The Social Skills and Attachment to Dogs of Children with Autism Spectrum Disorder
- First Online:
- Cite this article as:
- Carlisle, G.K. J Autism Dev Disord (2015) 45: 1137. doi:10.1007/s10803-014-2267-7
- 2.6k Views
Children with Autism Spectrum Disorder (ASD) have deficits in social skills, and interaction with service dogs has been associated with increased social skills for children with ASD. In this telephone survey of 70 parents of children with ASD, children owning dogs had greater Mean scores for social skills, using the Social Skills Improvement System Rating Scale, while those with some type of pet (not excluding dogs) had significantly greater skills for subscale item “assertion”. Parents described their children as attached to their dogs. Children owning dogs completed the Companion Animal Bonding Scale, and reported strong bonding with dogs. These findings suggest children with ASD may bond with their dogs, and pet ownership may be associated with increased social skills.
The field of study in human–animal interaction explores the bond between humans and animals. In a poll of adults (n = 3,014), when asked about relationships they have that are close, 94 % included dogs, 87 % included mothers, 84 % included cats and 74 % included fathers (Taylor et al. 2006). Eighty-five percent of dog owners are considered by their owners to be a member of the family. The benefits of human–animal interaction (HAI) are widely reported in the literature for typically developing children and adults. These benefits include companionship (Cohen 2002), unconditional love (McNicholas and Collis 2000; Melson 2001), stress relief (Hansen et al. 1999) and the social catalyst effect of facilitating interaction among humans (Triebenbacher 1998). Social interaction research in the field of human–animal interaction may be of unique importance for the population of children with Autism Spectrum Disorder (ASD), given one of the core challenges in ASD with social communication (American Psychiatric Association 2013).
The theoretical framework for human–animal interaction is supported by Biophilia, which proposes that humans are naturally drawn to and interested in nature (Wilson 1984). Through evolution, humans have developed an avoidance of aggressive animals and an attachment to friendly animals for a calming effect (Fine 2010). In an examination of this bond, investigators have sought to measure the effects of the strength of this attachment. In one study, time was identified as the predicting factor for the bond of people with their pets, using the Companion Animal Bonding Scale (CABS) (Cohen 2002). The more time spent in interaction, the stronger the bond. A study of typically developing children found that those with a greater attachment to their pets demonstrated greater pro-social behaviors, when compared with children reporting a weaker attachment (Vidović et al. 1999). In addition to this, typically developing children who spent increased time interacting with their pets and caring for them, were found to be more attached (Melson et al. 1991). The interest in animals by typically developing children and the bonding occurring during their human–animal interaction has been demonstrated.
Throughout the developmental phases of typically developing children, animals appear to be influential, rather than simply present in a passive manner (Melson 2001). Typically developing children have been reported to prioritize their relationships with their pets, identifying them as ones with whom they can share their feelings (Bryant 1990; Melson and Swarz 1994). In a study of typically developing children ages preschool through grade five, children described their pets as a source of emotional support and indicated their pets acted as a social lubricant by aiding their interaction with others (Triebenbacher 1998). This animal companionship also appears to provide other benefits. The presence of a companion animal has been associated with greater empathy and social confidence for typically developing children with pets, when compared with those without pets (Daly and Morton 2006; Endenburg and van Lith 2011). Due to the vulnerability reported in typically developing children during transitional phases, the social support provided by pets, through the actions of being a listening companion and confidant, pets may prove valuable during these challenging times. Less is known about the interaction of children with ASD and animals; however, a literature review by O’Haire (2013) found positive benefits in studies of children with ASD interacting with trained dogs, horses and other animals. The benefits for children with ASD identified in this review were increases in social interaction and communication, and decreases in stress and problem behaviors. For example, in 9 out of 14 studies increased social interaction was reported and findings in five studies included increased communication. These differences were noted when children with ASD interacted with therapists and peers in the presence of the animals and/or with the inclusion of the animals in therapy.
Nimer and Lundahl (2007) found that in animal-assisted activities, dogs were the animals most consistently identifiable in the literature with benefits. Studies utilizing dogs in therapeutic interventions have included those with special training, such as the dogs employed as service dogs with individuals and those used in Animal Assisted Therapy (AAT). Children with ASD living with service dogs have been found to demonstrate greater social reciprocity than children without service dogs (Wild 2012). In another study of service dogs placed with children with ASD, parents reported their children had increased happiness, decreased anxiety and fewer emotional outbursts, after arrival of the service dogs (Burrows et al. 2008). In a case study of a 12 year old boy with ASD working with a therapist, the child demonstrated increased positive social behaviors and decreased negative behaviors in the presence of the dog, compared to working with the therapist alone (Silva et al. 2011). A review of studies investigating trained dogs interacting with children with ASD found that children had increased verbal responses in the presence of the dogs (Berry et al. 2013). These positive trends toward the benefits of human–animal interaction for children with ASD through trained animals have led to the rationale for further examination of human–animal interaction between children with ASD and pets.
One study examined the association between pets and social skills of children with ASD using the Autism Diagnostic Interview-Revised (Grandgeorge et al. 2012). The pets consisted of dogs, cats, one rabbit and one hamster. In that study, two subset groupings of children, all having ASD, were compared. In the first subset, one group of children (n = 8) who had always lived with pets in the home had no difference in social skills, when compared with a matched control group of children (n = 8) who never lived with pets (Grandgeorge et al. 2012). In the second subset, a group of children (n = 12) who acquired a pet between the children’s ages of 4–5 years were compared with a matched control group of children (n = 12) who never lived with pets. In the second subset, the children who acquired new pets had significantly fewer social skill deficits for pro-social behaviors of “offering to share” and “offering comfort,” when compared with children not living with pets. Children in this study had a Mean Intelligence Quotient (IQ) of 42.
In the general population, not all people like pets (Kidd and Kidd 1987) so it may be expected that not all children with ASD will desire interaction with pets, nor will it be appropriate for all children. Some children with ASD may dislike animals or have allergies to specific types of animals. While dogs interacting with children with ASD during AAT are under the supervision of the adult initiating the therapy, dogs in the home may not be supervised 100 % of the time. More than 4.5 million dog bites occur each year in the United States and children are the most common victims (Gilchrist et al. 2008). Safety risks of dog bites in children with ASD have not been reported in the literature and are unknown. In other cases, the safety of the animal may be a concern. One study found children with ASD were more likely to have aggressive behaviors than children with intellectual and other developmental disabilities (Farmer and Aman 2011). There was no literature found identifying the safety risks of dogs living as pets in the homes of children with ASD; however, one study (n = 70) reported that 13 % of parents declined ownership of a dog due to their children’s behavior and the parents’ perception of safety concerns for a dog (Carlisle 2014). Welfare concerns for service dogs living in the homes of children with ASD have been reported as lack of rest and recreation, as well as accidental maltreatment of the dogs by the children (Burrows et al. 2008). In the heterogeneous group of children with ASD, the unique characteristics of each child may pose considerations for their interactions with animals.
A core diagnostic behavior of children with ASD includes deficits with social skills. Social skill benefits have been identified for adults and typically developing children interacting with pet dogs, while benefits for children with ASD have been associated with trained dogs. The aim of the present study was to determine if dogs in the home could be associated with benefits for children with ASD through increased social skills and bonding with the dog. The study also aimed to identify whether ownership of some type of pet (not excluding dogs) could be associated with increased social skills for children with ASD. Further, in examining the attachment of children with ASD and their dogs, the goal was also to identify whether or not the strength of the attachment between the children and their dogs would be associated with any differences in social skills.
Potential participants meeting the enrollment criteria for the present study were drawn from the data-base list of a Mid-western autism diagnostic and treatment center. These individuals were offered an opportunity to participate. Inclusion criteria was established as follows: children ages eight to 18 years, Intelligence Quotient (IQ) greater than 70 and a diagnosis of Asperger’s Disorder, Autistic Disorder, Pervasive Developmental Disorder—Not Otherwise Specified (PDD—NOS) or Autism Spectrum Disorder (ASD) (collectively referred to in this study as ASD) (Carlisle 2012). Children having a service dog were excluded. Due to the wide range in age of participants in this cross-sectional study, diagnosis of ASD was made in several ways. All patients received a best-estimate diagnosis of autism by a physician or psychologist through an evaluation that included a history form comprised of questions related to the Autism Diagnostic Interview-Revised (ADI-R), based on the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV or DSM-IV-TR). Eighty-three percent of those patients received an Autism Diagnostic Observation Schedule and/or ADI-R as part of their evaluation. Intelligence Quotient was evaluated using the following instruments: Leiter International Performance Scale-Revised (n = 26), Wechsler Abbreviated Scale of Intelligence (n = 24), Stanford Binet (n = 15), Wechsler Intelligence Scale for Children—III and IV (n = 3), Mullen Scales of Early Learning (n = 1) and the Wechsler Preschool and Primary Scale of Intelligence (n = 1). Out of 953 children, 285 met the inclusion criteria, and attempts were made to contact the primary caregiver (referred to as the parent) of each child. Current phone numbers were no longer available for 127 parents, 61 did not return telephone messages and 24 parents refused participation. Data were collected from 73 parent participants and three of these were excluded from analysis due to their highly unusual circumstances (Carlisle 2012). The final sample included 70 parents of children with ASD, and the identified children with ASD in the 47 households that owned dogs.
Participating parent ages ranged from 30 to 59 years with a mean of 41.9 years, and 61 were females (Carlisle 2012). Fifty-five of the parents had some type of post-secondary education. Fifty parents were married and the mean number of children was 2.9 per family. Racial and ethnic background was homogeneous with 61 Caucasian, two Hispanic, two Native American, two self-identifying as “Other” and one African American. Child diagnosis included 17 with Asperger’s Disorder, 18 with Autistic Disorder, 28 with PDD—NOS and seven with ASD. The mean age of the children was 13 years and 65 were male, while five were female.
This cross-sectional descriptive study was conducted using a telephone survey. The first hypothesis was that children with ASD living with a dog would have increased social skills, compared with those without a dog (Carlisle 2012). The second hypothesis was that children with ASD living with some type of pet (not excluding dogs) would have increased social skills, compared with those without any pet. The third hypothesis was that children with ASD having a strong attachment to their dog would have increased social skills, compared with those having weaker attachments.
Approval was obtained from the associated university Institutional Review Board with additional approval from the research committee of the Mid-western autism treatment center, which provided the list of potential participants (Carlisle 2012). Upon contact, inquiry was made regarding the child’s ability to respond to eight short questions about dogs using the telephone. Parents (n = 5) who identified their children as unable to complete this task were excluded. No child was offered enrollment without initial enrollment of the parent. In dog owning families with more than one dog, parents were asked to respond to questions based on the dog with which they believed their child was most attached.
Communication—eye contact, reciprocal conversation and appropriate gestures such as thanking someone.
Cooperation—helping others, sharing and taking turns.
Assertion—initiating an exchange with another individual in a social setting, such as asking questions and/or responding to another individual.
Responsibility—being accountable for something and responding to adults
Empathy—demonstrating comprehension of the experiences and emotions of others.
Engagement—initiating participation in activities and conversations with others, as well as making friends.
Self-Control—using restraint and compromise when interacting with others and encountering disagreements and conflict such as teasing.
Total Scores are calculated for SS and PB from individual items within the subscales. The parent forms are designed at a fifth grade reading level and take 10–25 min for completion. Parents were given at $10 gift card following completion of the survey. Their children were not required to participate for receipt of this compensation to the parents for their time.
Children were asked to respond to the Companion Animal Bonding Scale (CABS) (Poresky et al. 1987). Children in families with more than one dog were asked to respond to questions based on the dog in their family that they liked the most. In each case the children chose the same dog as the parent. The CABS is an 8-item instrument used to provide a rating of relationships between humans and their pets. The CABS consists of Likert-type scale response choices of “always,” “generally,” “often,” “rarely,” and “never.” To increase comprehension by children with potentially lower literacy levels, changes were made to the item wording in the CABS. “Companion animal” was replaced with “dog,” “responsive to” was replaced with “responsive to or likes,” and “relationship” was replaced with “relationship or friendship” (Carlisle 2012). Parents were offered the opportunity to listen to their children during the review of the study protocol, assent and collection of data by using their telephone in speakerphone mode. They were also offered the opportunity to write down the rating scale response choices for their child’s use, if they believed their child would benefit from a visual cue in responding to the survey. None of the parents were overheard attempting to influence their children’s responses. At the end of each child’s survey, the children were offered the opportunity to ask questions and/or to make any other comments regarding their dog and/or other pets in their home.
The demographic and CABS data were entered into SAS 9.3 for analysis (“SAS 9.3, 2011”). The SSiS-RS data were entered into the ASSIST software program, which is designed for data entry and analysis of the SSiS-RS (Gresham and Elliott 2008). The level of measurement was interval data. The distributions were normal by a visual check of histograms and by Q-plots, which showed no data points deviating far from the reference line. The equality of variances was established by the Folded F test, which yielded non-significant results; therefore, the assumption of homogeneity was met. For the first hypothesis examining SS differences between children in dog owning and non-dog owning families, a one-sided two-sample t test was used with a significance level of .05 for comparison of the categories (Carlisle 2012). Demographic variables such as income were examined for potential identification of any association(s) between the SSiS-RS and the two categories. This allowed for any adjustment due to potential confounding effects through their use as a covariate(s) in a multivariate model. For the second hypothesis examining the difference in SS for children who had stronger attachment to their dog, a t test was also used with a significance level of .05.
Descriptive statistics were used to examine demographic variables for the two categories of dog owners and non-dog owners. Confidence intervals of 95 % and Chi squared testing were used in computing comparison of the two categories. Pearson’s correlations addressed the relationships between SS and demographic variables, as well as attachment, as measured by the CABS.
Fifty-seven of the parents owned some type of pet including the following: dogs (n = 47), cats (n = 36), fish (n = 11), farm animals (n = 9), rodents (n = 5), rabbits (n = 3), reptiles (n = 3), arachnid (n = 1) and bird (n = 1) (Carlisle 2012). Thirteen parents owned no pets for a total of 70 in the sample. The single dog breed identified most often was Labrador retriever (n = 12), and other breeds such as terriers, sporting and hounds were reported for a total of 25 medium to large dogs. Ten dogs were classified by parents as being small mixed breeds, and a variety of toy breeds were identified for a total of 22 small dogs. The mean dog age was 4.4 years and the mean length of ownership time was 3.9 years. The mean age of the child when the identified dog was acquired was 9.1 years, and the mean number of hours per day that children spent directly interacting with their dog was 1.5, according to parent report.
Social Skills for Dog Owners
Comparison of SSiS-RS scores between children with and without dogs using a t test (one-tailed) (n = 70) (Carlisle 2012)
No Dog (n = 23)
Dog (n = 47)
Total social skills
Total problem behaviors
Social Skills for Pet Owners
Comparison of SSiS-RS scores between children with and without pets using a two-sample t test (one-tailed) (n = 70) (Carlisle 2012)
No pet (n = 13)
Pet (n = 57)
Total social skills
Total problem behaviors
Attachment to Dogs
Descriptive statistic of CABS by parent perception of child’s attachment (n = 47) (Carlisle 2012)
Not very attached
Pearson’s correlation of attachment measures and SSiS-RS scores (n = 47) (Carlisle 2012)
Social skills r (p value)
Problem behaviors r (p value)
Time interacting with dog
Total CABS score
Most of children responded readily and quickly to the CABS, without any difficulty, although the investigator was unknown to any of the children and the survey occurred using the telephone. Only 10 of the children in the present study required assistance with completion of the CABS, and this included repeating the questions and/or using written response options as a visual cue for answering. Two children were reported by parents to dislike talking on the telephone under typical circumstances, yet answered the CABS readily. The parents attributed this unlikely responsiveness to their children’s desire to talk about their dogs.
The rate of pet ownership among families in the United States is 66 %, according to the American Veterinary Medical Association (2012), yet 81 % of families in the present study owned pets (Carlisle 2012). Dogs were the most commonly identified type of pet, but the children were also reported by parents to have bonds to other household pets including rabbits and cats. In a study by Carlisle (2014), families of children with ASD were found to make the decision to own pets to aid their children in learning responsibility and to provide companionship.
While trained dogs have been associated with increased SS in children with ASD, limited evidence of this same benefit has been reported in the literature regarding dogs that live as pets, without any professional training. Increased SS in children with ASD have been associated with interaction with other trained animals such as horses and dolphins, as well as guinea pigs in a classroom setting (O’Haire 2013). The trend toward increased SS found in children living with dogs in the present study indicates similar findings may be true. The present study mirrors these earlier findings, wherein children who lived with some type of pet (not excluding dogs) had significantly greater SS for “assertion.” The positive relationship between children and animals may be acting as the social catalyst. When animals are the variable identified with this pro-social type of behavior, it is referred to as the social catalyst effect and this is well documented among adults.
A core characteristic of children with ASD relates to their desire for consistency within their environment (Kanner et al. 1943), and the finding of a significant association of greater SS with longer ownership of their dog provides a first look at a phenomenon that will require further examination. Conversly, an earlier study by Grandgeorge et al. (2012) found that the new arrival of a pet, rather than a pet that had always lived with the child, was associated with increased SS. The children in that study had cognitive impairments with a mean IQ of 42, while children in the present study, had IQ scores greater than 70. These conflicting findings may indicate that intellectual disability plays a role in the interaction of children with ASD and their pets. Additionally, while there was no difference in the SS of children associated with older dogs, longer ownership was significantly associated with fewer PB in the present study (Carlisle 2012). Child age was not a factor in these associations, but it is unknown whether the child’s familiarity with the presence of their dog, or the temperament of a mature dog may have played a role.
Children with ASD have deficits in pro-social behaviors (Hobson and Lee 1998) and sensory integration (Iarocci and McDonald 2006). This includes a struggle with the ability to comprehend complex facial expressions and non-verbal communication with humans (Dawson et al. 2005; Hobson and Lee 1998; Langdell 1978). There is a lack of necessity to perform these skills when interacting with dogs, and this may provide a rationale for the attachment of children with ASD to dogs (Solomon 2012). The present study supports this with 92 % of the children in dog owning families described by parents as “attached” or “very attached” to their dogs (Carlisle 2012). The issue of multiple pets in the homes of the families studied creates a potential confounder of the findings. However, the reality of most multiple pet-owning families is that particular family members bond with particular pets in ways and degrees different from other pets in their family. This was definitely the case in the families studied; therefore, it was very easy for both the parents and the children who participated to identify with which dog the child with ASD was most strongly bonded. The fact that they did and responded to the instruments in reference to this dog may have strengthened the findings because the children made a clear distinction between this dog and other dogs in the home. Given the higher pet ownership rate in these families than in the general population (81 versus 66 % respectively), large numbers of families would be eliminated from the research if only families owning one pet were included.
While the CABS has been cited for use among adults and children for over two decades, its use had not been explored for children with ASD. Given the apparent ease of completion and comparable rating of attachment by parents and children, further exploration of the use of the CABS may aid in determining its reliability as a useful tool in identifying pet bonding for older children with ASD who do not have an intellectual disability. The significantly stronger bond with small dogs in the present study may reflect less sensory stimulation experienced by just size of the dog. Further investigation of the phenomenon of dog size is needed to test this assumption. The SS benefit found in this study associated with living with some type of pet, including cats and rabbits, reinforces the possibility of a preferred relationship to smaller animals by many children. While significant, this association was weak and will require further investigation with larger samples sizes.
Most children with ASD interacted daily with their dogs and reported being “attached” to their dogs. The identification of greater SS in the subscale item of “assertion” for children with ASD living with some type of pet, and a trend toward greater SS for children living in dog owning families was found (Carlisle 2012). These findings, along with the highly heterogeneous characteristics of children with ASD, may indicate that some children would benefit through increased SS more from one type of pet than another. Future investigations focusing on the relationships of children with ASD and different types of pets may aid in identifying those pets associated with the greatest SS benefits for children with ASD, and those with which the children may be most likely to form a bond.
Several limitations are acknowledged for the present study. Parents who agreed to participate may have had different experiences than those who did not participate. Phone numbers for 187 potential participants from the data-base were no longer in working order, 61 parents did not return recruitment phone calls and 24 parents refused participation (Carlisle 2012). Dog ownership is common among families with children; however, not all people desire dog ownership or find interaction with dogs pleasant. Also, the high prevalence of ownership of some type of pet in the present study provides a confounding effect that limits the ability to isolate findings for dog ownership. Generalization for children of parents who do not own pets is limited due to the small number of non-pet owning families in the present study.
The present study excluded children with ASD having IQs less than 70 and younger than age 8 years to increase the likelihood of the children’s ability to complete the CABS; however, this limits the identification of findings to these groups (Carlisle 2012). Children outside these inclusion criteria may have benefited from living with a dog. In addition to this, it is acknowledged that the comprehension required by the children in the present study for categorization using Likert scales may have been difficult, and the results may have reflected this challenge. Self-reporting can be challenging for children with ASD; however, children in this study reported the same strong bonds to their dogs that the parents reported. It is acknowledged that a different rating scale was used for parents and children. While children self-reported and parents were not overheard during the telephone survey to be attempting to influence their children’s responses, children may have been influenced to respond in a manner not accurately representing their personal perception. As this was a first attempt to investigate pet attachment for children with ASD, further investigation will be needed to examine the reliability of the use of the CABS in this special population.
Size of the dogs was categorized in the study based on parent description of the size and breed of the dogs. It is acknowledged that an objective weight of the dogs was not possible in this telephone survey. The availability of this measure in future studies would improve the validity of a dog size variable.
Finally, there is an inherent self-selection bias in human–animal interaction research and possible that parents in the present study chose dog ownership because their children had more social skills. As it would not be ethical to place a dog with a family that does not like dogs, this is a known limitation. This study was also limited to a cross sectional investigation and further research, including randomized control trials, may aid in minimizing this bias.
This preliminary survey provides an introduction to the study of pet ownership in families of children with ASD. Test scores in the present study showed that children with ASD in dog owning families had a trend toward greater SS than children with ASD in non-dog owning families. Children with ASD in families owning some type of pet (not excluding dogs) had greater pro-social behaviors in “assertion,” when compared with children with no pets. Further investigation is needed to explicate the direction of the relationship between “assertion” and pet ownership. It is unknown whether children with stronger skills in “assertion” are more likely to have pets, or whether these skills are enhanced through pet ownership. Understanding this link would be important for children who have deficits in these social skills. In addition to this, the strong bond between children with ASD and their dogs, yet greater social skills associated with any pet, indicates further research is needed to elucidate the potential benefits of other types of pets in the home.
The author acknowledges the University of Missouri Thompson Center for Autism and Neurodevelopmental Disorders for the use of their data set in recruitment. A grant to support this study was awarded from Sigma Theta Tau—Alpha Iota Chapter. This paper was prepared from the author’s doctoral dissertation, acknowledging the support of her dissertation committee members, Dr. Rebecca A. Johnson, Dr. Lawrence Ganong, Dr. Debra Gayer and Dr. Micah Mazurek. A poster presentation of the data was made at the Midwest Nursing Research Society conference in Chicago, Illinois on March, 2013.