Objectives: The present study examines social interaction in miniature and standard poodles, breeds previously described as being less social overall than other breeds, using an owner-report survey with questions adapted from the human Autism Diagnostic Observation Schedule (ADOS). The primary aim of this study was to determine the existence of subgroups within each breed that demonstrate significantly less social behavior than other dogs within the breeds. We also sought to confirm factor loadings determined by a previous study utilizing the owner-report survey and to compare identified subgroups in terms of these factors.
Methods: Survey questions adapted from the ADOS to pertain to dog social interaction behavior were included as part of an online survey used to collect data. Questions were structured as likert-type scales, ranging from 0 to 2 to 0 to 4, depending on the question. Data analyzed was restricted to fully completed surveys for AKC or CKC-registered miniature and standard poodles.
Results: Cluster analyses using standardized data were utilized to determine the existence of a subgroup of dogs within each breed that demonstrates significantly reduced social behavior as compared to other dogs within the breeds. For both breeds a k-means cluster analysis for two clusters identified one small and one large cluster differing significantly in social behavior. In both breeds, the smaller clusters represent subgroups of dogs with lower scores on the social interaction questions. A previous study identified three factors examined by the survey: initiation of reciprocal social behaviors, response to social interaction, and communication. Although the sample sizes in each breed for the present study prohibited confirmatory factor analyses, experimental factor analyses revealed similar loadings for individual questions onto each of the three factors. One-way ANOVAs or Welch’s variance-weighted ANOVAs were used to compare the identified clusters in terms of factor scale scores. The less social subgroup of both miniature and standard poodles demonstrated significantly decreased scale scores across all three factors, as compared to the more typical subgroups. Thus, for both breeds, the dogs identified as having reduced social behavior differed from the remainder of the dogs on all the three social domains assessed by the survey questions.
Conclusions: The identification of these subgroups demonstrating reduced abilities across several domains of social interaction suggests a potential genetic component in social behavior that could be further examined in both dogs and humans. Future investigation will involve behavioral observation and genetic analysis for potential mutations underlying atypical social behavior in miniature and standard poodles.