Study on Human Perception of Emotion in Dogs: Results

Q: Was the level of dog experience associated with observational techniques?
A: Yes, compared to less-experienced individuals, individuals with greater dog experience were more likely to attend to the dogs' ears and facial expressions and less likely to attend to the legs and tail. In addition, the number of observed features and behaviors increased with experience.

Q: Can you show me some examples of videos?
A: Yes, here are two examples (video playback requires Javascript-enabled browser):

The above dog was primarily perceived as fearful by an initial expert panel. In the general sample, the probability of categorizing the above dog as fearful was 6% among those who had never owned a dog, 36% among owners, 59% among those who had worked professionally with dogs for less than ten years, and 56% among those who had worked professionally with dogs for ten or more years. In contrast, the probability of perceiving the dog as happy decreased as experience increased (50%, 31%, 15%, and 11%).

Q: Who participated in this study?
A: Participants included 152 individuals who had never owned a dog, 1,462 individuals with dog-owning experience, 307 dog professionals with less than ten years of professional experience, and 242 professionals with ten or more years of experience. Among the professionals, 70% were dog behavior professionals, such as trainers.

Q: Was the level of dog experience associated with perceptions of emotion in dogs?
A: Yes, the level of dog experience was a significant predictor of emotion ratings and categorizations. Differences by experience were more evident in interpretations of negative than positive emotional displays (as pre-assessed by an initial expert panel). Less-experienced individuals tended to provide more positive emotion ratings of negatively-valenced behavior than more-experienced individuals. In addition, they were more likely to categorize negatively-valenced behavior as happy, rather than fearful. As shown in the figure below, the likeliness of categorizing negatively-valenced behavior as fearful increased dramatically with experience.

An expert panel unanimously categorized the above dog as feeling happy. In the general sample, the probability of categorizing the dog as happy was 89% among individuals who had never lived with a dog, 96% among owners, 96% among individuals who had worked professionally with dogs for less than ten years, and 98% among individuals who had worked professionally with dogs for ten or more years. As mentioned previously, differences by experience were more evident in interpretations of negatively-valenced than positively-valenced behavior (see previous video).

Q: How do you know what the dogs in the videos were feeling?
A: We can never state with certainty exactly what the dogs were feeling. However, all videos were pre-assessed by a highly experienced panel of dog behavior professionals. In addition, the behaviors that the expert panel and many participants described as indicative of fear have been associated in previous studies with stress and fear in dogs. These behaviors include, for example, shaking, yawning, salivation, freezing, panting, paw-lifting, vocalizing, crouching into a low posture, flattening of the ears, and low tail position. See Resources for list of references.

Likelihood of fearful categorizations of negatively-valenced behavior by level of dog experience (Low-Exp = never owned a dog, Owners = dog-owning experience, Prof<10 = professional dog experience for less than ten years, Prof10+ = professional dog experience for ten or more years)