Articles Articles (Essential)

Behaviorally Challenged Dogs: Can We Save Them All? [1]


Photo by Eileen McFall: dog playgroup at Austin Animal Center.

True to their mission statements and the highest ideals of the animal protection movement, shelters committed to a No Kill orientation seek to provide a brighter future and a second chance to every animal entrusted to their care. Yet while eliminating the killing of 100% of companion animals who enter shelters is a worthy goal, the reality is that shelters invariably take in a small number of animals who are irremediably physically suffering. For these animals, such as those who have been hit by cars and sustained life-threatening injuries currently beyond the reach of veterinary medicine or are in multiple organ system failure and are therefore facing imminent death, killing represents true mercy and meets the dictionary definition of “euthanasia.”[2] While statistics indicate that this group generally represents less than 1% of animals entering the typical American animal shelter, this means that up to 4% of animals still being killed at those shelters reporting at least a 95% live release rate do not suffer from conditions that justify true euthanasia. Who are these animals and why do they still face death, even at those shelters hailed as having successfully implemented the No Kill Equation?

While innovations within the field of animal sheltering have resulted in life affirming alternatives to killing for almost  all categories of animals entering shelters—alternatives such as sterilization for community dogs and cats, palliative and hospice care for non-rehabilitatable animals who are not mortally suffering, as well as foster and onsite supplemental care for orphaned neonatals, among others  [3]—there is still one group  of physically healthy animals for whom no reliable safety net exists. These animals are being killed not for reasons of mercy, but human convenience: dogs with “behavior” challenges, including perceived aggression.

Continued innovation into how these animals are diagnosed and rehabilitated in a shelter setting is essential to ensuring their fair treatment and protecting their lives, as well as to ensuring the purest expression and ultimate achievement of the philosophical orientation underlying the No Kill movement. As such, overcoming the challenges these companions animals present remains one of the remaining, and most pressing, frontiers of the No Kill movement.  [4]

This position paper is intended to provide an overview as to where that effort now stands: how many dogs fit this category, how many are currently being killed, what efforts have proven successful at rehabilitation, what experts in behavior rehabilitation believe is possible, how we can overcome the limitations of the shelter setting to better serve these animals, and the proper philosophical orientation around their killing that will ensure our continued innovation on their behalf. It is intended to supplement several No Kill Advocacy Center guides, including “Defining No Kill,” [5]  “What Shelters Owe Traumatized Animals,” [6] the “Matrix,” [7] and several upcoming guides, including SOPs for behavior dogs. As such, it represents the continuation of an ongoing discussion about how shelters protect all the lives still at risk and, more importantly, provides a path to ending the killing of all dogs, with the exception of those who are irremediably physically suffering.

When it comes to saving dogs with “behavior” issues in shelters, including perceived aggression, how many—what percentage—can we place? [8] Is it 90% of shelter dogs as some advocate? Is it 95% as others do? Or is it higher? In fact, based on four recent studies, the pioneering work of behaviorists, and the results of some of the most successful and progressive shelters in the country, a live release rate of even 99% for these dogs is not high enough. Thankfully, the path to ending their killing altogether is now clear.

Dogs are being falsely labeled as “aggressive” and killed

As an increasing number of shelters across the nation achieve unprecedented success by focusing reform efforts not on the public, but on their own practices, two important lessons are to be drawn. First, that historical narratives and assumptions used to explain the “necessity” of shelter killing have, more often than not, been wrong. Second, that success requires us to challenge conventional wisdom, including the role those historical misperceptions as well as inadequate sheltering practices have played in fostering deadly, but ultimately avoidable, outcomes. Poor cleaning and handling protocols which can result in life-threatening illness, inadequate socialization and exercise regimes which result in stressed animals prone to both disease and anti-social behavior, and the impoundment of community animals who are not social with humans and, therefore, whose full range of needs can never be met in a captive environment, are just a few of the failings of traditional sheltering which the No Kill philosophy has challenged and is working to successfully overcome. Innovation to eliminate the killing of dogs with behavioral challenges demands that we once again summon the same courage and conviction which is to credit with rates of shelter killing that are at an all-time low and with an unprecedented number of communities throughout the nation reporting 95% or better live release rates.  [9]

To that end, it is incumbent on shelter administrators, shelter reform advocates, and public officials to acknowledge and respond to what the latest scholarship indicates are the many failings not only of the means by which the temperament of dogs in shelters is currently evaluated, but just as important, the inadequacies of the setting in which these tests are conducted: the shelters themselves.

As studies are now increasingly demonstrating, the very nature of shelters themselves and the policies of those shelters undermine the mental health of dogs and set them up to fail. At the same time, the temperament tests used to evaluate those dogs are, at best, deeply flawed and thoroughly lack predictive validity. At worst, shelters look for reasons to fail dogs. In short, if a dog is showing behavior issues, including perceived aggression in the shelter, it’s rarely the dog; it’s almost always the shelter.

Dogs are being labeled as “aggressive” (and, as a result, killed) by shelter staff employing methods that fail to take into account several critical factors; namely, that dogs under duress in shelters often act in ways that are dissimilar to their behavior out of one; that the shelter environment itself—loud, stressful, with inappropriate housing, strange smells, and lacking adequate socialization—can cause the behavior or prevent full rehabilitation; that dogs have experienced a recent trauma (including separation from their families) which can temporarily impact their behavior; that there may be a medical origin for the perceived aggression that is not taken into account; and that there are other possible solutions and alternative placements in lieu of killing.

A recent study, for example, found that dogs who are kept physically isolated from other dogs as they are in many shelters are stressed and stressed dogs react negatively, especially when they meet other dogs. [10]  Accordingly, “dogs housed alone and unable to come into contact with conspecifics [other dogs] developed a higher percentage of behavior problems.” They are often deemed to fail tests for “aggression” or they are labeled “dog aggressive” and killed. By contrast, dogs who are allowed to play with other dogs in the shelter via playgroups—or are cohoused with other dogs—are less likely to fail those tests.

Compounding these concerns is the flawed nature of the tests themselves. A recent study published in the Journal of Veterinary Behavior concluded that temperament evaluations in shelters are no better than a coin toss. [11] In fact, a coin toss may be better: “a positive test would at best be not much better than flipping a coin, and often be much worse, because many of the dogs who test positive will be false positives.” In other words, dogs are being killed as “unadoptable” or “aggressive” based on faulty testing.

Moreover, shelters often look for reasons to fail dogs, which would justify their killing. Accordingly, the authors suggest that,

[I]nstead of striving to bring out the worst in dogs in the stressful and transitional environment of a shelter and devoting scarce resources to inherently flawed formal evaluations that do not increase public safety, it may be far better for dogs, shelters, and communities if that effort was spent maximizing opportunities to interact with dogs in normal and enjoyable ways (e.g., walking, socializing with people, playgroups with other dogs, games, 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, be more indicative of the typical personality and behavior of dogs, and may help make dogs better candidates for adoption.

While both these studies received a lot of notice for these findings, there’s an even more powerful conclusion to be drawn which has received less focus. Both should put to rest, once and for all, the false notion that dogs in shelters are in shelters because there is something wrong with them: “Nothing in the prevalence estimates we reviewed suggest that overall, dogs who come to spend time in a shelter (and are not screened out based on history or behavior at intake or shortly thereafter) are dramatically more or less inclined toward problematic warning or biting behavior than are pet dogs in general.” In short, dogs in shelters are not “damaged.”

But if shelter dogs are not more inclined towards biting than pet dogs, what are the implications for the No Kill movement? Given that far less than 1% of pet dogs bite people, the conclusion is inescapable: shelters are killing dogs falsely labeled “aggressive.” Even those shelters which have eliminated the killing of 90% of dogs and which claim to be “No Kill” are killing dogs who have nothing wrong with them. In fact, looking at bite rates that require hospitalization, only 0.01% of dogs (or roughly 1 in 10,000) bite with enough force to cause an injury.  [12]

These studies mirror the findings of the most progressive and successful municipal shelters (and those running “open admission” shelters under contract) in the country. Doug Rae, the director of the Humane Society of Fremont County in Colorado, is instructive. The Fremont Humane Society runs the animal control shelter for seven cities under contract. Rae also previously ran a private shelter in Rhode Island and “open admission”/municipal shelters in Arizona, Indiana, Maryland, and Pennsylvania, including those taking in as many as 30,000 animals per year. Says Rae:

Over the years, I have rarely seen a truly ‘aggressive’ dog. The vast majority are simply scared. I was speaking with a Board member for a shelter on the West Coast : and she said that her shelter saves 96-97% of the animals and that 3 or 4% are dogs who are aggressive and need to be killed. In my experience, the percentage of truly aggressive dogs I have seen in small to very large shelters is well under one quarter of 1%.

Likewise, Austin, Texas, shelter leadership is working to tighten and improve behavior protocols and build more bridges with placement partners even though it has achieved live release rates as high as 98.8%—which includes dogs killed for medical reasons—acknowledging that even such a high success rate includes the killing of “behavior” dogs who are not mortally suffering and would thus not be irremediable.

The path to save them all  [13]

Given that temperament testing of dogs has little predictive ability, that shelters are stressful environments, and that shelter policies undermine the mental welfare of dogs, how can shelter administrators more accurately assess the need for behavior rehabilitation in dogs? They do so by eliminating extraneous and potentially deadly variables that skew results.

If a dog is exhibiting behavior issues in the shelter, getting them out of the shelter and into rehabilitative foster homes invariably resolves those issues for the vast majority of dogs. Where the shelter is not solely to blame, getting the dog out of the shelter allows for rehabilitation efforts to be more focused on the persistent behavior issue as opposed to those which are also shelter-induced. It also allows rehabilitation efforts to be more effective, and more effective much more quickly. Finally, for those dogs who may truly be aggressive to humans and whose prognosis for rehabilitation, at least at this time in history, is poor to grave, long-term protective placement is a viable option given the small number of dogs to which they would apply.

Managers at the Fairfax County Animal Shelter in Virginia sought to save the lives of more dogs by putting in place a pilot program to determine if their temperament testing had predictive ability, if the shelter environment itself was contributing to perceived behavior problems including aggression, and if time spent out of the shelter would allow for rehabilitation of dogs in ways that might be difficult in a municipal shelter environment. [14] Prior to the program, dogs who failed the shelter’s temperament evaluation were killed (many who were not evaluated were also killed). One of the most common reasons for killing these dogs was perceived “aggression towards humans or animals” arising from testing results.

As part of the program, medium to large dogs who failed the evaluation were instead sent into foster care for further evaluation and training. The end result: 90.4% of these dogs were successfully adopted. More importantly, none of the placed dogs ended up needing any rehabilitation; simply getting out of the shelter resolved any problematic behavior staff saw in the shelter.

For dogs who actually do require rehabilitation, a review of the literature on resilience in humans and applied to dogs through the pioneering work of the No Kill Advocacy Center, bioethicist Jessica Pierce, and Dr. Karen Overall, the Editor-in-Chief of the Journal of Veterinary Behavior: Clinical Applications and Research, [15] demonstrates that dogs are incredibly resilient, that there is no such thing as “irremediable psychological suffering” in dogs, and that all dogs with behavior trauma can be appropriately rehabilitated and/or placed through criteria that depends on the severity of the trauma.

In addition, successfully rehabilitating dogs suffering from even severe behavior problems doesn’t take as long as conventional wisdom would suggest. While none of the dogs in the Fairfax program could be classified as “severe,” all of them had behavioral challenges while in the shelter. This included dogs with, among other things, barrier reactivity, fear-based aggression, resource guarding, kennel stress, prey drive, and bite history. Some of the dogs also had secondary issues including extremely high energy, possible dog aggression, dog selectivity, fear of men, undersocialization, separation anxiety, and reactivity. Despite this, 40% of the dogs were in foster care for up to one week, 48% were in foster care between eight days and one month, and the remainder were in foster between one and eight months. In other words, 88% of the dogs were in foster care 30 days or less before shelter staff determined they were ready for adoption.

Even in severe cases, such as dogs facing profound trauma from physical abuse, dog fighting, hoarding, and puppy mills, the average amount of time for rehabilitation in dogs taken in by another shelter was only 12 weeks. [16] That doesn’t mean some dogs don’t take longer. It also does not mean there are not challenges, even immense challenges, to placement. Some of these dogs spend a very, very long time in treatment before they are adopted.

But what it does mean is that dogs are incredibly resilient, we can build resilience in traumatized dogs, and we can—and eventually will—end the killing of all dogs for behavior reasons. In fact, the path to do so is now clear. Putting aside those who meet the definition of “irremediable physical suffering” which is often less than 1% of all dogs, a 99% live release rate for dogs is too low. And that isn’t even considering a sanctuary placement option. [17] In fact, if Doug Rae is correct, and the most recent scholarship on the issue would indicate he is, less than 1% of dogs fall outside our ability to rehabilitate right now and for whom long-term protective placement such as a sanctuary is a viable option. For the rest, cohousing, playgroups, redesign of shelters, foster care, basic training, improvements in rehabilitation regimens (i.e., moving toward true behavior modification and away from rudimentary concentration on either negative punishment or positive rewards in a way that is nothing more than stimulus control operant conditioning), drug therapy, and other techniques would allow shelters to ultimately place them.

For the very small percentage of dogs where a sanctuary is warranted, we can learn from them in ways that are now precluded by their killing. Dogs who are experiencing behavior problems, including aggression, have much to teach us about resilience and resolution, lessons we can only learn by placing such dogs in safe, nurturing and protected environments that allow us to observe and gauge their behaviors and, ultimately, promote their healing and rehabilitation. The knowledge we would gain by doing so would allow us to translate what are now enigmatic, seemingly insurmountable, behaviors into effective, life-affirming solutions.

A line in the sand

For some, “Save Them All” isn’t an ultimate destination on a roadmap to a brighter future, a roadmap that requires diligent revision and updating to ensure its accuracy. Rather, it is a marketing ploy, a way to emotionally manipulate people into donating their money to organizations with already over-bloated coffers, even while they champion a conception of “No Kill” that only applies to 90% of animals while ignoring the needs of the remaining 10%. As such, they will argue that our approach is setting the bar too high; that there are “lines in the sand” when it comes to perceived aggression we should not and cannot cross, familiar refrains once stated about each and every one of the now widely accepted innovations of the No Kill Equation. Consider that 10 years ago, roughly 120,000 dogs and cats were being killed in the state of Michigan. That number now stands at about 27,000, a decline of over 75%. Indeed, over 50 of Michigan’s 80 sheltered communities have live-release rates of at least 90%, with many at 95%, and some reporting success as high as 98-99%. Every one of these achievements would have been impossible but for shelter directors willing to step over what were once dogmatically championed—but in the end proven to be nothing more than imaginary and arbitrary—”lines in the sand.”

Despite the tremendous improvements occurring in the field of animal sheltering as a result of the No Kill movement, the animals still being killed matter just as much as those who no longer face death, and for many of them, such as behaviorally challenged dogs, our duty is compounded by the fact that we—as humans—are often responsible for their condition through our neglect, abuse, and undersocialization. [18] Relieving us of that burden by killing these dogs does not result in redress for them.

A better and ethically consistent future in animal sheltering inevitably awaits us if the No Kill movement can continue to do what it has always done until every last animal entering our nation’s shelters—whatever the species, whatever the challenge—no longer faces killing: overcome the flawed but mutable traditions we have inherited from prior generations.

The sooner we recognize the need for change and further innovation, the sooner we will find the motivation and tools to bring that brighter future into reality. The sooner we bring that brighter future into reality, the fewer the number of dogs who will die. And the sooner we could move past yet another in a long line of ultimately non-existent “lines in the sand.”


[1]  The No Kill Advocacy Center is moving away from terms like “lifesaving” or “saving lives” to describe when animals leave shelters alive because for most animals entering shelters, this description is inaccurate. If an animal has been hit by a car or is suffering from a serious disease and enters a shelter which provides that animal with veterinary assistance that prevents death, that animal has been saved by a shelter. Likewise, when a rescue organization takes an animal from death row at a shelter, that animal has also been saved because the rescuer intervened to prevent the animal from being killed by someone else.

But when the term “lifesaving” is used to describe a shelter choosing to adopt out an animal instead of killing that animal, killing is implied to be a natural outcome of animal “homelessness” that must be overcome, which it is not. Homelessness is not a fatal condition—or at least it shouldn’t be. Moreover, the vast majority of animals who enter shelters are healthy and not in danger of dying but for the threat the shelter itself poses. Shelter employees cannot accurately be described as having “saved” an animal when the only threat the animal faced was the one that they themselves presented. In short, if someone was threatening to kill you, and chose to let you live instead, would you describe that person’s actions as having “saved” you?

[2]  The No Kill Advocacy Center defines “irremediable physical suffering” as an animal who has “a poor or grave prognosis for being able to live without severe, unremitting pain even with prompt, necessary, and comprehensive veterinary care,” such as animals in fulminant organ system failure. No Kill Advocacy Center, Defining No Kill,

[3]  No Kill Advocacy Center, No Kill 101,

[4] There are others, including the killing of wildlife based on arcane regulations, reactive health department mandates, and the unethical policy equivalent of “breed discriminatory legislation,” the belief that wild animals deemed “non-native” should be killed as a matter of course.

[5]  See

[6]   See

[7]  See

[8]  This position paper only addresses aggression in dogs, as cats (and other animals) do not pose similar safety risks. While it is not ethical to kill any animal for behavior reasons, there is no need to delay finding homes for cats deemed “fractious.” They can be sterilized and returned to their habitats if they are not social with humans and are used to living outdoors or they can be adopted out immediately if they are. Simply put, people will adopt cats with “catitiude.” In fact, over 15 years ago, the “open admission” shelter in Tompkins County, New York, eliminated any “behavior category” for cats and thus any killing of cats for “behavior,” “aggression,” or being “feral.” This is not to say that cats who experience behavior issues in the shelter do not warrant changes in shelter housing, shelter treatment, and behavior intervention to address those needs. They do. Our point here is only that they can be adopted out despite those issues because resolution of behavior challenges is almost always done by getting them out of the shelter. Moreover, for those who do need further treatment, we do not believe that it precludes immediate adoption and believe treatment in the home will be more effective and focused, as it is for many dogs. See fn. 14.

[9]  See

[10]  Flower, S., The Effect of Play Group on the Behavior of Shelter Dogs,

[11]  Patronek, G., Bradley, J., No Better Than Flipping a Coin: Reconsidering Canine Behavior Evaluations in Animal Shelters,

[12]   See

[13]  See fn. 1.

[14]  Auerbach, K., Placing Medium and Large Breed Shelter Dogs with Behavioral Challenges in Foster Homes: Results and Outcomes, It is worth noting that because the shelter fell under the jurisdiction of the police department, the program emphasized public safety as a primary concern.

[15]  No Kill Advocacy Center, What Shelters Owe Traumatized Animals,

[16]  See

[17]  It should be noted that a sanctuary should not be seen as a place where one gives up on animals with “severe aggression” or “trauma.” Instead, sanctuaries should be viewed as an environment where the animal and public are  protected during long-term rehabilitation and barring that, where the dog is provided permanent placement that meets the needs of the individual for life.

[18]  This is true even if their behavioral pathology is endogenous and profound because there is redress.


Have a comment? Join the discussion by  clicking here.