Lindsay R. Mehrkam There is perhaps no other animal species so characteristically associated with play than the domestic dog. This is true not only of dogs who live in our homes as pets; it is also true when it comes to choosing a companion from the shelter. In our society, play is often synonymous with our view of dogs. For example, when dogs reject play initiation from a potential adopter in an out‐of‐kennel setting, they are less likely to be adopted (Protopopova and Wynne 2014). Shelters routinely share photos and videos of dogs engaging in fun, playful behaviors to promote interest in adoption. Parks are constructed specifically to promote off‐leash play between dogs and their owners, as well as with other dogs. As many owners and professionals know, there is no question that dogs engaging in healthy play with one another is one of the most engaging sights to observe. Not surprisingly, as shelter management practices continue to evolve, playgroups have become a widely popular and useful practice in many shelters. In a 2018 nationwide survey, approximately 83% of responding shelters in the United States reported using playgroups in some way (Shabelansky and Segurson 2018). Playgroups may be defined as organized opportunities for multiple dogs to interact off leash in an area outside of their kennels, usually in a designated fenced yard or indoor space. Beyond this definition, however, playgroups vary widely in both their form and function. It may come as somewhat of a surprise to many, however, that the concept of allowing dogs to engage in off‐leash play with one another in both public outdoor settings (e.g., dog parks) and shelter settings is somewhat controversial. Some argue that dogs do not need to play to have a good quality of life. Even more contentious topics include how to best manage play in dogs, how much and how often play should be a part of a shelter dog’s life, how play contributes to a dog’s quality of life in the shelter, and whether play impacts adoptability. As animal care professionals, we are guided by evidence‐based management techniques, tools, and practices, and playgroups should be no exception. This chapter will first examine the scientific literature on the benefits of play for dogs’ well‐being and provide evidence to inform professionals on both the rationale for incorporating playgroups and how to best do so. Throughout this chapter, a number of operational concepts essential to successful playgroups will be discussed. Terms defined include “consent,” “aversives,” “flooding,” “punishment,” “healthy play,” and “reinforcing.” These are all terms rooted in behavioral science, with specific definitions, and they have important applied value in playgroup contexts. Ask anyone what the function of predatory behavior is, and they will probably respond “to eat,” or what the function of copulation is, and we will hear “to produce offspring.” The answers about the function of play tend to produce a bit more diversity, such as “for fun,” “to expel excess energy,” “to practice,” or “to strengthen social bonds.” However, the purpose of play remains a bit of a mystery to scientists. In addition, because the structure of play appears frivolous and incomplete relative to other more “necessary” behaviors (e.g., aggression, mating, foraging), play hasn’t been the subject of much serious study for scientists in animal behavior and was once considered to be a sort of “behavioral fat” (Muller‐Schwarze et al. 1982), which animals only engage in to expend surplus energy. However, given that play occurs in so many individuals and species, we know it must have a biologically relevant function. For this reason, play has received renewed interest from scientists. It can incur immediate costs to the participants, including an increased risk of injury and predation (Fagen 1981; Harcourt 1991; Burghardt 2005) and loss of energy (Palagi 2007). Thus, it follows that because play has costs, it must also afford benefits (Hinde 1974; Bekoff and Byers 1981; Fagen 1981; Smith 1982). But what are they? Play is difficult to define, consists of a relatively small portion of an individual’s full behavioral repertoire, and does not appear to have a readily observable behavioral function. The apparent absence of proximate causes of play have led scientists to suggest that the primary benefits gained from play are delayed (over the span of a lifetime, developmental stages, or across generations of a species), rather than immediate (accessible in the direct environment and within a developmental stage). Several theories propose explanations for the evolutionary functions behind social play. One of the earliest is surplus resource theory (SRT), which states that animals use the excess resources and energy afforded to them during juvenile periods to explore their environment (Burghardt 2005; Pellegrini et al. 2007). SRT also attempts to specify the conditions necessary for play to develop in different species. Such conditions may include sufficient parental care, long periods of immaturity, and the ability to thermoregulate and engage in and recover from vigorous activity. From this view, play can be observed in animal species with surplus resources or at times of the year when resources are abundant. There is relatively little known about the immediate benefits of play to an animal (Bekoff and Byers 1998; Hall 1998), but continued research provides new clues and theories. Play can afford opportunities for behavioral and cognitive innovation and subsequent practice of newly developed behaviors and strategies (Bateson 2005; Bjorklund and Rosenberg 2005; Bruner 1972; Stamps 1995). It has also been proposed that “play is for practice” (Burghardt 2005); that is, social play affords opportunities for social learning and allows organisms to develop the physical and psychological skills needed to cope with stressful or unexpected situations (Spinka et al. 2001; Dugatkin 2014). Play has now become a behavior of great interest and value to many scientists and practitioners, although there is still not a clear scientific consensus about its benefits. In actuality, play may have multiple functions, depending on the species and the individual. Although play can take different forms, social play has been of particular interest to scientists. It is usually the focus of research on playful species because not only is it often considered an easily recognizable form of play, but it is also generally performed more frequently and for longer periods of time than object play (Burghardt 2005). Social play may be defined as play behaviors directed at conspecifics (Burghardt 2005), although social play between members of different species has also been documented (e.g., humans and dogs; Rooney et al. 2001). It may lead to the forging of long‐lasting social bonds among animals (Carpenter 1934), including primates (Mendoza‐Granados and Sommer 1995), and may serve to provide young males with coalition partners who may be important in their adult life (de Waal 1992). Social play may also promote opportunities to fine‐tune behaviors integral to successful competition, foraging, and mating, all of which have clear survival value for all species. Some have suggested that the function of rough‐and‐tumble play is to assert “dominance” due to “a variety of social and social/physical challenges [that] arise in the course of these activities” (McCune 1998). It has also been suggested that social play may aid in the development of cognitive skills (Bekoff 2000, 2004) by providing opportunities for the “self‐assessment” of developmental progress as compared to conspecifics (Thompson 1996). For example, play fighting in male rats may be beneficial to individuals by testing the propensity of potential competitors to fight (Smith et al. 1999), and courtship play in primates may help overcome female reluctance to mate with males (Pellis and Iwaniuk 1999). Immediate adaptive motor benefits, such as increasing the versatility of motor responses, have also been suggested in primates (Palagi 2007). It is well known that play is most common in larger‐brained vertebrates (Pellis and Iwaniuk 2004; Burghardt 2005). Various forms of play—including social, solitary, and object play—have been documented most widely in chimpanzees (Matsusaka 2004; Palagi et al. 2006), gorillas (Palagi et al. 2007), bonobos (Pan paniscus) (Palagi et al. 2006; Palagi 2008; Palagi and Paoli 2007), rats (Panksepp et al. 1985), domestic dogs (Bekoff 1974; Bauer and Smuts 2007; Ward et al. 2008), and wolves (Cordoni 2009). Many canid species are social, and spending time among conspecifics provides more opportunities for social play (Bekoff 1972; Fagen 1981; Burghardt 2005; Horowitz 2009). Occurrences of social play have been reported frequently in captive wolves (Fox 1970; Cordoni 2009), and it has been reported that captive animals play more than wild animals, presumably because they have the surplus energy required to engage in play behavior. The characteristic playfulness of domestic dogs may support this presumption; dogs afford rich opportunities for researchers to study a variety of species‐typical play behaviors and how the physical and social environment affect play. Dogs are in many ways an ideal model for understanding play both structurally (how it looks) and functionally (why it occurs, why it has evolved, and what it achieves for the individual animal and/or species). Relative to most other species, dogs have been identified as a species with characteristically high levels of interspecific and intraspecific play (Bauer and Smuts 2007; Ward et al. 2008; Cordoni 2009) and furthermore engage in high levels of play not only as juveniles but also long into adulthood. Generally, dogs have been selectively bred to be playful with humans as well as other dogs, as these are often considered desirable traits. In addition, home environments, rich with toys and ample space, often permit and promote play. Both exogenous and endogenous factors can influence play in dogs. The motivation to engage in play, for example, may reflect a breed’s propensity for exhibiting predatory motor patterns/sequences (Coppinger and Coppinger 2001). For example, Mehrkam et al. (2017) found that livestock guarding dogs (LGDs) were less motivated to play with a toy compared to herders and retrievers. Solitary play with a toy involves looking at and following the toy’s movement (e.g., orient, eye, stalk, chase) and possibly picking up the toy (grab‐bite). Herders and retrievers both exhibit intact but truncated predatory sequences, which have been selected for in these breed types so they can successfully perform their respective working roles (Coppinger and Coppinger 2001; Udell et al. 2014). By contrast, breeders select for the inhibition of motor patterns later in the sequence to enable LGDs to successfully perform their working roles. These findings suggest that there may be motivators of both social and solitary play that correspond with breed types and that correspond with breed‐typical motor patterns, as proposed by Coppinger and Coppinger (2001). Additional biological factors may influence dogs’ motivation to engage in play—as well as what motivates them to play with conspecifics. Mehrkam et al. (2017) found that neuter status predicted the occurrence and duration of social play: dyads containing dogs of mixed neuter‐status type (e.g., neutered and intact individuals) engaged in social play at higher levels than did dyads containing dogs of the same neuter status (e.g., two intact males or two intact females). Of the three breed types tested by Mehrkam et al. (2017) (LGDs, herders, and retrievers), no breed was significantly more likely to engage in play overall; environmental factors (i.e., which stimuli were presented prior to play) were a significantly better predictor of play levels than breed. Though this is somewhat contrary to Coppinger’s hypothesis, it is important to acknowledge that play, like many behaviors, is influenced by a combination of breed type, neuter status, and environmental factors. Across breeds, social play was no more likely to be facilitated by nonsocial stimuli (e.g., toys or release from brief confinement/separation) than by social stimuli (e.g., direct human attention); by contrast, solitary play was more likely to be facilitated by nonsocial stimuli than by social stimuli. Social play may thus reflect a motivation or function to “test” or reinforce social bonds between conspecifics when a valued resource (e.g., toy, human attention) is present. Strong motivators of human‐dog play may include toy‐related stimuli regardless of breed type (Mehrkam et al. 2017). Given the presence of human handlers in playgroups, an understanding of human‐dog play is also important to their successful management. The presence of an attentive familiar human caregiver has been repeatedly shown to facilitate social play between familiar dogs (Mehrkam et al. 2017; Mehrkam and Wynne 2021). Passive handlers can still influence social play levels as well, but not nearly as much. There can be different approaches and styles of handlers during playgroups, and this can likewise have differential effects on playgroup success. Following from both anecdotal observations as well as the literature, well‐timed and moderate vocal praise from handlers during appropriate dog‐dog play can further promote play without leading to conflict or aggression. Aside from its theoretical importance in the field of animal behavior, play has practical importance as well. It is well established that play can be an indicator of good welfare in animals (Held and Spinka 2011). For example, play has been associated with reduced cortisol in a variety of species and occurs in the absence of fitness threats (Spinka et al. 2001). Engaging in play may be self‐rewarding to the organism and produce pleasurable emotional experiences or afford other private (i.e., unobservable) benefits (i.e., animals play “for fun,” Held and Spinka 2011; Pellegrini and Smith 2008). It is also associated with positive emotions in animals and in humans as well as increased endogenous opioids in the brain (i.e., increased mental stimulation and reward activation) in animal models (Vanderschuren et al. 1997). Social play is associated with increased oxytocin in owned dogs (Romero et al. 2015), and opportunities to engage in toy play after a training session have also been associated with improved training outcomes (Affenzeller et al. 2017). Although much research is continuing in this area, there is ample evidence to support the notion that various play types may be reliable indicators of good welfare in a range of animal species, including dogs. Given these findings, dogs who have an opportunity to participate in well‐run playgroups may experience much‐needed psychological benefits that may help them to cope with stressful shelter environments. Although animals may reduce play during times of potential conflict or stress among the group (e.g., wolves in Cordoni 2009) or when there are limited resources, it has been reported that play may also occur during these times as a way to regulate stress. For example, Palagi (2007) found that play in primates seems to occur in anticipation of forthcoming stress associated with feeding. In addition, play has been suggested as a potential way to regulate stress associated with competition for owner attention in domestic dogs (Mehrkam et al. 2017; Mehrkam and Wynne 2021). Collectively, these findings offer an important reminder that just because dogs may engage in play doesn’t necessarily mean that they are free from stressful contexts; in contrast, it may be that play opportunities with other dogs are even necessary to regulate or respond to the stress of being in the shelter and that the occurrence of play may indicate successful coping or resiliency. This only further underscores the importance of having playgroups in shelter operations and behavior programs, as well as the resources to support them. The incorporation of well‐integrated playgroups into shelter enrichment programs may improve quality of life for dogs during their stay in the shelter—and also get them into homes more quickly, which is also expected to benefit their psychological well‐being. Adopters often report that play is a desirable behavior (Dogs Playing for Life n.d.; Protopopova and Wynne 2014). In addition, access to a play session with preferred toys before training sessions improved performance in training in pet dogs (Affenzeller et al. 2017). Whether social play with other dogs—rather than toys—also promotes improvement in performance in training is an interesting but currently unexplored question. The benefits of social play in dogs have only recently been examined in the context of environmental enrichment. One way quality of life for shelter dogs may be improved is through enrichment or, more specifically, environmental enrichment. Environmental enrichment may be formally defined as an animal husbandry strategy that seeks to enhance physical and psychological well‐being by providing stimuli that promote species‐typical behaviors, novel sensory stimulation, and behavioral choice (Shepherdson 1998). Today, playgroups are becoming more widely accepted as a social enrichment strategy for shelter dogs, but do all playgroups always meet these goals? To assess scientifically whether any enrichment strategy is effective, it must produce the intended behavioral effects on the animal being exposed to the intervention. While environmental enrichment strategies may have a common definition, they can come in many different forms. Different types of enrichment may include feeding, object‐based or tactile, visual, olfactory, structural, and social components. When implemented, playgroups are typically intended as a social enrichment strategy. By definition, social enrichment involves opportunities to solicit and access interactions with conspecifics, as well as with staff and volunteers. In addition, playgroups can provide increased exploratory opportunities and increased sensory stimulation (e.g., structural, olfactory, or visual enrichment). Remember that access to dogs, for many—but not all—is a primary reinforcer. However, sometimes, food—another primary reinforcer—will compete with a dog’s motivation to engage with conspecifics, thereby disrupting play. Enrichment should also promote behavioral choices and control without overwhelming the animals, which is especially relevant when it comes to playgroups. To ensure playgroups are enriching, the S.P.I.D.E.R. framework for enrichment programs can be applied (Mellen and MacPhee 2001). Originally developed by behavioral scientists at Disney’s Animal Kingdom zoological institutions, which implement robust enrichment programs, the components of this framework apply to shelter playgroups as well and include: Setting Goals, Planning, Implementation, Documentation, Evaluation, and Readjustment (see animalenrichment.org for more information). Social enrichment may be defined as strategies that promote the provision of conspecific interactions (Hoy et al. 2010), although interactions with other species (e.g., humans) have been recently suggested to meet these criteria as well. Increased opportunities for play, species‐typical social interactions, companionship, physical exercise, and socialization with both other dogs and people may occur through successful, well‐run dog‐dog playgroups. In addition, depending on how the playgroup handlers coordinate and manage inter‐dog interactions, playgroups can also afford increased opportunities for training and bonding to handlers and potential adopters. For example, Shabelansky and Segurson (2018) reported that most respondents from US shelters believed playgroups improve knowledge of a dog’s behavior (97%), improve quality of life/welfare (96%), and reduce stress (96%). Other perceived benefits included improvement of staff morale (87%) and reduction in length of shelter stay (70%). Species‐typical behaviors are those we would expect to see all individuals within a species perform in their natural habitat. Certainly, it can be somewhat tricky to define what a species‐typical behavior is for a domestic animal, as their “natural habitat” can be hard to define. Despite this challenge, the scientific literature suggests that inter‐dog play is indeed a species‐typical behavior for domestic dogs. In the absence of highly pressured artificial selection, free‐roaming village dogs engage in play behavior, as do New Guinea singing dogs (Dwyer and Minnegal 2021). Although certainly age, general health, and experiential factors (such as inadequate socialization) can impact an individual dog’s development of play, as a species, dogs do play. Therefore, playgroups as an environmental enrichment strategy do provide opportunities for species‐typical, natural behaviors. Mehrkam et al. (2014) explored whether human interaction from familiar caregivers could be a form of social enrichment for pair‐housed wolves and wolf‐dog crosses in a private sanctuary. The researchers measured levels of conspecific affiliation, including social play, and human‐directed affiliation, as well as general activity levels and stereotypic behavior. The results indicated that human interaction was enriching, evidenced by significant increases in conspecific affiliation (including social play) and human‐directed affiliation, increases in activity levels, and reductions in stereotypic pacing as compared to baseline sessions in which human interaction was not provided. Later, Mehrkam and Wynne (2021) replicated this procedure with pet dogs who lived in the same home and observed similar increases in social play during conditions in which owner attention was available compared to conditions in which the owner was absent or was present but ignoring their dogs. Thus, attention from caregivers, which often includes social and verbal praise and tactile stimulation, may facilitate positive emotional responses in captive canines (Mehrkam et al. 2014; Mehrkam et al. 2017; Mehrkam and Wynne 2021). In addition, play did not escalate to aggression in any of the dyads studied during experimental sessions, despite high‐intensity social play being regularly observed. Therefore, when conducting playgroups, handlers should be encouraged to deliver verbal praise and, when appropriate, petting to facilitate healthy play, especially among dogs who are familiar with one another. It is reasonable to assume that similar effects might be observed with dogs in playgroups who may be relatively unfamiliar with one another and their handlers; however, future research is needed to confirm this assumption. As with any enrichment or behavior modification strategy, there may be medical reasons why playgroups may not be approved or may be limited for an entire shelter’s population, a subset of the population, or an individual animal. Any candidates for playgroups should be selected only following the clearance of veterinary staff. In addition, although inter‐dog play is certainly a species‐typical behavior, one should not assume that all dogs are suited for dog‐dog playgroups or find play with other dogs positively reinforcing. It is important to remember that while some dogs may be reactive to other dogs across time and situations, other dogs in the population may exhibit discomfort or conflict toward other dogs due to behavioral deterioration experienced as a result of being in the shelter environment. Dogs may exhibit increased stereotypic behaviors and increased withdrawal or loss of interest in social contact with people and/or other dogs, as well as increased aggression and reactivity toward other dogs (Protopopova 2016). In some situations, dogs housed in a shelter may appear to be reactive toward other dogs due to the presence of a barrier but may not exhibit reactivity in a playgroup context. Thus, while it is necessary to select dogs carefully, shelters must also be careful not to exclude dogs who may benefit from playgroup based on easily modifiable problems or for reasons that may be simply artifacts of the shelter environment. Research on enrichment preferences in a variety of species demonstrates that although a strategy may indeed be species‐appropriate, individual learning history and experience must also be considered; not all individuals of a species may find the same enrichment strategy highly reinforcing/enriching or highly preferred, if preferred at all (Mehrkam and Dorey 2014; Mehrkam and Dorey 2015). This lesson also applies to playgroups, as a variety of lifetime experiences (e.g., lack of socialization with other dogs, stress, injury or trauma, or other aversive interactions with dogs and/or people) may lead individual dogs to find play with other dogs unenjoyable. It is worth noting that, over time, this could change with proper positive experiences/conditioning, but whether a shelter wants to pursue this will depend on resources and length of stay. Identifying appropriate candidates for playgroups is of great importance to success and will be discussed in more detail later in this chapter. The science of learning and behavior analysis is clear: tools are stimuli. Stimuli can be defined as any environmental change or event that can be perceived by an individual. To beconceptually systematic, it is best to think of these tools in this way, rather than as rewards, reinforcers (positive or negative), punishers (positive or negative), corrections, or interrupters. Scientifically, for stimuli to be reinforcers, they must cause an observed increase in the future likelihood of a behavior. Likewise, for stimuli to be punishers, they must cause an observed decrease in the future likelihood of a behavior. Furthermore, reinforcers and punishers are categorized by whether they were added (i.e., positive) or removed (i.e., negative) to the environment to have their effect on the organism. Therefore, whether we call the tools we use aversives, rewards, reinforcers or punishers, interrupters, or corrections depends on how they affect the future probability of that behavior, no matter how tempting it may be for us to use those labels beforehand or to make the error of assuming that just because we intend to use a tool for a certain reason, the intended effect will follow. Tools that are used often in playgroups include vocal praise, vocal corrections, treats, toys, break and/or pig boards, leashes, collars, harnesses, muzzles, carabiners (to lock or secure gates), water spray bottles, shake cans, compressed air, and air horns. Certainly, other dogs are used in practice as live “tools” to help facilitate play and other social interactions between other dogs. In a nationwide survey of playgroup practices, vocal corrections were the most common (86%), followed by verbal praise (82%), “leash‐on” (leaving the leash attached to the dog) (77%), and “body/social pressure” (71%). Use of control poles, prong or pinch collars, and electronic collars were least reported (Shabelansky and Segurson 2018). These tools can be used as interventions and combined with other techniques, including recall and obedience training, differential reinforcement, and timeouts or planned ignoring by handlers (see Figure 13.1). When determining which tools to use in playgroups, shelter professionals should keep in mind that reliance on aversives can lead to the association of both people and other dogs with unpleasant occurrences in off‐leash contexts, particularly if playgroups are the only contexts in which dogs would interact freely with people or dogs in the shelter environment. In addition, overinvolvement of handler interventions intended to interrupt undesired behaviors can be disruptive to healthy play interactions. This can also be further negatively reinforcing to the handler because intervening can remove the possibility of a potential conflict. Handlers should be cautious when correcting or interrupting dogs during play and ensure that they are only doing so for behaviors that are indicative of negative outcomes. Although preventing conflicts or fights is important for dog welfare, it should not come at the expense of sufficient opportunities for healthy social play. Figure 13.1 Recommendations for least intrusive to most intrusive stimuli used to manage shelter dog playgroups. Source: Courtesy of Mara Velez. Currently, there are two formal models with a specific structure for conducting playgroups, including written protocols and guidelines, published manuals, and associated training opportunities for shelter professionals. The first one established was the Dogs Playing for Life (DPFL) model, which incorporates a wide variety of tools and techniques to efficiently conduct playgroups. The second is the model advocated by the Shelter Playgroup Alliance (SPA) that emphasizes a LIMA‐based approach to playgroups (least intrusive, minimally aversive). Both models emphasize the role of playgroups as enrichment strategies aimed at promoting shelter dog quality of life and offer in‐person workshops in which representatives come to a shelter and help shelter staff and volunteers develop and implement playgroups or refine/consult on existing playgroup programs. Both DPFL and SPA emphasize treating each dog as an individual. Both emphasize a holistic approach to maintaining a good quality of life for shelter dogs, including meeting dogs’ physical, behavioral, and emotional needs. The two models differ substantially on the goals of the playgroup, what constitutes healthy play, the range of tools used to manage dog‐dog play, and the type of intervention methods used. Rather than adopting and strictly following the guidelines of one of these models, the vast majority of shelters run a modified version of one or a combination of the two. In a nationwide US shelter survey study by Shabelansky and Segurson (2018), the majority (83%) of responding organizations use playgroups and with some frequency (71% had more than three per week). However, the average number of dogs in each playgroup was small, indicating that most organizations do not provide the majority of their dog population with the opportunity to play with other dogs on a regular basis. Participating shelters indicated that their playgroups use both large and small groups of dogs and aversives/interrupters as well as treats, toys, and vocal praise. In practice, shelters should adapt their playgroups based on their resources (including staff and volunteers), population of dogs (and needs of individual dogs), and shelter environment. A playgroup may not follow either model with perfect integrity. In addition, a successful and enriching playgroup program is one that will continuously ensure a balance of both the physical and behavioral needs of the dogs. Dogs Playing for Life (DPFL) is a nonprofit organization developed by Aimee Sadler in 1998 at the Southampton Animal Shelter in Southampton, New York. It is the longest‐running and most widely used organized playgroup model in the United States. DPFL describes their program as a “training and behavior modification program for shelter dogs featuring play groups” (DPFL n.d.). DPFL emphasizes maximizing the efficiency of playgroups, using them with the goal of not only providing physical and mental enrichment but with the added goal and function of using playgroups as an operational and behavioral assessment tool to assist with both daily shelter operations and behavioral evaluations of dogs. Typically, relatively large playgroups (six or more dogs at any one time) are recommended (though this somewhat depends on the shelter) and widely used to maximize efficiency and allow for a range of conspecific interactions. In addition, handlers monitor inter‐dog interactions but generally intervene only when a fight is occurring. When intervening, handlers use a wide range of tools to deliver warnings and corrections to dogs and typically do not use toys, treats, or verbal praise often if at all. The DPFL model advocates an “every dog, every day” approach to playgroups and aims to use playgroups as a way to maximize operational efficiency, socialize the dogs, and expend excess energy in a healthy and interactive way.” The DPFL website states that playgroups are the foundation for teaching dogs to “behave in an appealing and attractive way for volunteers and adopters” (DPFL n.d.). One of DPFL’s main goals is to promote and increase the number of play and other species‐typical opportunities for as many dogs in a shelter as possible. Often, playgroups following a DPFL model will be larger (containing more dogs) than what appears to be the typical group size used by shelters (DPFL, personal communication). For example, the DPFL manual (n.d.) states that one of their longest‐running and most successful collaborating shelters averages approximately 60 dogs overall in playgroups within the 2.5 hours playgroups are run every morning. They note that “a handler can attend to many more dogs through offering playgroups, rather than attempting to leash walk each individual dog for a short period of time.” DPFL’s manual is available on their website (www.dogsplayingforlife.com). One main advantage of the DPFL model may be maximizing the efficiency of shelter resources. The DPFL playgroup model also allows for a potentially wider range of species‐typical behaviors, including agnostic and exploratory behaviors. DPFL encourages handlers to allow dogs to “have altercations or interactions that they can work out on their own” and “to learn from each other what is and isn’t appropriate” (DPFL n.d.). The emphasis is placed on monitoring the dogs’ behavior holistically in the context of the entire social interaction, rather than on discouraging specific problematic behaviors or precursors to aggression. The DPFL model has been widely successful in terms of implementation at a wide number of shelters. However, it is important to note that the methods used in the DPFL model may have a greater potential for flooding if dogs in the playgroup find other dogs aversive. One dog could find the presence of other dogs aversive for any number of reasons—fear, territoriality, pain, or guarding of an area of the yard/another play partner/a handler. In addition, since DPFL encourages a relatively more hands‐off approach to managing playgroups, inter‐dog conflict may be more likely to occur, even if that conflict does not often lead to serious fights. For example, it is not encouraged that DPFL handlers intervene in response to behaviors in isolation, even if they are indicative of higher arousal or potential conflict. Such behaviors might include pinning the ears back, holding the tail erect, raising hackles, mounting, growling, or showing teeth, even though many of these behaviors do occur during conflict. An additional disadvantage may be the level of experience/training/skill needed by handlers to safely manage a large group, recognize when it is appropriate to intervene, and know how to do so effectively and humanely. Together, these challenges could result in relatively limited generalizability of this approach across shelters or even across personnel within the same shelter. The use of aversives (e.g., verbal reprimands, spraying) are considered to be easier to immediately implement than delivering treats or praise for acceptable behavior; given the well‐documented power of aversives to suppress behavior, this may lead to a higher likelihood of staff relying on these tools, despite potentially harmful side effects on a dog’s learning history. The DPFL manual provides a summary of potential disadvantages of each training tool used in the DPFL program. There is some preliminary research to support the efficacy of the DPFL model. For example, stress‐related in‐kennel behaviors were observed to decrease in shelter dogs randomly assigned to a DPFL‐model playgroup compared to dogs in the same shelter randomly assigned to an on‐leash walk for the same duration of time per day across four days (Belpedio et al. 2010). This may be somewhat supported by Mehrkam and colleague’s most recent study, as they did not find significant differences in physiological stress measures between DPFL and control dogs overall (Mehrkam et al., unpublished results). The Shelter Playgroup Alliance (SPA) is a nonprofit organization founded by Mara Valez and includes shelter staff and behavior experts who developed a framework that emphasizes “a least intrusive, minimally aversive (LIMA)‐based approach.” The mission of SPA is to provide animal welfare organizations with education and tailorable guidelines and support materials that facilitate healthy inter‐dog interactions. The SPA describes itself as a model that promotes “safe and healthy play, limits the use of aversives, and minimizes the risk of injury and behavioral fallout.” In addition, an important feature of SPA is that it also emphasizes the use of strategies outside of playgroups in their enrichment programs for shelter dogs. In this way, playgroups are viewed as only one part of a shelter enrichment program. SPA’s shelter playgroup guidelines aim to promote choice and control for shelter dogs, as well as adherence to the ethical guidelines adopted by certifying bodies such as the International Association of Animal Behavior Consultants (IAABC 2018) and Certification Council for Professional Dog Trainers. The SPA Inter‐Dog Playgroup Guidelines (2019) recommend much smaller playgroups (two to six dogs per playgroup). The SPA guidelines also state that “LIMA‐based playgroups are devoted to providing dogs with opportunities to consent, physical and mental safety, and avoid the use of aversives.” The LIMA ethical standard was conceptualized to provide competency‐based criteria for animal behavior professionals. Trainers and consultants who follow the LIMA ethical standard often refer to the Humane Hierarchy created by Dr. Susan Friedman. In particular, this involves maintaining low levels of arousal, limiting the number of dogs in playgroup, maximizing handler‐to‐dog ratios, and prescribing alternate enrichment strategies for dogs in addition to or instead of playgroups. The SPA guidelines further state that “LIMA‐based playgroups control the environment in a way that sets the stage for dogs to be successful, while utilizing tools that increase desired behaviors, rather than those that temporarily suppress unwanted behaviors.” The SPA guidelines also emphasize that handlers consider the body language of the dogs throughout the entire interaction to ensure that all dog participants are enjoying inter‐dog play. Specifically, the SPA guidelines recommend groups of two to three dogs and a 2:1 dog‐to‐handler ratio for “low‐risk tolerance” organizations. For “higher‐risk tolerance” organizations, the guidelines recommend four to six dogs per playgroup with a 3:1 (or greater) dog‐to‐handler ratio. SPA does recommend that a 3:1 ratio can be a low‐risk option if the group consists of a stable population whose dog skills are well established. In addition, introducing dogs through protected contact and with leashes on is recommended. The use of food and toys during playgroup is also permitted, so long as the delivery of both food and toys are kept under good handler control during the playgroup and appropriate for the individual dogs in the group (e.g., no guarding history) as well as the density of dogs in the play area. The SPA guidelines acknowledge that “while tools that discourage behaviors may be necessary in some situations, LIMA‐based playgroups aim to minimize their use rather than rely on them.” LIMA‐based playgroups include dogs that find spending time with other dogs positively reinforcing and removes dogs that do not for reasons that might include becoming fearful, overaroused, or stressed or showing behavior indicative of a desire to leave. Like DPFL, SPA also offers workshops and mentorships for shelter professionals wanting to learn more about how to successfully implement their playgroup model at their organization. An advantage of the SPA playgroup model includes aiming to limit stimuli that may be potentially aversive or punishing to an animal. Although it can be argued that experiencing unpleasant consequences is a natural part of any organism’s learning history (Vollmer 2002), there is also ample evidence to show the negative effects such stimuli can have on both dogs and their trainers, especially when applied repeatedly (Casey et al. accepted for publication; Cooper et al. 2014). In addition, the SPA model maximizes stimuli that are intended to be rewarding or reinforcing; importantly, such reward‐based training practices have been shown to be beneficial for dogs, ranging from pet dogs to shelter and working dogs (Blackwell et al. 2008; Cooper et al. 2014; Protopopova and Wynne 2015; Ziv 2017), among other species (Ferguson and Rosales‐Ruiz 2001). The disadvantages of the SPA playgroup model may include reduced efficiency due to fewer dogs per playgroup and greater need for resources (e.g., proper harnesses, treats, treat pouches), increased staff and/or volunteer demand, and reduced opportunities for both off‐leash exercise and a wide range of species‐typical social behaviors, including play. Thus, high‐intake shelters may have a difficult time implementing this model; however, SPA states that their 66‐page guidelines (available at www.shelterdogplay.org) were developed based on behavioral science with input from well‐educated animal behavior and training professionals with experience in open‐ and limited‐admission shelters and that “the guidelines are intended to be flexible to meet the needs of each shelter.” Most recently, a multi‐site study by Mehrkam et al. (unpublished results) found that shelter dog playgroups that utilized treats and vocal praise from handlers and used a LIMA‐based approach resulted in improved in‐kennel behavioral welfare, improved indicators of adoptability, and significantly lower cortisol levels than either a control shelter not using playgroups and two shelter playgroups that used aversives regularly. Furthermore, the researchers found that post‐playgroup salivary cortisol levels were higher only in dogs that participated in a small (two‐to three‐dog) playgroup that used aversives; interestingly, dogs in the control group, large playgroup with aversives, and small playgroup with treats and vocal praise all exhibited decreased salivary cortisol levels after playgroup. This suggests that the increased rate of aversives used per dog in the small playgroup using aversives may have contributed to increased physiological stress in these dogs. With respect to behavioral outcomes, only the small playgroup using rewards significantly improved indicators of behavioral welfare and adoptability in the kennel. In contrast, the control (no playgroup) dogs exhibited significantly decreased behavioral welfare and adoptability. It is important to note that there was substantial within‐shelter variation in cortisol and behavior across individual dogs, suggesting that shelter dogs may indeed respond to playgroups differently based on individual experiences. The results of this study suggest that playgroups—especially reward‐based playgroups such as those recommended by the SPA model—may include reduced inter‐dog conflict, improved behavioral outcomes, and lower stress levels for dogs. Although social play is a natural, species‐typical behavior for most dogs, playgroups in the shelter are surprisingly controversial—both in terms of whether playgroups are effective in promoting shelter dog welfare and also what is the best way to run them. There is also concern about the relative risks of increased inter‐dog aggression and disease transmission. Part of this debate is due to the fact that every shelter is unique, and what may work well in one shelter may be extremely challenging in another. A shelter that has limited volunteers, high turnover, or frequent disease outbreaks, for example, may not feel confident in its ability to effectively and safely run playgroups seven days per week. The results of an online nationwide survey by Maddie’s Fund suggest that playgroups in US shelters are generally safe for dogs and people (Shabelansky and Segurson 2018) and fights, bites, and injuries are not as prevalent as some critics of shelter playgroups might argue. This is also consistent with research in dog parks, although fights in dog parks seem to occur more frequently, likely due to less oversight or intervention observed in dog parks. For example, one study found that less than 5% of all dog‐dog play interactions in two dog parks were intervened in by owners, including interactions that led to serious altercations (Mehrkam et al., unpublished results). Playgroups can provide needed enrichment, social interaction, and exercise for a large number of dogs efficiently, and thus are a particularly attractive tool for high‐dog‐intake shelters. Interaction among dogs brings unavoidable risk, and, accordingly, one should not assume that play comes without potential harm. Inter‐dog play has the propensity to escalate into minor altercations fairly often, although the evidence suggests that severe fights resulting in injury or even death are relatively rare. Shabalensky and Segurson (2018) appear to report that 12% of shelters, however, do report two to five bites to people during the course of a year. It is important to note that such severe outcomes tend to be relatively rare; nonetheless, they should be accepted as a potential outcome and risk of a playgroup—or, truly, any time multiple dogs occupy the same space. Despite the risks, the potential benefits of playgroups are many, and, as described by the Association of Shelter Veterinarians (ASV), they should be considered “when the benefits can be maximized and the risks adequately addressed” (ASV 2019). As is best practice for enrichment programs, “successful” must be operationally defined by identifying behavioral goals. Playgroups generally should not be implemented at a shelter unless the behavioral goals are agreed upon by all team members and include input from key behavior, medical, and operations staff. Some goals might include the following: Certainly, more than one goal may be selected for any one playgroup, and, moreover, the same shelter may aim for a different goal per group of dogs—or even for individual dogs within a group. Once goals are agreed upon, it is time to plan the playgroup program. First, acknowledge that whenever multiple dogs interact in a confined space there are potential risks. A plan for conflict or fights should be planned, prepared for, and articulated among the shelter staff before playgroup begins. Should a fight or incident occur, such events should be clearly communicated verbally within the relevant teams of the shelter and documented in writing, ideally along with a contingency plan. Everyone involved in playgroups, whether staff or volunteer and regardless of role, should be on the same page to ensure consistency. This can be accomplished by ensuring that personnel (i) have reviewed the playgroup protocols and incident reports, (ii) had an opportunity to contribute to the writing and/or to comment on the plan when it was being created, (iii) are held accountable for following the plan, and (iv) have the ability to give feedback on the plan and recommend adjustments after implementation. Active planning includes antecedent arrangements for both the physical and social environment. First and foremost, play areas should be safe, secure (e.g., with high fences), and large enough to promote play among multiple dogs. Equally important is that play areas have places for dogs to avoid interaction, rest, or engage in other species‐typical behaviors (e.g., sniffing, digging) if they choose. Shelters may consider elevated platforms or other structures with ramps and/or stairs, which afford the opportunity for dogs to both climb to an elevated position out of reach of other dogs and hide underneath. Benches, picnic tables, and kiddie pools are also useful options regularly used in these ways by dogs in playgroup. Ideally, play areas should be multiple connected areas rather than a single area so they can be closed off if needed, and they should have separate entrance and exit pens to reduce conflict and promote safe interactions between dogs during greetings, as well as clear pathways for dogs and handlers to avoid unnecessary altercations or confusion among handlers. It is also recommended that play yards contain gates that swing both ways, containers to place toys in that are out of reach to dogs (to reduce resource guarding), carabiners to secure dogs to fences in emergency situations, and interruption tools (air horns, break sticks, spray bottles) in weather‐resistant containers. Play areas should be regularly inspected and well maintained to prevent injuries and escapes and to promote sanitation. There is some debate about whether toys or treats should be in the play area. For example, DPFL recommends against toys or treats “or any other valued resource since they can become a potential distraction or issue.” However, it is possible that treats and other appetitive stimuli can be used safely if kept securely contained in a treat pouch attached to the handler and delivered intentionally and directly to specific dogs at appropriate times. Handlers should not, for example, give treats to dogs when they are mobbing or jumping on them, but can provide treats as a reinforcer for successful recalls to the handler or taking breaks during play. In fact, shelter dogs who participated in playgroups that primarily utilized rewards such as food and vocal praise had significantly lower cortisol levels, lower behavioral indicators of stress, and higher behavioral indicators of adoptability (Mehrkam et al., unpublished results). Resource guarding can also be maintained by a variety of consequences, depending on the individual dog’s history (Mehrkam et al. 2020). For example, dogs may exhibit resource guarding not only to retain access to the guarded item (positive reinforcement), but also to remove the perceived threat via negative reinforcement. (This is, in part, why animals sometimes guard items or areas that no longer contain food, water, or toys.) It is recommended that flat, washable collars without buckles and quick‐release snaps are used during playgroup to prevent entanglement between dogs. Staff should ensure that collars are fitted properly on each dog prior to playgroup. Muzzles (basket and nylon are most recommended) should also be used regularly as a precaution for any dogs who may have exhibited reactivity toward other dogs or handlers but may still be considered potential candidates for playgroup. Six‐foot rope, nylon, or leather leashes, as well as BioThane® (BioThane Coated Webbing Corp., North Ridgeville, OH) longlines (to prevent tangling), are also all recommended. Gentle Leader® headcollars (PetSafe®, Knoxville, TN) are not recommended during playgroups as they can create quick and intense pressure on the dog’s head and neck and present an immediate danger to the dog if the leash gets caught or stuck on an object in the play yard. Other tools include two‐way radios (e.g., for handlers to communicate across a large play yard over the sound of many dogs, to communicate with other staff inside the shelter, or to move dogs in and out of playgroup), an airhorn, compressed air, citronella deterrent spray, shake cans, spray water bottles, stainless steel water bowls used to make a loud, clanging sound, and squeaky toys. Shock or electronic collars should never be used in playgroups; a growing body of research has indicated they are no more effective than reward‐based training and also lead to welfare concerns (e.g., Cooper et al. 2014). In a playgroup, the social environment consists of the handlers and other dogs, and, sometimes, adopters or other community members or dogs walking outside of the play area. Given the various aspects of the playgroup environment, attributing roles is key before entering the play yard, with one group leader per play area. One of the most common recommendations made for best practices in managing dog‐dog play interactions is with respect to size. Size considerations are relevant in two ways: the first consideration is the quantitative size of the playgroup (i.e., number of dogs present), and the second consideration is whether to allow dogs to be in playgroups together based on their physical size. There is no optimal or universally agreed upon number of dogs to include in any one playgroup at a single time. It is estimated that the average playgroup size for most shelters is eight dogs (Shabelansky and Segurson 2018). Johnson et al. (2013) reported that dogs that participated in three‐dog playgroups (n = 24) and dogs that participated in five‐dog playgroups (n = 12) both showed improvements across behavioral measures of in‐kennel arousal and excitability between pre‐ and post‐playgroup observations at the shelter. More specifically, the researchers reported that dogs in the three‐dog playgroup showed improvements across all measures except for position in kennel, where no change was observed. While it should be noted that the authors conducted very brief (approximately 10s) behavioral observations, they also acknowledged that future research on playgroups should be conducted to more comprehensively assess if playgroups do improve in‐kennel behavioral arousal levels resulting in improved adoptability and overall dog well‐being. Inter‐dog play most often occurs between two dogs (Käufer 2014). DPFL does not generally recommend restrictions on size limits of playgroups or handler‐to‐dog ratios. SPA recommends a maximum of four to six dogs per playgroup, with a 2:1 ratio of dogs to handlers, and also recommends ensuring that a minimum of two handlers be present at all times. There is an increased risk of aggression in dog parks when more than two dogs are involved in a play interaction (Mehrkam et al., unpublished results). Interestingly, however, there appears to be no correlation between average size of the typical playgroup and an increased risk of fights or injuries (Shabelansky and Segurson 2018). Mehrkam et al. (unpublished results) found there was no significant overall effect of size of playgroup alone on behavioral welfare indicators. It may actually be the case that playgroups larger than two or three dogs offer more choice and control to individual dogs because they have more potential play partners to choose from (including if one dog doesn’t reciprocate play solicitation or engagement from another dog), so long as space is sufficient for dogs to avoid an interaction or to engage in other exploratory behaviors if they prefer. Overall, the ideal‐sized playgroup is one that matches the space and number of handlers available while considering the composition of dogs present in the playgroup as well as the shelter’s population. In other species, including humans, crowding may increase the likelihood of aggressive interactions and should be avoided. It is recommended that the number of dogs per playgroup should be determined based on (i) the number of experienced, trained handlers available to manage the playgroup, (ii) the physical size of the play space to prevent crowding effects and mobbing, and (iii) the tools and methods used to manage the dogs. According to the ASV, handlers should optimize human and animal safety by limiting the number of dogs in playgroups based on staff/volunteer training, skillset, and shelter resources. In addition, the nature of the dogs themselves (e.g., age, disposition, temperament, physical ability) is essential to consider. A playgroup consisting of older calm dogs or a group of juvenile small, playful dogs may be able to be substantially larger than a group of rowdy medium or large juvenile dogs. Playgroups can be beneficial for dogs of all physical sizes. There is some debate, however, as to whether dogs of different relative sizes should be part of the same playgroup. Healthy dog‐dog play is possible between dogs of different relative sizes and has been observed in both playgroup, dog park, and in‐home settings (Mehrkam et al., unpublished results); however, it is estimated that fewer than half of shelters utilizing playgroups allow dogs of various weights and sizes in one playgroup (Shabelansky and Segurson 2018). One reason for this is that incidents between small and large dogs do occur. When they do, such incidents can pose a high risk of serious injury, especially to the smaller dog. Predatory aggression is a possibility when there is a substantial size difference, and the risk of injury can increase in off‐leash settings. Given that this behavior can result in serious injury or death to smaller dogs, pairing up unfamiliar large and small dogs may not be worth the risk. In addition, dog owners who experienced conflicts in dog parks reported that large dogs sustained physical bites more often than small dogs (J. Berg, personal communication). Clearly, this suggests that large dogs can also be at risk of physical injury in off‐leash settings. These risks should be kept in mind by handlers if they plan to combine dogs of different relative sizes together in playgroups. The SPA and the ASV both recommend matching or pairing dogs based on their size or considering relative size when selecting playgroup participants. Nonetheless, it could be harmful to withhold a potentially enriching interaction between two potentially successful play partners based on size alone. In addition, it has been observed that dog owners do discourage play between dogs of different relative sizes more frequently than play between dogs of the same relative size, even when the probability of play‐induced aggression is not significantly different (Mehrkam and Wynne, unpublished data). Therefore, if handlers do combine small and large dogs together, handlers should be aware that they are not discouraging or interrupting healthy play or other social interactions solely because of this potential size bias. While only 48% of shelters allowed dogs of various weights/sizes in one playgroup, dog size was one of the least commonly reported reasons for exclusion from shelter playgroups, so this may not present a major barrier to their implementation (Shabelansky and Segurson 2018). Nonetheless, it may be a precautionary strategy to hold separate large and small dog playgroups, even if there is truly no empirical difference in risk factor, especially if it makes handlers more confident in their ability to adequately manage playgroups (see Figure 13.2). In addition, by controlling/eliminating the variable of size from perceptions of whether social play is healthy or leading into dangerous territory, handlers may have one fewer aspect to consider in a constantly dynamic scenario. Systematic study of this issue is certainly warranted. Playgroups can be beneficial for dogs of all ages, sizes, breeds, and temperaments. However, it is important for handlers and other shelter staff to evaluate whether each individual dog is willing to participate in playgroup and finds the experience enriching or reinforcing; such evaluations are sometimes referred to loosely as “consent tests,” or, more scientifically, “choice tests” or “preference assessments” (see Box 13.1). Inclusion in playgroups will depend on the individual dog. Dogs should not be excluded due to blanket restrictions, such as breed. Dogs that display on‐leash reactivity or barrier reactivity can still be candidates for playgroups. Those behaviors are not considered accurate predictors of aggression or sociability in an off‐leash scenario such as playgroups (DPFL n.d.), although this is largely based on anecdotal observations. This would be a worthwhile and important claim to verify in future research on playgroups, especially given the increasing attention on the validity of shelter behavior assessments. Figure 13.2 Small (A) and large (B) dog playgroups. Shelter playgroups can vary widely in their form and function. Source: Courtesy of Rachel Cohen‐Maso (A) and Dogs Playing for Life™ (B). Shabelensky and Segurson (2018) gathered useful data on playgroup population metrics. For example, they found that, on average, 50% of a shelter’s dog population participated in playgroups. The percentage of dogs taken out to participate in playgroups was significantly higher in low‐intake shelters versus high‐intake shelters. The most common reasons why dogs were not allowed in playgroup activity were aggression to other dogs (78%) and being post‐surgical (73%), followed by being “in quarantine” (58%), “lack of dogs with matching play styles” (56%), and illness (54%). Hyperactivity, dog size, and breed restriction were least commonly reported as reasons for exclusion. In addition, if a dog has shown offensive aggression to dogs in the past (e.g., ignoring cues of other dogs, pursuing a dog that is actively retreating and/or hiding, seeking out opportunities to aggress when not provoked), they should be eliminated as a playgroup candidate (DPFL n.d.). Dogs who are extremely fearful, do not show signs of progressing socially with repeated playgroup experiences, and/or are deteriorating as a result of the playgroup experience should also be excluded from playgroups. In these cases, such dogs may show clear behavioral indicators of not enjoying the playgroup experience (e.g., hiding throughout the duration of playgroup, cowering, staying in corners, showing reluctance to enter the play yard after repeated exposures, increasing defensive behaviors in response to play). Finally, it is possible that human and personnel factors (e.g., availability of sufficiently trained staff, access to training resources) may also play a role in why the majority of dogs in most shelter populations are not included in playgroup.
13
Play and Playgroups
13.1 Introduction
13.2 The Science of Play
13.2.1 Social Play in Canids
13.2.2 Psychological Benefits of Play
13.2.3 Playgroups as Enrichment
13.3 Tools and Terminology
13.4 Shelter Playgroup Models
13.4.1 Dogs Playing for Life
13.4.2 Shelter Playgroup Alliance
13.5 Canine Playgroups in Shelter Settings
13.5.1 Setting Goals for a Successful Playgroup Program
13.5.2 Arranging the Physical Environment
13.5.3 Arranging the Social Environment
13.5.4 Selecting Playgroup Candidates
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
