Abstract
Breed is a ubiquitous classifier for both companion and working dogs, with profound implications for perceived and/or anticipated behavior, cognition, physiology, athleticism, and "temperament". Owners, breeders, shelters and others are frequently asked to report on the known or suspected breed(s) of their dogs in various scenarios affecting dog and human wellbeing. However, there is a paucity of research on breed reporting accuracy. Using Dog Aging Project data, we compare genetic breed results with owner-reported ancestry for 5673 dogs. We find that 80% of dog owners identified their dog's breed as matching the genetic reports. Among those few who reported disagreement with the genetic results, the majority indicated having dogs belonging to a breed not included in the reference panel. Additional causes for the 6% of samples reported as disagreeable include a stricter owner definition of match for single-breed dogs (e.g. 99.9% vs 100%) and genetic results showing a mix of two closely-related breeds for an owner-reported single-breed dog. Our findings indicate that people are usually aware of their dog's genetic ancestry, which is important for the validity of using breed as a covariate in population-level studies, and for managing life history outcomes for dogs relative to breed-related traits and biases.