11.2.1 Human Social Interaction

Our relationship with dogs is unique. Through a combination of domestication, artificial selection, and socialization, dogs have developed the ability to form bonds with us, and our social influence on them is strong (Morey 1994). When we consider this shared history with dogs, it should not be surprising that in the inherent social isolation of the animal shelter, human‐animal interaction has been one of the most studied enrichment interventions (see Gunter et al. 2021 for a review).

The impact of one‐time, in‐shelter interactions involving 20–30 minutes of petting or play on behavioral measures and cortisol levels has been assessed in a number of studies (Hennessy et al. 1997; Hennessy et al. 1998; Menor‐Campos et al. 2011; Shiverdecker et al. 2013; Dudley et al. 2015). In some of these studies, the sex of the human influenced the impact of the petting, with women having a greater effect on cortisol reduction than men (Hennessy et al. 1997; McGowan et al. 2018). In total, these interventions report reductions in cortisol immediately following the interaction, suggesting that they decrease dogs’ stress. Behaviorally, Shiverdecker and colleagues (2013) found that when dogs were left alone in a novel enclosure, they vocalized significantly more than when they were being petted, played with, or left with a passive stranger.

Willen et al. (2017) also examined the effects of 30‐minute interactions across two successive days, similar to what staff or volunteers would provide in the shelter, to detect possible cumulative benefits. They found that, as demonstrated in prior studies, dogs’ cortisol reduced after interacting with a person, but multiple interactions did not produce additional benefits. While cortisol decreased following 30 minutes of interaction, those reductions did not persist an hour later. Behaviorally, dogs vocalized, panted, and tried to escape the interaction room less and were seen to be wagging their tails more often in these sessions as compared to dogs that were left alone in a novel room or remained in their kennel.

More recently, McGowan et al. (2018) found that just 15 minutes with a person could decrease dogs’ heart rate, increase heart rate variability (HRV), and reduce standing by the dog during the interactions. The authors found that the amount of time the dogs interacted with the volunteer was influential, such that when dogs spent less than 50% of their time in contact with the volunteer, they experienced less of a benefit from the intervention. This suggests that physical contact with the enrichment, in this case the human, is important in decreasing biological markers of stress reactivity. Willen et al. (2017) also found evidence that 15 minutes of petting reduced cortisol, but it only occurred for some dogs, notably those brought to the shelter as strays, with reductions comparable to what was found with 30‐minute interactions. Dogs that were owner relinquished did not experience this effect.

It seems likely that the activity that a person engages in with the dog during these interactions affects the dogs’ in‐kennel behavior. Protopopova et al. (2018) examined the effects of two 15‐minute‐long interventions of reading and exercise that occurred daily for 2 weeks. When dogs received exercise, they moved back and forth in their kennels less often immediately preceding the intervention; but immediately after, dogs’ back and forth movement increased but jumping on the kennel door lessened. For dogs that were read to by a volunteer, undesirable behaviors associated with an increased length of stay were reduced prior to the intervention. Immediately after, dogs were also moving back and forth in their kennels more often but barking and jumping on their kennel doors had decreased.

The results of Protopopova et al. (2018) are mixed in that dogs’ back and forth motion in the kennel increased after both the exercise and reading interventions, a behavior previously shown to be associated with longer lengths of stay (Protopopova et al. 2014). When parsing out the interventions’ positive effects, a greater reduction on door jumping was seen with exercise, but an overall larger decrease in undesired behaviors (including both door jumping and barking) was found when volunteers read to the dogs. As the authors note, more research is needed with longer observations, both before and after the interventions, to better elucidate their effects. In the meantime, shelters may consider individualizing the activity of the interaction based on the behavior of the dog, using calm interactions for dogs that need to reduce their barking, while more active dogs that are jumping on kennel doors and rubbing the walls may experience more benefit from exercise.

Interactions of 15–30 minutes occurring weekly or multiple times a week for several weeks have also been explored, such as those described by Bergamasco et al. (2010), Valsecchi et al. (2007), Normando et al. (2006), and Hennessy et al. (2002). Bergamasco et al. (2010) investigated a 25‐minute intervention in which shelter dogs played with a person and toy, were walked on leash, groomed, trained, and received food, praise, and petting three times a week for eight weeks. Although Bergamasco and colleagues found little effect of the interactions on dogs’ cortisol levels or HRV, dogs’ sociability over the eight‐week intervention improved such that they were more likely to approach the experimenter and initiate interaction in the kennel before leaving on their leashed walk.

Valsecchi et al. (2007) tested an intervention of training, play, and petting every other day for two months and found dogs’ obedience, docility, and sociability improved compared to unenriched dogs. Dogs in both the intervention and control groups showed reductions in cortisol; the authors suggested that the regular sampling and, consequently, interaction with humans may have contributed to the overall cortisol decrease. Similarly, Normando et al. (2006) reported that 15‐minute, weekly interactions with a volunteer over five weeks led to changes in dogs’ social behavior. Dogs receiving the intervention were out of sight less and wagged their tails more often than non‐intervention dogs when the experimenter was present, and these effects persisted after the intervention ceased. Thus when considering the impacts of these interventions, it is worth noting that dogs’ interest in people increases as they spend more time with us, which may aid in their likelihood of adoption and ultimate welfare of leaving the shelter and living in a home.

While interaction durations of 15 and 30 minutes have received the most attention in the literature, longer one‐time interactions have also been explored. Coppola, Grandin, et al. (2006) explored the impact of a 45‐minute interaction of petting, play, and grooming on dogs’ cortisol levels immediately following the interaction as well as one and two days and one week later. Interestingly, Coppola and colleagues found that dogs in the contact group did not differ two hours following the interaction with the experimenter but did so when measured the following day, with no other differences observed at later time points. This study’s finding of an effect on the day following human interaction remains the only evidence to date of a delayed reduction in stress for shelter dogs, suggesting the need for additional studies to determine if longer in‐shelter interactions confer distinctive advantages.

Interventions in which dogs leave the shelter for few days or weeks may provide an increased benefit to dogs when compared to in‐shelter interactions (Hennessy et al. 2006; Fehringer 2014; and Gunter et al. 2019). This is likely due to the prolonged nature of the interventions as well as the environment that they occur within. Hennessy et al. (2006) studied dogs fostered in a three‐week prison program, during which they lived with and received obedience training and social interaction by incarcerated handlers. Post‐intervention, the dogs more readily responded to cues such as sit, down, and walk. They also jumped on strangers and barked less, but yawned more, in a novel situation post‐intervention than dogs that had remained at the shelter. While cortisol was relatively unaffected for both groups, the effects on behavior are cautiously optimistic evidence for the use of such programs for longer‐stay dogs or those needing additional socialization.

Time in a home might confer even greater benefits to dogs awaiting adoption. Fehringer (2014) reported that placement in a foster home resulted in lower cortisol compared to in‐shelter levels, and that the dogs’ cortisol steadily declined over the first three days in foster care. In a recent study by Gunter et al. (2019), dogs that stayed for one or two nights in a volunteer’s home were found to have lower cortisol than dogs that remained in the shelter; and while dogs’ cortisol increased upon return, it was no higher than baseline levels in the shelter before fostering. Additionally, dogs’ longest bouts of uninterrupted rest were highest while in the home but remained longer upon return to the shelter than before fostering. In all, these studies provide compelling evidence for the use of stays out of the shelter as an important enrichment tool with measurable benefits.

Limiting visitors in kenneling areas may also improve dogs’ proximate welfare by reducing the unpredictability of human traffic, improving dogs’ ability to control environmental contingencies; however, the impact on their ultimate welfare, such as reduced adoption visibility and potentially longer lengths of stay, should also be considered. Wells and Hepper (2000) found that during days with more visitors to the shelter, dogs spent more time at the front of their kennels, standing and barking. Hewison et al. (2014) prohibited potential adopter access during afternoons when the animal shelter was otherwise open for viewing dogs. They found that noise levels were over 12 dB lower during the intervention as compared to pre‐intervention levels with positive changes in behavior as well. Sedentary behavior increased in both the mornings before opening and afternoons during the intervention, locomotor activity also decreased, and the frequency and duration of repetitive behaviors were reduced in the afternoon—suggesting that scheduled breaks during the day from the arousing bustle of people could afford dogs an opportunity to relax and recharge without leaving the shelter.

11.2.2 Canine Social Interaction

While not as well‐studied as interactions with people, researchers have examined the effects of social contact with other dogs, either through housing manipulations (Mertens and Unshelm 1996; Wells and Hepper 1998; Dalla Villa et al. 2013; Walker et al. 2014) or interactions out of the kennel (Belpedio et al. 2010; Flower 2016; L. Gunter et al., work in preparation).

Mertens and Unshelm (1996) measured behavioral differences between individually and group‐housed dogs and found that dogs housed together were less noisy and better behaved. In that study, dogs that were singly housed barked, whimpered, and howled more than dogs that were group‐housed. The researchers observed that solo dogs were more likely to display aggressive behavior at the kennel front toward other dogs. While conflicts did arise when dogs were housed in groups (which were as large as 30 dogs), of the 211 dogs participating over the three‐month study period, only five incidents occurred, resulting in what the researchers called “light wounds.” Of particular interest, 10% of dogs that were singly housed during the study displayed stereotypies, while no dogs in the group‐housed condition did.

Because of the need for a human presence in the group‐housing condition to deescalate potential conflicts, two staff members monitored the group‐housed dogs (with dogs kenneled singly at night and during feeding). Perhaps as a result of this extra human presence, dogs in this condition more often approached an unknown observer than did singly housed dogs during a behavioral test. Follow‐up research is needed to address limitations of this study, including potential shelter and population differences (this study took place at two different facilities). These findings suggest that group‐housing in which dogs as well as people are present could be helpful in improving sociability.

While group‐housing of the size tested by Mertens and Unshelm (1996) may not be feasible for most shelters due to space, canine sociability, or staff resources, pair housing may be a viable alternative. Dalla Villa and colleagues (2013) studied the effects of group (four to five dogs in an enclosure) versus pair housing on long‐stay shelter dogs and found that dogs’ behavior improved under pair‐housing conditions. Dogs’ trotting, walking, and standing decreased in pair housing, while more lying down occurred, suggesting that dogs were spending more time resting when living together in pairs. While social behavior occurred rarely in either type of housing, agonistic behavior (i.e., defensive or threatening social behavior) was not observed between paired dogs, while such behavior made up nearly one‐fifth of social interactions of group‐housed dogs.

Walker et al. (2014) examined the impacts of removing one dog from a co‐housed pair after the pair had been living together in the shelter for several weeks. After separation, the remaining dogs’ running, changes in posture, and stretching all increased, suggestive of more activity and possible restlessness. Other behaviors that are considered stereotypic in nature and indicative of decreased welfare, such as circling and figures of eight, also increased. Many of these behaviors did begin to decline over the six‐day post‐separation period, although play (now without a conspecific present) remained reduced. While no changes in cortisol were found, S‐IgA levels were elevated after dogs were separated from each other. Interpreting the impacts of this short‐term separation on dogs’ S‐IgA is based on just a few studies, but the elevated levels do suggest that the acute stress of separation negatively impacted dogs’ immune function. Additionally, trends for longer latencies to the middle and near negative positions in a cognitive bias test, though not significant, also support a finding that separation had a negative impact on dogs’ underlying affective state. In agreement with observations by Mertens and Unshelm (1996) and Dalla Villa et al. (2013), Walker et al. (2014) found that co‐housed dogs spent little time behaving agonistically (0.1%) and instead were much more affiliative (3.2%).

It is difficult to determine whether the simple presence of other dogs is, in fact, canine social interaction or merely visual stimulation, and whether such a presence is enriching to dogs or a detriment to their welfare. Wells and Hepper (1998) tested two housing conditions for dogs: one in which the dogs were able to view other dogs in directly opposing kennels versus dogs that looked out to an empty kennel row. Wells and Hepper found that dogs more often were at the front of their kennels when they were able to see other dogs and more often at the back of the kennel when they could not. No differences were observed in the dogs’ behavior (i.e., standing, sitting, resting), including vocalizing.

Certainly, dogs being at the front of their kennels is helpful for adopter viewing and can lead to shorter lengths of stay (Protopopova et al. 2014). However, single housing is sometimes associated with more reactivity at the front of the kennel (Mertens and Unshelm 1996). Thus, housing dogs singly with nearby conspecific visual contact but no social interaction may have more complex effects on welfare than is currently understood, and further empirical exploration of this commonly used shelter housing arrangement is needed to inform best practices.

It is possible that social interactions outside of the kennel may be more enriching than simple visual contact. Belpedio et al. (2010) examined the effects of 30 minutes of off‐leash, canine social interaction as compared to dogs remaining in their kennels. Saliva for cortisol analysis was collected each morning as a baseline as well as 30 minutes and 3 hours post‐interaction, and dogs’ in‐kennel behavior was recorded 1 minute prior to saliva collection. The authors detected no differences in cortisol levels, but, collectively, stress‐related behaviors, which included jumping, barking, and whining, occurred more frequently with dogs in the kenneled group during the study. It should be noted, though, that dogs that received social interaction did lick, pant, pace, spin and yawn more than dogs left in their kennels, suggesting a mixed effect, if any, of this particular intervention on dogs’ overall welfare.

Flower (2016) investigated the impact of one‐time, off‐leash canine interactions on dogs’ performance on a behavioral assessment with more promising results. When dogs’ behavior was assessed with an unfamiliar dog on‐leash, dogs that had interacted with other dogs prior to the assessment displayed more playful, submissive behavior toward the unknown dog than dogs that had not had the interaction experience. While these results are preliminary, they suggest that this type of brief social contact for dogs could be beneficial in supporting appropriate on‐leash behavior when seeing other dogs in the shelter.

The benefits of canine social interaction may extend beyond positive changes in behavior. The effect of differences in housing and social contact with other dogs on cortisol and S‐IgA levels has been examined (L. Gunter et al., work in preparation). Traditional kennel housing was compared to a novel housing design with smaller, glass‐fronted enclosures. Two social conditions were compared within each housing type: a condition with no social contact with other dogs and a condition with three 15‐minute conspecific sessions a day. Dogs experienced each of the four experimental conditions for three days. (This design provided a true control for social interaction in a given housing type.)

The authors found that dogs’ S‐IgA was higher when living in the novel‐designed kennels (irrespective of social condition), and lower when receiving daily canine social contact regardless of housing, with no interaction effect between housing and social condition. (It is worth noting that when videos of the social contact sessions were behaviorally coded, dogs spent more time with people than the other dogs or the environment during these sessions.)

While no main effects of housing or social contact were found with dogs’ cortisol values, cortisol was highest when dogs were living in novel housing and receiving no social contact and were lowest when living in traditional kennels and receiving no social contact. While these results provide evidence that dogs’ contact with other dogs can influence their stress, immune function, and ultimately, welfare, more research is needed to better understand how characteristics of the social contact as well as the level of activity engaged in during these sessions can influence the intervention’s impact.

11.2.3 Object Enrichment

When considering placing objects within the kennel, it is useful to revisit our definition of enrichment. Beds, chews, balls, ropes, and soft toys have varying functional value to dogs, so we must consider the species as well as the individual’s history and preferences in determining how relevant an object will likely be. It is not surprising, then, that investigations into object enrichment for shelter‐housed dogs have been met with mixed success (Wells and Hepper 2000; Wells 2004; Pullen et al. 2010; Kiddie et al. 2017).

Wells and Hepper (2000) explored the impacts of a bed and a suspended Nylabone® (Nylabone Products, Neptune City, NJ) chew, both placed at the front of the kennel on dogs’ behavior. They found that the bed at the front of the kennel caused the dogs to spend more time there, whereas the chew did not; however, dogs’ bed usage decreased when the bed was placed in the front as compared to its typical position in the back of the kennel. While dogs initially sniffed the Nylabone, fewer than 20% of the dogs were seen chewing, pawing, or tugging at it. These findings suggest that while a bed is of interest to shelter dogs, where it resides impacts use.

Subsequently, Wells (2004) provided dogs a variety of objects: squeaky and non‐squeaky balls, Nylabone chews, ropes, and Boomer Balls® (Company of Animals, Broomfield, CO). In a design similar to typical shelter practices, objects were provided for longer durations, six days, with dogs’ location in the kennel and behavior recorded. Overall, dogs rarely interacted with the objects (<10% of all observations) with a reduction over time, suggesting a possible habituation effect. Dogs did show a preference for the Nylabone over the other toys by spending the most time with it; and dogs moved more and stood less when provided the Nylabone, squeaky ball, and non‐squeaky ball. As the author recommends, if object usage is limited and habituation likely, rotation is recommended for the highest probability of engagement.

Pullen et al. (2010) explored two categories of objects: (i) robust toys, including a Boomer Ball, rope, nylon and rubber tug, and rubber toy; and (ii) toys that are more destructible, such as a vinyl bone, soft and plush toys with squeakers, and a non‐squeaky tennis ball. Robust toys were concurrently presented in combinations of hanging and on the kennel floor (trial 1), while the destructible toys were only presented alongside the robust toys, all on the floor (trial 2). Only 35% of shelter dogs interacted with the robust toys during the robust‐only trial, spending on average just two minutes with them. Dogs that contacted the robust toys, spent more time interacting with them on the floor (vs. hanging) with a shorter latency to interact. When the toys were hanging, however, dogs interacted longest with the rope toy. Conversely, when shelter dogs were given the option of soft, destructible toys in trial 2, more than three‐quarters of dogs spent 25% of the 15‐minute session interacting with them, mainly ignoring the robust selections. They also interacted longer with the softer toys, spending the most time with the squeaky bone, soft, and plush toys, and the least time with the tennis ball.

Kiddie et al. (2017) investigated three types of low‐cost interventions to improve shelter dog behavior: partitions blocking visual contact of adjacent dogs, whole coconuts (to potentially play with, chew, or break open), or cardboard beds as possible relief from the plastic mesh bottom of a crate. Dogs were observed for two 30‐minute sessions on non‐consecutive days in the baseline, intervention, and post‐intervention periods with each type of enrichment. While no behavioral differences were found between treatments, dogs were observed to lie down, sit, and yawn less with any enrichment (as compared to baseline and post‐intervention). Of the 36 dogs in the study, all but 1 dog destroyed the cardboard bed—treating it as a chewable object rather than potential bedding. Kiddie et al. (2017) found that dogs interacted with the coconut and cardboard less over time, which could be related to their eventual deconstructed states. However, it is also probable that, as Wells (2004) found, novelty plays a role in object enrichment.

11.2.4 Auditory Enrichment

Several studies have explored the use of auditory enrichment to change dogs’ behavior and reduce noise in shelter kennels (Wells et al. 2002; Kogan et al. 2012; Bowman et al. 2015; Brayley and Montrose 2016; Bowman et al. 2017), focusing on certain genres of music or types of sound, often those that are perceived as pleasant and mood‐enhancing to humans (Rickard et al. 2005).

Wells et al. (2002) investigated the impacts of classical, heavy metal, and pop music; human conversation on the radio; and a control condition of no auditory enrichment on the dogs’ behavior. Each was played for four hours with dogs experiencing all music types. Dogs spent more time resting, less time standing, and more time being quiet in their kennels when classical music was played as compared to other genres of music, human conversation, or no auditory enrichment. Conversely, with heavy metal, dogs were observed barking more frequently compared to any other condition. Similarly, Kogan et al. (2012) exposed dogs to 45 minutes a day of songs from two of the same musical genres: classical and heavy metal along with a track from “Through a Dog’s Ear” by Leeds and Wagner (2008), wherein classical piano music is simplified to create a more soothing rendition. Along with a control condition with no added sound, auditory stimulation was provided to the dogs three times a week for four months. As seen by Wells et al. (2002), classical music led to the most time spent resting. Dogs vocalized more often when no music was provided and were less vocal when classical music was played, but differences were observed between classical selections. Both Wells et al. (2002) and Kogan et al. (2012) observed detrimental behavioral effects with heavy metal music, including increased body shaking, barking, and less time resting.

One consistent limitation of the aforementioned studies is the duration of the auditory intervention. Bowman et al. (2015) addressed this concern when they tested the effects of a classical music compilation compared to a no‐additional‐sound control. Both conditions lasted for six‐and‐a‐half hours a day over seven days with dogs’ in‐kennel behavior observed twice daily for one‐and‐a‐half hours, first in the morning and then again in the afternoon.

When classical music was played dogs spent more time sitting and laying down, with less time standing and vocalizing. While no changes in cortisol were found with the addition of music, changes in heart rate and HRV, which indicate stress reduction, were identified. Bowman et al. (2015) also investigated whether dogs habituate to auditory stimulation, another point of interest to shelters. They found that the behavioral and HRV effects of classical music began to diminish in as quickly as one day, suggesting that while classical music may be effective in improving welfare, rotating the selections that dogs are exposed to each day is important in maintaining those benefits over their stay.

In a follow‐up study, Bowman et al. (2017) investigated four other genres of music in addition to classical music, using compilations of pop, soft rock, reggae, and Motown. Each of the five genres was played once for six hours with a no music (control) condition tested before and after. Regardless of genre, dogs spent more time lying down and less time standing. Barking was not impacted by any genre, but dogs were 142 times more likely to bark after the music stopped, whichever genre it was. The succession of genres in this study more closely resembles a five‐day intervention of mixed genres without the inclusion of any potentially stimulating music (i.e., heavy metal)—and as such, the more meaningful comparison between music and silent control here suggests that varied, auditory enrichment is preferential to shelter noise unabated.

Brayley and Montrose (2016) tested many of the same genres in previous studies, but with the novel addition of an audiobook intervention (a more systematic version of Wells et al.’s [2002] inclusion of conversation on the radio). In two‐hour presentations, dogs experienced Beethoven, 80s pop music, “Through a Dog’s Ear,” a reading of the The Lion, the Witch, and the Wardrobe, and a control condition of no additional sound. When the audiobook was played, dogs spent more time resting and less time walking, sitting, or standing compared to all other conditions. Barking and other vocalizations, such as howling, growling, and whining, also occurred least often with the audiobook, but some differences between the conditions were observed.

Considering that regular conversation tested by Wells et al. (2002) demonstrated no impact on dogs’ behavior, the consistent effect of the audiobook in this study is curious. The authors suggest that the professional delivery and tempo of the book’s narration may have led to greater attending by the dogs than just casual conversation.

11.2.5 Olfactory Enrichment

Many of the reasons to explore auditory enrichment in the shelter, likely apply to olfactory interventions as well. One key feature to these interventions, however, is the species‐specific relevance of olfaction in the daily lives of dogs (Nielsen et al. 2015). Over the past decade and a half, researchers have explored the impacts of odors and pheromones on the behavior and physiology of shelter dogs (Graham et al. 2005a; Tod et al. 2005; Binks et al. 2018; Hermiston et al. 2018; Uccheddu et al. 2018; Haverbeke et al. 2019).

Graham et al. (2005a) investigated lavender, chamomile, rosemary, peppermint along with a no‐odor control, each diffused in front of and behind the dogs’ kennels for four hours a day over five consecutive days. Exposure to both lavender and chamomile led to increases in dogs’ resting and decreases in movement and vocalization. Conversely when rosemary and peppermint were diffused, resting decreased while dogs’ standing, moving, and vocalizing occurred more often.

Binks et al. (2018) evaluated cloths scented with ginger, coconut, vanilla, and valerian placed in the dogs’ kennels along with two control conditions: a non‐scented cloth and no cloth present. For each condition, dogs’ behavior was recorded for two hours a day over three successive days. With all odors, dogs vocalized and moved less and rested more compared to both controls, demonstrating a positive effect of these olfactory interventions. Moreover, dogs reclined with their eyes closed most often with coconut‐ and ginger‐scented cloths compared to the no‐cloth control. However, it is worth noting that the non‐scented cloth also reduced dogs’ vocalizing and movement and increased resting behavior more so than having no cloth in the kennel, suggesting more investigation may be needed to parse out the novel effect of the cloth and the odors themselves. (Use caution when implementing olfactory interventions in the shelter that involve essential oils; see resources from ASPCA’s Animal Poison Control Center at www.ASPCApro.org.)

Pheromones differ from odors in that they are species‐specific chemosignals that affect the behavior of conspecifics. Dog‐appeasing pheromone (DAP) is a synthetic version of a pheromone produced by lactating female dogs shortly after giving birth and is processed through the canine vomeronasal organ, a specialized organ for detecting non‐volatile chemosignals that is part of many species’ olfactory systems. Tod et al. (2005) compared the behavioral effects of a seven‐day diffuser treatment of DAP versus placebo on dogs living in separate kennel blocks at a shelter. They found that when an unknown person walked past the dogs’ kennels on the final day of treatment, the average barking amplitude (loudness) of the DAP group was lower than that of the placebo; however, the peak barking amplitude did not differ, indicating that barking in both groups had similar peaks in their loudness. Likewise, when Hermiston et al. (2018) sprayed DAP directly in dogs’ kennels 30 minutes prior to an unknown dog walking past the dogs’ kennels, the kennel block of the DAP‐treated group was more than 6 dBs lower than dogs’ kennels that were untreated, a more than 30% reduction in volume. However, barking frequency and behavior did not differ between groups.

11.2.6 Visual Stimulation

To our knowledge, only one study has explored the effects of visual sensory stimulation, namely, television monitors, as a form of canine enrichment in the shelter. Graham et al. (2005b) investigated the impact of monitors placed at the front of dogs’ kennels, with video of other dogs, unfamiliar animals, and humans as well as a blank screen, on the dogs’ behavior. Each intervention type lasted four hours a day for five days. Regardless of image type, dogs’ vocalizing and movement decreased with more time spent at the front of their kennel. Similar to findings on toy engagement (Wells 2004), dogs spent only 10% of their time looking at the monitors as compared to controls, and their interest waned over time, suggesting the more species‐specific, interactive enrichment previously described may be preferred by dogs and provide greater welfare benefits.

11.3.1 What to Measure

In terms of assessing enrichment in the shelter, there are two main classes of behavior to measure. The first is the dog’s direct engagement with the enrichment. A variety of measures can be used to assess direct engagement, each answering different questions about the intervention (see Box 11.1). The other class of behavior is whether the enrichment impacts important behaviors for that individual animal, such as producing more resting behavior or a reduction in barking.

Determining if, how long, and in what way dogs engage with the enrichment is essential for determining, at a preliminary level, if that enrichment should be continued or not: is the dog engaging and in a desirable fashion? Or is the dog not engaging? If engagement is confirmed but the enrichment does not bring about the desired behavior change, one can assess if the lack of effect was due to it not impacting the behavior at all or just not to the degree desired.

Of course, one of the main goals in providing enrichment is to improve the animals’ welfare and assessing the impact on dogs’ behavior is critical. For example, consider a 15‐minute front‐of‐kennel treating program, where volunteers toss hot dogs when passing by the dogs’ kennels to create a positive association with visitors. To assess intervention efficacy, the shelter could measure the frequency or duration of dogs lunging or barking at the front of their kennels. A shelter that has many shy, fearful dogs might implement a similar program but measure behavior change differently. In this case, staff could measure dogs’ frequency of approach behavior when they pass by kennels during morning rounds.