Conspecific observational learning by adult dogs in a training context

The experimental procedure consisted of two phases: Phase 1, intended to ascertain that dogs could not perform the selected exercise when requested by their handlers; Phase 2, to assess whether they would perform the same exercise after the observation of a conspecific demonstrator. The tests were performed outdoors in a fenced training area and one out of two dynamic exercises were selected for each dog: jumping on a trunk or hop on a slide for children. The outcome of Phase 2 was coded into a binary variable as successful or unsuccessful and a generalized linear model with binary logistic link function was used to analyze outcomes. The model included the dog's age as a covariate and the experimental group, level of experience, sex and all two-ways interactions as fixed factors. The dogs' probability to replicate the action increased significantly after demonstration by a conspecific compared to a control group and such probability improves with age. This study supports the usefulness of the intraspecific observational learning in adult dogs for training purposes.