at such a dog in a “No kill” shelter and do a “Breed management” or “Public safety” pull. This is a dog that you know will bite someone someday and be a headline in a paper OR will sit at the shelter til it dies of old age because it is unadoptable. You pull those dogs and you have them euthanised that same day. Sad? yes. necessary? sometimes yes

    True love for dogs means you evaluate them CORRECTLY and fairly with breed allowances and kennel fear allowances. You foster them out for a MIN of 2 weeks to truely assess them properly, you adopt them out to screened homes that suit them and yes you kill the ones who will be a public danger- NOT an Assess a hand danger

    Comment by Diane Richardson — April 12, 2014 @ 6:33 am

  18. Having had personal dealings in a shelter environment..I can 100% agree with this article. Changes need to happen…

    Comment by Alberta Doberman Rescue — April 13, 2014 @ 12:37 am

  19. I assessed a dog that was saved after Hurricane Katrina from Louisianna. All I can say is Thank God for Assess-A-Hand.
    Not only did he bite the hand in one second he tore it out of my hand a shook it feriously. When he finished his food he acted like the loving dog he was before.

    Comment by Anne — April 16, 2014 @ 3:51 pm

  20. “I don’t agree with all of Sue Sternberg’s ideas, but she IS placing safe dogs in “average” homes, and is motivated to develop some protocols to assist other shelters in doing the same.”

    I call BS. Show me the numbers. The SFSPCA followed up on dogs that “passed” a Sternberg protocol *administered by Sternberg* and those that failed. Guess what they found?

    The “kill anything with a pulse” approach does not just pack the dumpsters with perfectly nice animals, it “passes” animals that will later go on to bite, maybe because the test protocol or tester missed something, or maybe — get this — because of the horrifically false confidence the test engenders, the complacency that follows about appropriate placement and followup. Because every animal with a mouth has the potential to bite, and it the humans around it are careless enough, ignorant enough, or stupid enough, it will.

    There are no idiot-proof dogs, so it is *always* beholden on a shelter, rescue, or breeder to follow best practices in weeding out or converting idiots.

    Comment by H. Houlahan — April 21, 2014 @ 9:24 am

  21. The test is only as good as the tester. Many who don’t read dogs are assessing dogs. If we continue with ‘save them all’, dogs will be less and less welcome in society as more and more ER visits happen to kids. People are getting dumber about animals in our world, so placing risky animals with naive people is counter productive. Personal attacks on individuals who are rolling up their sleeves are not helpful to the animals, even if it feels good.

    Comment by Tricia — May 4, 2014 @ 6:10 am

RSS feed for comments on this post.

Leave a comment


Recent Posts

Recent Comments


website design by Black Dog Studios