9.1.1 Decision‐Making

It’s not surprising that the quality of a decision about an animal is assumed to be directly related to the amount of information gathered about that animal, especially when it comes to complex challenges like predicting future behavior. The more puzzle pieces gathered, the more complete the picture. For example, staff in Australian shelters felt that they could make better decisions if they were given more time to get to know the dogs in their care (Mornement et al. 2010). Goold and Newberry (2017b) suggest that shelters should wait as long as possible before deciding on the fate of an animal. Not only does that allow for increased behavior observations of the dog in various scenarios; there is also evidence that a dog’s behavior is particularly unstable during the first few days or weeks in the shelter environment (Goold and Newberry 2017a, 2017b; Kis et al. 2014). For instance, Goold and Newberry (2017b) found that shelter dogs increased their friendly behavior toward humans over the first two weeks in the shelter, with the most sizable change happening in the early days after arrival. The magnitude of the change varied across individuals; dogs who were the most social on the first day in the shelter were not necessarily the most social on day 15. They note that individual differences in the day‐to‐day variation of shelter dog behavior have been largely unstudied.

However, there is a substantial cost to gathering information. It takes time and resources. The quest for information should not come at the cost of holding dogs in the shelter for longer than necessary. Lengthening a dog’s stay is not only likely to have a detrimental effect on that dog, it also prevents housing, supplies, and personnel from being directed toward other dogs in need. A shelter’s overarching goal should be to identify adoption candidates as quickly as possible and move them into adoptive or foster homes or into a relocation program. Likewise, dogs who are not adoption candidates should be identified and swiftly moved along a path to outcome; dogs with treatable behavior problems that make them immediately unadoptable can be placed in a behavior modification program, and dogs with untreatable behavior problems or unacceptably poor welfare should be humanely euthanized.

Striking an optimal balance between speed and accuracy when making decisions is no easy feat. Effective decision‐making involves collecting information in an efficient manner and gauging when enough information has been accumulated to be confident in the decision (Drugowitsch et al. 2015). How much evidence is required depends in part on the consequences of inaccurate decisions. In the case of decisions about shelter dogs, the consequences are considerable. As mentioned above, mistaken judgments about a dog can result in his needless death. Errors in the opposite direction can result in a dangerous dog being placed in the community. Both types of miscalculations are likely to weigh heavily on shelter personnel, possibly leading to compassion fatigue (Hill et al. 2020; Scotney et al. 2015).

Not all decisions are necessarily fraught with uncertainty, though. Dogs that are brought to the shelter with credible reports of severe aggression toward people or other dogs should be euthanized as quickly as legally possible. There is no benefit to holding these dogs in a stressful shelter, during which time staff may become attached to them. Dogs that come in with trustworthy reports from a surrendering owner and dogs that are highly social with people and dogs, seemingly unfazed by the stresses of the shelter, can be fast‐tracked for adoption. Be aware, though, that quick decisions based on only a few observations (“thin‐slicing”) are most likely to be accurate when made by behavior experts; thin‐slicing by non‐experts is notoriously prone to error (Croskerry 2006). Dogs who fall in between these two extremes—the ambiguous ones—are the dogs for whom decisions should be based on an adequate accumulation of behavioral observations.

9.1.2 The Observational Assessment Model

Traditionally, many shelters have placed their faith almost exclusively in standardized behavior evaluations to reveal information critical to pathway planning. But, as discussed in Section 9.2.9, behavior evaluations have not proven highly accurate nor precise when used to predict behavior, particularly aggression, after adoption. It has been suggested that standardized evaluations, in part because of their provocativeness, result in a significant number of dogs exhibiting aggression who ultimately would not do so in a home (Patronek and Bradley 2016). Thus, rather than relying on a single “snapshot” of behavior, a more rational approach is to implement what has been called a “mixed assessment model” (Rayment et al. 2015) or a “longitudinal observational assessment model” (Goold and Newberry 2017b) (hereafter referred to as the observational assessment model). Such an approach relies on collecting behavioral data from all of the players who are likely to know or learn something about a dog in a variety of contexts, in addition to conducting a standardized evaluation. Shelter personnel, volunteers, adoption counselors, and even members of the public may interact with the dog and, without a formalized process for collecting their observations, those data are lost. Moreover, observing the dog in naturally occurring situations, such as walking by a children’s playground or meeting a potential adoptive family, can uncover information that wouldn’t be revealed in a standardized evaluation. This approach, by definition, involves multiple sources weighing in on the dog’s behavior in numerous contexts, which is expected to reveal behavioral consistencies and, thus, is more likely to capture the dog’s personality. A drawback, however, is that most of the people observing the animal and generating these data are not trained experts in canine behavior. Currently, there are very few studies addressing the integrity of such information. Including a standardized behavior evaluation in the assessment process is advisable, as this ensures that behavior staff also interact with the dog.

A distinct advantage of the observational assessment model is that if concerning behavior is exhibited in one context, it can be further explored in other situations. This is especially important for behaviors that might warrant euthanasia. Shelters should only make euthanasia decisions based on a single incident of aggression if the aggression is truly egregious, such as a severe attack on a person or other dog. Otherwise, problematic behavior should be safely confirmed in other situations to guard against potential errors. As more research accumulates, it may prove feasible for shelters to also substantiate behavioral observations with correlated physiological measures, such as heart rate and heart rate variability via heart rate monitors (Bergamasco et al. 2010; Rayment et al. 2013; Vincent and Leahy 1997), activity levels via accelerometers (Hoffman et al. 2019; Jones et al. 2014), and arousal via leash tension meters (Shih et al. 2020).

Dowling‐Guyer et al. (2011) express the legitimate concern that this model may be difficult to implement in shelters because of the limited situations and stimuli to which shelter dogs are exposed. However, many shelters already rely on behavior information collected from more than one source, including surrendering owners, daily care staff observations, and a brief medical/behavior intake exam (Marder 2015). Even staff at high‐intake shelters are likely to be able to observe dogs with ambiguous behavior in at least six to nine situations; Goold and Newberry (2017a, 2017b) felt this was an adequate number of observations of shelter dogs’ interactions with people and other dogs to provide an accurate indication of future behavior. More potentially worrisome is the concern that busy shelter staff do not have the time to record their observations or that behavior staff do not have the time to compile and assimilate the data. The good news is that with new technologies emerging, such as mobile apps that feed directly into a shelter’s database, staff can quickly and efficiently enter behavior data. Records can be autofilled with the dog’s previous data so that only significant changes in behavior are flagged for the behavior team.

9.2.1 Information from Relinquishing Owners

It is a commonly held belief among shelter professionals that owners won’t be honest about the behavior of the dog they are surrendering (Duffy et al. 2014; Stephen and Ledger 2007). The reasons for this distrust are many. One perspective is that owners may exaggerate their dog’s behavior problems because they feel the need to justify why they are giving up the dog. Alternatively, owners might downplay behavior problems to improve their dog’s chances for rehoming. Yet another possibility is that once owners have made the decision to surrender their pet, they are uninterested in providing information and will respond to questions carelessly. In addition, relinquishing owners have been shown to have weaker attachment to their pets, leading Kwan and Bain (2013) to speculate that perhaps relinquishing owners are less observant and, therefore, less knowledgeable about their dog’s behavior. However, DiGiacomo et al. (1998) found that many owners struggle with the decision to surrender, often resorting to the shelter after expending considerable effort to keep or rehome the animal, which suggests that some owners do at least have a strong sense of responsibility toward the pet and probably a reasonably accurate perspective on their behavior.

Contrary to those concerns, research indicates that owners often do disclose information that is valuable for making outcome and placement decisions. Segurson et al. (2005) asked people surrendering their dogs to complete the 103‐item Canine Behavioral Assessment and Research Questionnaire (C‐BARQ). Half of the people were told that their responses were confidential and would not be shared with shelter personnel; the other half were told that their responses would be used by the shelter to match their dog to a new owner. The researchers found that when owners believed the information would be kept confidential, they were more likely to disclose that their dog exhibited aggression toward family members or fear of strangers. There was no difference between the groups in their willingness to report that their dogs showed aggression to strangers, aggression toward or fear of dogs, non‐social fear, separation‐related behaviors, or attachment/attention‐seeking behaviors. Both groups reported higher incidences of problem behavior than a control group of owners surveyed when bringing their dogs to a medical referral clinic.

In contrast, a subsequent study that used a shorter 42‐item version of the questionnaire, called C‐BARQ^(S), found no difference between the responses of owners who were told that their information would be kept confidential, owners who were told that their information would be shared with the shelter, and owners who weren’t told anything about how their information would be used (Duffy et al. 2014). The most apparent difference between the studies, other than the length of the questionnaire, was that Segurson et al. (2005) approached owners face‐to‐face, explaining the questionnaire and asking them to complete it, whereas the questionnaires were simply made available and self‐administered in the Duffy et al. (2014) study. In general, face‐to‐face interactions are more susceptible to social desirability biases; people feel pressured to give answers that they believe the other person wants to hear (Durmaz et al. 2020; Lelkes et al. 2012; Richman et al. 1999). Thus, it’s possible that self‐administered questionnaires are more likely to result in accurate disclosures of problem behavior, but more research is needed to fully understand the best way to gather such information from owners.

As a tool to predict how a dog is likely to behave in an adoptive home, owner reports do not appear to be terribly useful on their own. Stephen and Ledger (2007) followed up with adopters at two‐ and six‐weeks post‐adoption and found only a few consistencies with what owners had reported about their dogs at surrender. At two weeks, fear, anxiety, or aggression toward veterinarians, aggression to unfamiliar dogs or unfamiliar people, stealing food, chewing furniture, mounting, and anxiety when left alone were the behaviors correlated with surrendering owner reports. By six weeks there were fewer correlations. Duffy et al. (2014) found that adopter reports only corresponded to surrendering owner reports for aggression toward unfamiliar people, urination when left alone, and destructive chewing. McGuire et al. (2020) specifically asked surrendering owners about their dogs’ tendency to guard food, toys, or chew items and found no positive relationship with resource guarding reported in the adoptive home. They did confirm, though, that dogs who were reported by their owners to not guard resources were similarly unlikely to guard in the new home.

If a Good Samaritan brings a stray dog to the shelter, it is worth collecting whatever behavioral information the finder might know about the dog. They presumably captured and handled the dog, at least minimally, or they may have even kept the dog at their home for a time while they tried to find the owner themselves. Those interpreting this information should take into consideration the amount of time the person spent with the dog, the circumstances under which the information was gathered, and the nature of the information reported. Likewise, with increasing numbers of dogs being relocated from one shelter to another, the author advises that the receiving shelter request any behavioral records from the originating shelter, bearing in mind that the value of those records is a function of the shelter’s behavior assessment process and the time the dog spent there.

9.2.2 Medical Staff Observations

Every dog coming into a shelter typically undergoes, at the very least, a medical intake exam, so medical personnel can be an economical and valuable source of information about the dog’s reaction to unfamiliar people, physical handling, restraint, and potentially unpleasant experiences. However, it is important to keep in mind that how a dog behaves during a medical exam may be influenced by his discomfort with the procedures being performed during the exam and his history with medical care. Some dogs who are tolerant of or even friendly toward people in every other situation may become defensively aggressive or extremely fearful when handled by medical personnel, especially if they are in pain, because of previous unpleasant experiences in medical contexts (Mills et al. 2020).

There is reason to suspect that an animal’s behavior during a veterinary exam may not be reflective of behavior in other circumstances. Research substantiates that veterinary exams are stressful for most dogs (Riemer et al. 2021; Travain et al.2015). Döring et al. (2009) found that close to 80% of owned dogs exhibited fearful behavior during a standardized veterinary exam. Dogs unaccompanied by their owners experience significantly more stress (Csoltova et al. 2017; Stellato et al. 2020), especially when physically manipulated (Stellato et al. 2019). Owner‐surrendered dogs admitted to a shelter are essentially in a comparable situation, with a stranger rather than the owner taking them for their exam. Furthermore, intake exams typically occur shortly after admission to the shelter, when dogs’ stress levels are found to be highest, as indicated by elevations in cortisol (Coppola et al. 2006; Hennessy et al. 1997; Hiby et al. 2006).

While there is no published research specifically examining the value of behavioral information gleaned during a standard veterinary exam in the shelter environment, Lind et al. (2017) found no correlation between veterinarians’ assessments of stress and a behavior evaluation of owned dogs at a veterinary clinic. However, there is at least some indication that dogs who are frightened during a veterinary exam are likely to also show fear in other circumstances. Edwards et al. (2019) found that dogs reported by their owners to exhibit fear when examined by a veterinarian (55%) are also likely to exhibit fear in new situations (68%) and suffer from pain/touch sensitivity (35%) and non‐social fear (37%).

Behavior observations during a medical intake exam may be primarily useful for learning how the dog responds to stressful events, developing a handling plan for the dog during future medical procedures, and creating a behavior modification plan to reduce sensitivity to handling or restraint if necessary. Behavior information from medical personnel should be considered in light of the circumstances of the interaction and in conjunction with other observations in different contexts. Box 9.1 contains sample questions that medical staff can answer to collect critical data on the nature of the interaction and Box 9.2 contains a word bank to consistently characterize responses.

9.2.3 Daily Care Staff and Volunteer Observations

Daily care staff and volunteers often have more direct contact with a dog than anyone else in the shelter system. They typically interact with the dog during routine cleaning, watering, and feeding. In some shelters, daily care personnel and volunteers are also responsible for providing in‐kennel enrichment (beds, toys, chew items, etc.), transferring the dog to and from outdoor exercise areas or play yards, supervising playgroups, and walking and socializing with the dog.

However, behavior observations collected by daily care staff who haven’t received targeted training on that task may not be very accurate. In one study, dogs underwent a standardized behavioral evaluation and, simultaneously, shelter staff completed a questionnaire about each dog’s behavior toward adults, children, dogs, and cats; their reaction to being handled; how well they walked on leash; and so forth (van der Borg et al. 1991). Staff correctly predicted only one‐third of the behavior problems ultimately reported by adopters one to two months after adoption, whereas the behavioral tests predicted almost 75% of the problems.

One understandable explanation for the poor accuracy of daily care staff observations is that people may be reluctant to disclose negative information about a dog because they don’t want to jeopardize the dog’s chances for adoption. Therefore, daily care staff and volunteers may be biased toward providing positive feedback, as was the case in the van der Borg et al. (1991) study, in which staff predicted that dogs would not exhibit problem behaviors in the adoptive home when, in fact, they did. Goold and Newberry (2017a) found that close to half of the shelter staff they studied mistook aggressive behavior for excitement and frustration, but rarely did they mistakenly classify non‐aggressive behaviors as aggression. The staff ratings, therefore, indicated a tendency to judge dogs’ behavior positively.

In addition to an optimistic judgment bias, staff and volunteers who are untrained in animal behavior may misinterpret behaviors or be unable to recognize more subtle variations in behavior. Research on the validity of inexpert judgments of dog behavior provides mixed support for this. For instance, in comparisons of dog professionals, dog owners, and non‐dog owners rating behavior from video footage, familiarity with dogs wasn’t necessary for the accurate identification of friendly or happy dogs, but fearful dogs were better recognized by the professionals (Wan et al. 2012). Tami and Gallagher (2009), however, found that both experts and non‐experts were equally capable of identifying fear in dogs. When it came to identifying other types of behavior, being an expert didn’t seem to help that much. Tami and Gallagher (2009) found that both experts and non‐experts were equally poor at differentiating play from aggression. Wan et al. (2012) found that even highly proficient experts had trouble agreeing on when dogs were exhibiting behavior suggestive of “sadness” and “anger.”

It is particularly fascinating that as people gain more familiarity with dogs, they pay more attention to the whole dog. Experts reported that they were more likely to rely on multiple features of the dog when making their judgments than non‐experts (Wan et al. 2012). An eye‐tracking study showed that non‐experts focused more on dogs’ heads while dog trainers paid more attention to dogs’ bodies (Kujala et al. 2012). This attention to the whole dog may enable experts to more readily interpret dogs’ overall affective state. Expert dog trainers in the Tami and Gallagher (2009) study provided more information about the dogs’ motivations and emotions than the non‐experts. Similarly, in the Kujala et al. (2012) study, experts were more likely than non‐experts to make inferences about the dogs’ mental states, such as “the dog was reserved” or “the dog was friendly and playful.”

Fortunately, when non‐experts are provided with training, their proficiency improves. Flint et al. (2018) showed that providing people with written descriptions, pictures, and video examples of mild‐to‐moderate and high‐to‐extreme levels of fear significantly increased their ability to accurately identify levels of fear in new videotaped examples. Likewise, non‐experts who were provided with a one‐hour training session on reading and interpreting dog behavior were almost as adept at assessing dogs for their aptitude as military working dogs as professionals with 9–22 years of assessment experience (Fratkin et al. 2015).

Clearly, if we wish to collect accurate behavioral information from shelter staff and volunteers, targeted training is required. Training programs should emphasize the importance of attending to each of the relevant component features as well as the whole picture of the dog, enabling people to differentiate aggression from play and frustration and to detect subtle variations in emotional states, including fear, sadness, and anger. The ideal instructional format for learning to perceive and interpret dog behavior has not been investigated, so it is probably best to combine written descriptions with still images and video examples. There is evidence to suggest that incorporating all three media capitalizes on each of their strengths. Simplifying the task with photographs and written text may enhance initial acquisition of knowledge, whereas providing more complex video examples improves the learner’s ultimate performance (Arguel and Jamet 2009; Ganier and de Vries 2016; Höffler and Leutner 2007; van Hooijdonk and Krahmer 2008).

9.2.4 Walks and Field Trips

Shelters often incorporate walks by staff or volunteers into the routines of their dogs. Walks are believed to improve the health and mental well‐being of the dogs and reduce stress associated with being in their kennels (Coppola et al. 2006; Perry et al. 2020). In a study that specifically looked at the relationship between the impressions of volunteer dog walkers about certain dogs and the same dogs’ behavior during a standardized behavior evaluation, Shih et al. (2020) found that dogs who pulled less on the leash during walks were rated as more social during the evaluation. Other than that, the researchers found little correlation between volunteer reports and the evaluation results, although that may be because the judgments asked of the volunteers were quite zoomed out, such as “I don’t think this dog is suitable for a non‐experienced adopter.” The researchers felt that the volunteers might have been more tolerant of the dogs’ behavior than the evaluators and that they may have been reluctant to disclose behavior that would be perceived as negative.

Notwithstanding the findings of the Shih et al. (2020) study, a walk can reveal much about a dog’s behavior: his enthusiasm for getting out of his kennel, his comfort with having his collar held and a leash applied, how he reacts to other dogs as he walks through the shelter, his interest in interacting with the handler versus checking out the environment, his confidence in novel and outdoor locations, his inclination to pull on the leash, and his willingness to return to the kennel. Walks that take place off shelter property are even more informative. How does the dog react to passersby, children, joggers, cyclists, skateboarders, and/or other potentially provoking or frightening stimuli (see Figure 9.1)? Providing these details, along with strategies for managing any problematic behavior observed on walks, may promote regular dog walking post‐adoption. While encouraging adopters to walk their dogs doesn’t necessarily reduce return rates (Gunter et al. 2017), owners who walk their dogs report a greater sense of obligation toward their pets (Gunter et al. 2017; Hoerster et al. 2011), so educating them about dog‐specific walking strategies based on information from shelter walks could promote the human‐animal bond post‐adoption.

9.2.5 Socialization Sessions

We know that one‐on‐one interactions with people can be highly beneficial for reducing stress in shelter dogs (Coppola et al. 2006; Dudley et al. 2015; Hennessy et al. 1998; Shiverdecker et al. 2013; Willen et al. 2017). In addition to being important for stress reduction, socialization sessions can provide useful information about how a dog interacts with people. When gathering information, ask questions of the socializer that will inform good adoption matches and provide helpful information for new adopters. Is the dog motivated to interact with people? Does the dog enjoy petting? Is the dog comfortable playing with people? Is he content to stay in proximity? Does he respond to common verbal instructions, such as “Sit,” “Down,” and “Come?” Is the dog eager to perform behaviors to earn treats? What kind of treats most motivate him? Like reports provided by daily care staff and volunteers, information garnered from socialization sessions may be impacted by people’s level of behavioral expertise and by their desire to paint the dog in a positive light.

Another source of socialization information that should not be overlooked is adoption counselors, sometimes known as “matchmakers.” During meet‐and‐greet sessions with potential adopters, matchmakers are privy to observations that may not be attainable in other situations. How did the dog react to meeting an entire family? Did he gravitate to the adults or the children? Did he show a preference for males or females? How did he handle children’s loud voices and erratic movements? Was he comfortable being handled by an “average” adopter? Did he exhibit behaviors that have been shown to be particularly appealing to adopters, such as playing with a toy and lying down in proximity to people (Protopopova and Wynne 2014)? If the adopters rejected the dog, was it because of his behavior? The answers to these questions are extremely helpful not only to the behavior team but also to the matchmakers to prepare for future meet‐and‐greet opportunities.

9.2.6 In‐Kennel Behavior

The behavior of dogs in their kennels is typically monitored to detect potential welfare concerns but is not usually considered useful in predicting behavior outside of the kennel environment. In fact, a prevailing belief is that even friendly dogs may exhibit aggression when confined to a kennel. Goold and Newberry (2017a) found support for this. When dogs (at least those under six years of age) were in their kennels, the probability of displaying aggression toward people was highest compared to a variety of other situations. Curiously, they also found that when dogs were in their kennels, aggression toward other dogs was lowest. Anecdotally, this has not been the experience of many shelter professionals.

Clay et al. (2019) claim that monitoring in‐kennel behavior allows for early detection of potential problems ultimately revealed during a standardized behavior evaluation. They found that observing dogs in their kennels for one hour a day for the first five days in the shelter predicted the dogs’ behavior in an evaluation performed on the sixth day. Dogs who exhibited behaviors associated with fear, anxiety, and arousal (whining, panting, tenseness, lowered body, standing by the wall, etc.) in their kennels were likely to show the same behaviors during the evaluation. Aggression was rare both in kennel and in the evaluation. Over the course of the five days, fearful behavior declined for most dogs, and the dogs spent more time at the front of their kennels or resting in their beds, suggesting that they were acclimating to the environment. Dogs who were eventually euthanized were less relaxed and jumped up more in their kennels than dogs who were adopted. While monitoring in‐kennel behavior for an hour each day is not feasible for most shelters, the same information may be gathered through daily care observations and routine quality of life monitoring, possibly combined with accelerometer data, as accelerometers become more affordable and sophisticated (see Chapter 19).

9.2.7 Playgroups

Increasingly, shelters are offering playgroups for their dogs (see Chapter 13). Not only is participation in playgroups an excellent form of enrichment and physical exercise for dog‐friendly dogs, there is limited evidence that it also results in fewer stress‐related behaviors once the dogs return to their kennels (Belpedio et al. 2010; Johnson et al. 2013). Participation in playgroups has also been shown to influence how dogs respond to other dogs during a standardized behavior evaluation. Dogs who joined a playgroup session prior to undergoing an evaluation reacted more positively to a stimulus dog than control dogs who were evaluated without a playgroup session, although the two groups of dogs were studied at different times, so other factors could have accounted for the disparity (Flower 2016). Gfrerer et al. (2018) also found that regular interactions between adult dogs reduced the likelihood of aggression toward dogs in a standardized behavior evaluation. Over the course of eight weeks, adult military working dogs, housed in much the same way as shelter dogs, were provided with three hours of interaction with other dogs weekly. Most interaction occurred in a playgroup context, but some dogs could only safely interact leashed or behind a fence. These dogs subsequently exhibited less aggression toward a model dog and a live stimulus dog than control dogs who had not been exposed to other dogs (Gfrerer et al. 2018). Lacking long‐term follow‐up on these specific dogs, it is impossible to know whether the playgroup experience produced only a short‐term reduction in conspecific aggression, perhaps due to habituation, or a lasting change resulting from learning how to interact non‐aggressively with dogs.

In addition to the benefits of shelter playgroups for the participating dogs, playgroups can provide worthwhile information for matching a dog to an appropriate adopter. How does the dog react to being off leash in an enclosed area? Is the dog more interested in interacting with people or dogs, or does he prefer to explore the environment? How playful or energetic is he? Does he get along well with all dogs or only certain individuals (see Figure 9.2)?

The dog’s behavior toward other dogs while in a playgroup can be particularly enlightening because dogs may react differently toward other dogs when uninfluenced by a leash and handler (McConnell and London 2009). How does the dog respond to corrections from other dogs, and how appropriately does the dog deliver corrections to other dogs? Depending on the situation, the dog may also reveal how he responds to people intervening in his interactions with other dogs (verbal corrections, physical restraint, etc.) and to aversive stimuli (water spray, startling noise, compressed air, citronella spray, etc.) should their use be necessary to interrupt inappropriate behavior or a dogfight. On the other hand, a limitation of playgroup observations is that a dog’s motivation to play with other dogs may completely overwhelm his desire to interact with people. Also, some dogs simply don’t enjoy playgroups, and their anxiety in playgroup situations may inhibit their behavior in general or cause defensive behavior not provoked in other contexts. Thus, it is important to assess dogs’ behavior toward people and dogs outside of the playgroup context as well. An on‐leash assessment is particularly important, as this is how most dogs encounter other dogs and people in their everyday lives.

9.2.8 Foster Stay and Adopter Returns

Observations of a dog’s behavior as he transitions from the shelter to a home environment are extremely helpful in matching the dog to a new adopter. A stay in a foster home is presumed to be the most accurate reflection of how the dog will behave in an adoptive home, at least during the time just after adoption. For the purposes of information gathering, an adoption return may also fall into this category.

Many shelters have volunteer foster programs that provide a temporary stay for dogs who need medical or behavioral rehabilitation before being returned to the shelter for adoption. A stay with an experienced foster volunteer can be invaluable for determining if a behaviorally challenged dog is safe to place or can experience a good quality of life in a home. In a review of the records of a large, open‐admission municipal shelter in the United States, Patronek and Crowe (2018) discovered that adult dogs returning from a foster stay were 20 times more likely to be adopted (or transferred to another shelter or rescue group, which implies that the dog met some sort of criteria for adoption suitability) than dogs who had remained in the shelter. Similarly, 85% of adopted dogs who were returned within 30 days were quickly and successfully readopted. While it could be that dogs are better behaved as a result of their experience when they return to the shelter (Hennessy et al. 2006), Patronek and Crowe (2018) surmise that the shelter obtains valuable information about the dogs’ time out of the shelter to facilitate good matching with adopters. Regrettably, they didn’t delve into the process by which the shelter gathers and documents that information.

Foster programs are considered beneficial for promoting adoption, either through increased visibility to potential adopters or by so‐called “foster fails,” when the fosterer ends up adopting the dog. Mohan‐Gibbons et al. (2014) examined a program called Adoption Ambassadors in which the foster volunteer is responsible not only for caring for the dog but also for identifying an adopter and facilitating the transfer. Significantly fewer fostered dogs were subsequently returned by their adopters than dogs who had been adopted directly from the shelter. However, it took longer for dogs in the foster program to be adopted than dogs in the shelter. This could be because fosterers are pickier when it comes to selecting adopters or because dogs in foster homes are not exposed to as many potential adopters as they would be in the shelter. It’s also possible that the longer the dog spends in foster, the better prepared he is for a successful adoption. Without a full understanding of the factors at play, placement in a foster home should be considered carefully. Dogs who are likely to fit into anyone’s home may be best served staying at the shelter, where they will be adopted quickly, whereas dogs who may need more careful matching for a successful adoption could be prioritized for foster homes.

Adoption of dogs who have been fostered may be more successful because of the higher quality of information provided by fosterers. Potential adopters also have the chance to observe the dog in more realistic situations than staff do in the shelter. Nearly one‐quarter of the adopters of foster dogs in the Adoption Ambassadors program said they based their decision to adopt in part on information imparted by the fosterer, compared with only 3% of adopters who said they relied on information from shelter personnel (Mohan‐Gibbons et al. 2014). Gilchrist (2019) examined the agreement between fosterers’ and adopters’ impressions of their dogs and found that fosterers’ responses about their dogs on the C‐BARQ matched much more closely with the subsequent adopters’ responses, even six months post‐adoption, than with those provided by shelter staff.

Coordinating a sustainable foster program is resource‐ and time‐intensive, and not all shelter professionals are convinced that transitioning a dog from the shelter to a foster home and potentially back to the shelter is good for the dog’s well‐being. However, research reveals that getting the dog out of the shelter, even for one to two nights, reduces stress and improves quality of life by promoting uninterrupted rest during the time away from the shelter (Fehringer and Dreschel 2014; Gunter et al. 2019). While the benefits do not continue after the dog returns to the shelter, cortisol levels were no different than they were before the dog left (Gunter et al. 2019), suggesting that the transition back to the shelter is not unduly distressing. Unfortunately, a “sleepover” is unlikely to reveal much useful information about a dog’s behavior, simply because most dogs spend that time catching up on their sleep.

Many shelters transferred their animals into foster homes during the COVID‐19 pandemic and, with all the potential benefits, the hope is that robust foster and sleepover programs will become a sustainable component of care for homeless dogs (Morgan et al. 2021). Dogs experiencing high levels of distress or displaying undesirable behavior in the kennel or those suffering from long lengths of stay may benefit the most from time out of the shelter to reduce their stress and enable them to reveal how they would behave in a home environment. See Box 9.3 for an example survey for collecting behavior information from fosterers (and associated word bank in Box 9.2).

9.2.9 Standardized Behavior Evaluations

As mentioned previously, shelters have long relied on regimented behavior evaluations (sometimes erroneously referred to as “temperament tests” or “personality tests”) as a consistent, practical way to gather information about a dog in a short period of time (Marder 2015; Reid et al. 2004). In an evaluation, the dog is exposed to a battery of tests designed to simulate experiences commonly encountered by a pet dog (Mornement et al. 2009). Some of the experiences are potentially provocative, such as unpleasant handling, and possibly scary, such as being approached by a threatening stranger. Most evaluations try to capture information about the dog’s sociability with people and with dogs, his tolerance of handling, his activity level and playfulness, his propensity for aggression, his reaction to novelty, and his ability to recover from a frightening event. A list of standardized behavior evaluations that have received scientific attention appears in Table 9.1.

Despite their widespread acceptance (Mohan‐Gibbons et al. 2012), there is little scientific evidence to support the use of behavior evaluations as predictive sources of information. A good behavior evaluation ought to fare well on three important criteria: (i) feasibility, (ii) reliability, and (iii) validity (Taylor and Mills 2006). A discussion of each of these criteria follows. For a comprehensive review of the relevant published literature on behavior evaluations, see Patronek et al. (2019).