Our informal survey of shelters with training programs suggests that most emphasize basic manners training. Behaviors taught include “sit,” “down,” walking on-leash without pulling, interacting with people without jumping on them, and sitting at the front of the kennel to greet people. Some shelters teach additional behaviors intended to facilitate positive interaction with potential adopters, such as name recognition, eye contact, hand targeting, and trick training (e.g., “shake”). Some shelters that house dogs for longer periods also include more complex behaviors like “stay,” walking past other dogs without interacting or reacting, waiting to eat food or pass through doorways until released, going into a crate and settling, “drop it,” “leave it,” and sitting automatically in front of doors/gates or when people approach (see Figure 9.1). Ideally, training occurs while dogs remain available for adoption so that their chances for placement are not compromised.

Most shelters rely on specially trained volunteers to conduct training sessions. A popular model is to offer tiered training to volunteers. Lower-level volunteers focus on socialization and teaching simple tasks, while higher-level volunteers employ their advanced skills to take on sophisticated training projects, including agility and nosework, or to work with more behaviorally challenged dogs. Some shelters are fortunate enough to have capable staff who can allot time to training dogs. Other shelters have dedicated behavior staff that focus exclusively on training, behavior modification, and coaching volunteers who assist them with their duties. Local dog trainers sometimes donate their time to help shelter dogs and, at the same time, acquire valuable training experience. Some shelters also offer scheduled training classes specifically for volunteers and staff working with select shelter dogs. Whenever possible, we recommend holding classes in an area where adopters can watch. Staff can demonstrate effective handling and training skills while showcasing dogs available for adoption.

The majority of shelters we interviewed employ reward-based methods almost exclusively for establishing new skills and discouraging unwanted behaviors. The popularity of such methods is in line with findings that dogs are more responsive and less stressed by training procedures that use appetitive over aversive incentives (Hiby et al. 2004; Herron et al. 2009; Deldalle & Gaunet 2014). Reward-based training is also thought to foster positive interactions between dogs and people (Marston & Bennett 2003; Wells 2004). Food is the most common form of reward used in shelters—with good reason, as Feuerbacher and Wynne (2012) demonstrated that social interaction (praise and petting) was an ineffective reinforcer for shelter dogs. The integration of conditioned rewards (a “marker,” like a clicker) is encouraged, but most programs emphasize the use of treats and verbal praise. Despite claims to the contrary (Pryor et al. 2002), most shelters report that clicker training is too complex for volunteers to implement and generalization to the home environment is poor. Thorn et al. (2006) also observed quicker responsiveness and better retention with the use of a verbal marker over the clicker when teaching shelter dogs to sit.

Whether these training programs have the desired effect of staving off behavioral deterioration is still an open question. Numerous studies have confirmed that shelters are stressful environments for dogs (Hennessy et al. 1997, 2001; Coppola et al. 2006; Hiby et al. 2006) and that regular human contact, in the form of socialization and obedience training, reduces stress in many shelter dogs (see Figure 9.2). Tuber et al. (1996) demonstrated that the mere presence of a person reduced the stress experienced by dogs placed in a novel environment. Hennessey and his colleagues showed that human interaction involving gentle stroking and soothing talk diminished the impact of the stress experienced in the shelter environment (Hennessy et al. 2002). In another study, dogs that received one 45-min session of gentle stroking, play, walking, and reward-based obedience training on their second day after arrival at a shelter resulted in reduced salivary cortisol-mediated stress responses compared with dogs that did not experience the session of human contact and training (Coppola et al. 2006). Likewise, Shiverdecker et al. (2013) found that one 30-min session of human contact in a quiet environment, whether the person was passive or actively engaged in petting or playing and training, substantially reduced plasma cortisol levels, vocalizing, and stress panting. To improve quality of life in the shelter and facilitate adjustment to a new home, Tuber et al. (1999) recommend daily human contact in a “real-life” room, combined with teaching dogs to relax quietly in crates and to sit when people approach their kennels.

Studies focused on how shelter dog behavior influences adopter choices have produced conflicting findings. Some have found that physical appearance trumps behavior for shelter dog adopters (Luescher & Medlock 2009; Protopopova et al. 2012). However, in another study, members of the public stated that when choosing a dog to adopt, they would place more emphasis on the dog’s behavior than on its appearance (Wells & Hepper 1992). People specified that they prefer dogs that come to the front of the kennel, sit, interact, and do not bark. According to Wright et al. (2007), a dog engaged in friendly behavior toward a human is perceived as more adoptable than the same dog exhibiting neutral or aggressive behavior. Weiss et al. (2012) discovered that some adopters were even attracted to dogs that jumped up on them. Thus, though it seems logical to assume that teaching basic manners to shelter dogs will make them more appealing to adopters, some “impolite” behaviors may actually be desirable in a shelter dog. It might be advantageous for staff to avoid discouraging any behavior that is inherently social.

Although some studies suggest that training programs improve dogs’ chances of adoptive placement, convincing evidence to support that claim is sorely lacking. Braun (2011) reported that the number of dogs remaining long-term in one Viennese shelter was cut in half after the shelter instituted a dog-walking program. (The volunteers were coached by a professional dog trainer, which implies that the dogs received training during their walks.) At a shelter in Illinois, college undergraduates trained dogs to sit and generalized the response to novel people (Thorn et al. 2006). Staff and volunteers anecdotally reported that the trained dogs were easier to handle, came to front of the kennel to sit, barked less, jumped up less, and engaged in less spinning, pacing, and lunging.

Luescher and Medlock (2009) studied the impact of daily sessions in which a trainer taught randomly selected shelter dogs to sit (see Figure 9.3). The dogs were also rewarded for not jumping up and not barking when people approached their kennels. In addition, shelter staff delivered food treats whenever passing by the kennels to encourage the dogs to come to the front. The dogs that received training were 1.4 times more likely to be adopted than untrained dogs, although there was no effect on length of stay. However, staff and volunteers were not blind to which dogs were trained so they may have biased adopters toward choosing dogs that had participated in the program.

Reasoning that social, friendly dogs would appeal to adopters, researchers at a Florida shelter looked at the effect of teaching dogs to make eye contact with people (Protopopova et al. 2012). Dogs were randomly assigned to one of three groups: (i) dogs that received training, (ii) control dogs that were not removed from their kennels, and (iii) control dogs that were removed from their kennels for the same length of time each day as the training dogs but were fed treats according to a time schedule, independent of their behavior. Each dog in the training group was taught to offer eye contact during 15-min daily sessions over the course of 6 days. Ten different trainers were used to encourage generalization. Although researchers confirmed that the training dogs did indeed learn the eye contact response, these dogs were no more likely to be adopted than the two groups of control dogs. There was also no effect on length of stay in the shelter.

Dogs at an Ohio shelter were randomly assigned to an enrichment group (foraging enrichment and basic manners training) or to a control group (regular shelter activities) (Herron et al. 2014). The reward-based training, conducted daily by nine different trainers, focused on establishing desirable behaviors, such as sitting, making eye contact, and coming to the front of the cage, and discouraging undesirable behaviors, such as jumping up and barking. After 6 days, there were more dogs in the training group that sat, lay down, and barked less than in the control group and fewer dogs that jumped up. However, like the previous study, this research failed to demonstrate the impact of training on adoption rates.

Most of the shelters we surveyed reported that because they have insufficient resources to provide training to all dogs in their shelter, they target specific subgroups. Dogs with minor or moderate behavior concerns often receive attention, such as those prone to jumping up, becoming highly aroused, or mouthing people. Some shelters choose instead to target dogs that are challenging to place for reasons unrelated to behavior, such as pit bull types. Patronek and Glickman (1994) propose that increasing public demand for dogs over 1 year of age would have the most significant impact on adoption rates. If this is true, shelters with training programs might be wise to focus on another subgroup—older juveniles and adults. Perhaps studies that examine the impact of training programs on adoption rates for specific subgroups, such as these, would yield more compelling results to support the idea that training enhances adoptability. On the other hand, it might be most informative to evaluate the effect of training programs on return rates for difficult-to-place dogs instead of adoption rates or length of stay. Dogs that are more challenging to place may also be at greater risk of re-relinquishment after adoption, and basic manners training may reduce the likelihood of returns (Wells & Hepper 2000).

Behavior modification programs within the general shelter environment

Shelter dogs can present with the full gamut of behavior problems. Information about a dog’s behavioral tendencies should be obtained from the owner at intake, if available, by staff during daily shelter operations and through a standardized behavior evaluation (see Chapter 6 for more detailed information on intake and assessments). Bear in mind that while owners are willing to disclose some information about their dog’s behavioral history, one study suggests that unless they believe the information to be confidential, they will likely withhold accounts of owner-directed aggression and stranger-directed fear (Segurson et al. 2005). Private shelters in the USA rate aggression to people as the most serious behavior problem and, in our informal survey of shelters with behavior programs, few shelters work with dogs that are aggressive to people except in very specific circumstances. Problems such as guarding food, aggression to dogs, and fear are considered most responsive to treatment in the shelter environment.

Food guarding

Guarding food, bones and other highly valued consumable items is a prevalent behavior among pet dogs (Luescher & Reisner 2008), and all widely used standardized behavior evaluations in US shelters, including Assess-A-Pet™ (Bollen & Horowitz 2008), Match-Up II™ (Dowling-Guyer et al. 2011), and SAFER® (Weiss 2007), incorporate at least one subtest to elicit it. Though Marder et al. (2013) found that most owners do not consider such behavior problematic in the home, limited-admission shelters report that aggression over food is one of the most common reasons they refuse admittance (Center for Shelter Dogs 2012). Mohan-Gibbons et al. (2012) determined that not only was food guarding the most common behavioral reason for considering dogs unadoptable but also that only 34% of the 77 shelters they surveyed attempted to modify the behavior. Thus, many dogs in shelters are likely euthanized for food guarding.

Of the shelters with structured behavior modification programs that we surveyed, most work with dogs that exhibit aggression over food. Treatment typically consists of regular desensitization and counterconditioning (DSCC) sessions designed to help dogs feel more relaxed when people interfere with them while they are eating from a food bowl (see Figure 9.4). Examples of this type of approach are described by Wood (2011) and by the ASPCA Virtual Pet Behaviorist (http://www.aspca.org/pet-care/virtual-pet-behaviorist/dog-behavior). Wood (2011) reports that 89% of the dogs entering Humane Society of Boulder Valley’s food-guarding program graduate and are successfully adopted, based on a 1-year follow-up check. Exercises that accustom dogs to trade items or to drop items from their mouth on cue are sometimes used as an alternative approach or included as an adjunct to DSCC sessions (Donaldson 2002).

Mohan-Gibbons et al. (2012) experimented with placing 96 shelter dogs that exhibited food guarding on the SAFER® behavior evaluation. While awaiting adoption, these dogs were provided food ad libitum on the assumption that guarding might be diminished or eliminated if the resources were unlimited. Adopters were provided a set of instructions designed to condition the dog to feel comfortable eating when people were nearby. Follow-up telephone calls were conducted by shelter staff at 3 days, 3 weeks, and 3 months to assess the frequency and severity of food guarding. Of the 60 adopters who were contacted at least once, only six reported observing any aggression over food, chew items, or toys. Compliance with the instructions was very poor, with only a third of owners “free-feeding” their dogs during the first 3 weeks or providing food in a puzzle toy. By the 3-month mark, 31 of 35 owners reported that they could pick up the food bowl while the dog was eating. Slightly more than half said that knowing about the dog’s history of guarding caused them to take more precautions around the dog than they normally would have. During the 3-month follow-up period, no dogs bit their owners over resources and none were returned to the shelter because of food aggression. This research suggests that dogs showing resource guarding can be successfully adopted with little to no intervention.

In another effort to determine the impact of food aggression on adoption success, 97 adopters of dogs from a Massachusetts shelter were contacted after the dog had been in the home a minimum of 3 months (Marder et al. 2013). Twenty of the 97 dogs had exhibited resource guarding during the Match-Up II behavior evaluation. However, only half of these dogs (n = 11) reportedly showed aggression over food in the home. Seventeen dogs that did not show food guarding during the behavior evaluation did so in the home. This means that the likelihood of a dog that displayed food aggression in the shelter also showing the behavior in the home was only 55%, whereas the likelihood of a dog that did not display food aggression in the shelter also not showing the behavior in the home was 78%. Of the dogs that did guard food in the home, most exhibited the behavior infrequently, and the behavior tended to be restricted to growling or showing teeth, rather than snapping, lunging, or biting. Few adopters labeled their dogs as food aggressive, and 86% indicated that they would adopt the same dog again (compared with 89% of adopters of a dog that did not show food aggression in the home).

These studies suggest that euthanizing dogs because of food-guarding behavior in the shelter environment may be unnecessary. First of all, the behavior is amenable to treatment prior to placement. Second, if shelters do not have the resources to devote to treating dogs that guard food, few of these dogs end up exhibiting the behavior in the home. When they do, the aggression is usually not severe, and adopters do not consider the behavior problematic.

Although these findings may warrant a reexamination of policies regarding the disposition of dogs that guard food, shelters should not deem them evidence for adopting out all such dogs haphazardly. Many of the shelters that report successful adoptions after treatment choose not to work with dogs if they display severe guarding (such as hard bites or few warning signals) (Wood 2011). Likewise, Mohan-Gibbons et al. (2012) excluded dogs from their study that showed behaviors they considered indications of severe guarding, such as leaving the bowl to bite the fake hand or biting multiple times. Marder et al. (2013) restricted their study to dogs that had been in the adoptive homes for at least 3 months; however, there were no severe guarders returned to the shelter before the 3-month mark (A. Marder, pers. comm.).

In summary, shelters should not consider food aggression an automatic death sentence. If resources permit, treatment can often diminish or eliminate the problem before adoption. If resources do not permit, dogs that exhibit mild to moderate guarding may be successfully adopted without treatment.

Intraspecific aggression

Dogs that show aggression toward other dogs pose a significant challenge because they can be difficult to manage in the shelter environment, and they require special placement to ensure that adopters are willing and able to accept the responsibility. According to a study conducted by the Center for Shelter Dogs, aggression toward other dogs ranks in the top three reasons why private shelters refuse dogs’ admittance (Center for Shelter Dogs 2012).

Many shelters do not attempt to modify the behavior of dogs that exhibit intraspecific aggression, but those that do often rely on obedience training techniques that will allow adopters to manage the behavior. Dogs are taught to walk calmly on-leash and/or to maintain eye contact with a handler while in the presence of other dogs. Some shelters go one step further by using DSCC to condition dogs to relax and anticipate good things when other dogs are near. For instance, a contingency can be arranged such that whenever the aggressive dog perceives another dog, it receives especially tasty treats or a favorite toy. This is repeated until the aggressive dog shows a change in emotional response and appears to anticipate good things when other dogs appear. This does not necessarily lead to the dog being able to interact with other dogs without becoming aggressive, but at least the dog can be walked past other dogs without causing a fuss. A variant on this approach, called Behavior Adjustment Treatment, combines DSCC with a negative reinforcement contingency in which the aggressive dog is rewarded for calm behavior by moving him away from the other dog for a brief period (Stewart 2012).

Orihel and Fraser (2008) demonstrated the effectiveness of a DSCC program for interdog aggression in shelter dogs. They identified 16 dogs that exhibited aggression to other dogs in a standardized behavior evaluation. Nine treatment dogs received daily 30-min sessions in an outdoor area. During each session, the dogs were repeatedly approached by stimulus dogs and rewarded for sitting or making eye contact with their handler. Aggressive behaviors were interrupted by using a leash and a head halter (Halti®) to direct the dog’s head and body away from the stimulus dog. Seven control dogs spent 30 min each day in the same outdoor area but received no treatment. After 10 days, all dogs were reassessed. Seven of the nine treatment dogs were significantly less aggressive on retest, whereas five of the seven control dogs were worse. Unfortunately, there was no sustained effect once treatment was ceased. This confirms that the severity of dog aggression can be reduced, but continued treatment is necessary to maintain behavioral change. Alas, 30-min daily sessions that require two handlers are too onerous to be feasible for most shelters.

An increasing number of shelters address interdog aggression problems by implementing a playgroup program, such as Playing for Life!™ (Sadler 2014). Dogs are provided with frequent off-leash access to other dogs, typically in a large, outdoor, fenced area. Sadler, who was instrumental in setting up playgroups at Southampton Animal Shelter Foundation and Longmont Humane Society, claims that by providing shelter dogs with repeated socialization opportunities, not only do they learn to relax, interact, and sometimes even play with other dogs, they may also be less likely to exhibit other undesirable behaviors such as kennel reactivity, generalized fear, pulling and lunging on-leash, poor impulse control, mouthing people, and other manifestations of unruliness. Dogs that are aggressive to other dogs often need a gradual introduction to playgroups. They are first exposed to carefully selected dogs that are especially tolerant and playful, they may require correction for inappropriate aggressive behaviors, and they may need to be muzzled for a period of time. Tangible rewards for appropriate behavior are typically considered unnecessary because the goal is for socializing to become so reinforcing that the dog is no longer motivated to behave aggressively. In addition to the direct benefit of reducing interdog aggression, shelter staff get to know the dogs in a more naturalistic setting so they can better describe the dogs’ personalities and behavioral tendencies to adopters. Especially compatible dogs can be identified and group-housed together or even offered for adoption as a “package deal.” Regular playgroup experience may also ward off behavioral deterioration and improve a dog’s quality of life while in the shelter. Play sessions provide the dog with physical exercise, mental stimulation, fresh air, and temporary relief from the stress of kennel life. All of these factors are likely to improve a dog’s adoption potential (see Figure 9.5).